WO1997042607A2

WO1997042607A2 - Apparatus and method for generating a sheet metal bending plan

Info

Publication number: WO1997042607A2
Application number: PCT/US1997/007473
Authority: WO
Inventors: Kensuke Hazama; Kalev Kask; Satoshi Sakai; Moshe Edward Schwalb
Original assignee: Amada Metrecs Co., Ltd.; Amadasoft America, Inc.
Priority date: 1996-05-06
Filing date: 1997-05-06
Publication date: 1997-11-13
Also published as: US7197372B2; US6327514B1; DE69737913T2; US7266419B2; EP1038272B1; US20060106757A1; US20020038163A1; DE69737913D1; EP1038272A2; US5971589A; WO1997042607A3

Abstract

An apparatus and method is provided for managing and distributing design and manufacturing information throughout a factory in order to facilitate the production of components, such as bent sheet metal components. In accordance with an aspect of the present invention, the management and distribution of critical design and manufacturing information is achieved by storing and distributing the design and manufacturing information associated with each job. By replacing the traditional paper job set-up or work sheet with, for example, an electronically stored job sheet that can be accessed instantaneously from any location in the factory, the present invention improves the overall efficiency of the factory. In addition, through the various aspects and features of the invention, the organization and accessibility of part information and stored expert knowledge is improved.

Description

APPARATUS AND METHOD FOR MANAGING AND DISTRIBUTING DESIGN AND MANUFACTURING INFORMATION THROUGHOUT A SHEET METAL PRODUCTION FACILITY

RELATED APPLICATION DATA

This application claims the benefit of U.S. Provisional Application No. 60/016,958,

filed May 6, 1996, entitled "Apparatus and Method for Managing and Distributing Design

and Manufacturing Information Throughout a Sheet Metal Production Facility," the

disclosure of which is expressly incoφorated herein by reference in its entirety.

COPYRIGHT NOTICE

A portion of the disclosure of this patent document contains material which is subject

to copyright protection. The copyright owner has no objection to the facsimile reproduction

by anyone of the patent disclosure, as it appears in the U.S. Patent and Trademark Office

patent files or records, but otherwise the copyright owner reserves all copyright rights

whatsoever.

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention generally relates to the field of manufacturing and to the

production of components, such as sheet metal components. More particularly, the present invention relates to an apparatus and method for managing and distributing design and

manufacturing information throughout a factory in order to facilitate the production of bent sheet metal components.

2. Background Information

Traditionally, the production of bent sheet metal components at, for example, a

progressive sheet metal manufacturing facility, involves a series of production and

manufacturing stages. The first stage is a design stage during which a sheet metal part

design is developed based on a customer's specifications. A customer will typically place

an order for a particular sheet metal component to be produced at the facility. The

customer's order will usually include the necessary product and design information so that

the component may be manufactured by the factory. This information may include, for

example, the geometric dimensions of the part, the material required for the part (e.g., steel, stainless steel, or aluminum), special forming information, the batch size, the delivery date,

etc. The sheet metal part requested by the customer may be designed and produced for a

wide variety of applications. For example, the produced component may ultimately be used

as an outer casing for a computer, an electrical switchboard, an armrest in an airplane, or part

of a door panel for a car.

During the design stage, a sheet metal part design may be developed by the design

office of the manufacturing facility using an appropriate Computer- Aided Design (CAD)

system. Based on a customer's specifications, a 2-dimensional (2-D) model of the sheet

metal part may be developed by a programmer with the CAD system. Typically, a customer

will provided a blueprint containing one or more drawings of the component and the critical

geometric dimensions of the part. The blueprint may also indicate any special forming or

marking to be included in the part, as well as the location of holes or other types of openings

on the surface(s) of the sheet metal part. The design programmer will often use this blueprint to develop a 2-D model on the CAD system. The 2-D model may include a flat

view and one or more other perspective views of the sheet metal part, with bending line

and/or dimensional information.

Before actual bending of the sheet metal part takes place, the part must first be

punched and/or cut from initial stock material. Computer Numerical Control (CNC) or

Numerical Control (NC) systems are typically used to control and operate punch presses and

plasma or laser cutting machinery to process the stock material. In order to facilitate

processing of the stock material, a Computer-Aided Manufacturing (CAM) system or

CAD/CAM system can be used by a design programmer to generate control code based on

the 2-D model. The control code may comprise a part program that is imported to and

utilized by the punch press and/or cutting machinery to punch or cut the sheet metal

component from the stock material.

The next stage in the production process is a bending plan stage. During this stage, a bending plan is developed by a bending operator at the shop floor. The operator will

normally be provided with the blueprint or 2-D drawing of the component, along with one

or more samples of the cut or punched stock material. With these materials, the bending

operator will develop a bending plan which defines the tooling to be used and the sequence

of bends to be performed. The bending workstation may include CNC metal bending

machinery, such as a CNC press brake, that enables the operator to enter data and develop

a bending code or program based on the bending plan.

Once the bending plan is developed, the operator will set up the workstation for

initial testing of the bending sequence. During this testing stage, the punched or cut stock

material will be manually loaded into the press brake and the press brake will be operated

to execute the programmed sequence of bends on the workpiece. The operator will analyze - 4 - the final bent sheet metal part and inspect it for conformance with the customer's

specification. Based on the results of the initial runs of the press brake, the operator may

modify the bending sequence by editing the bending program. The operator may also

provide feedback to the design office so that the sheet metal part design can be appropriately

modified. Further testing will typically be conducted until the bent sheet metal component

is within the required design specifications.

One of the final stages in the production process is the bending stage. After the

bending plan has been developed and tested, the bending operator will set up the required

tooling at the bending station and operate the press brake based on the bending plan and the

stored bending program or code. Job scheduling is also performed in order to ensure that

the necessary amount of punched or cut stock material will be available on time at the

bending station, and so that other jobs will be completed by the requested delivery dates.

Job scheduling may be developed or modified by a shop floor foreman during the earlier

stages of the production process and/or concurrently throughout the entire process. After

the final bent sheet metal parts have been produced, the parts may then be assembled and

packaged for shipping to the customer.

The conventional production and manufacturing process described above suffers

from several drawbacks and disadvantages. For example, although the design and

manufacturing data for each customer's order is normally archived physically (e.g., by paper

in a file cabinet) or electronically (e.g., by storing on a disk or magnetic tape), such data are

normally stored separately and not easily retrievable. Further, in most factory settings, the

distribution of critical job information takes the form of a paper job or work sheet that is

distributed throughout the factory floor. As a result, data is often lost or damaged, and it is

difficult to search for both the design and manufacturing data relating to a previous or similar job. In addition, due to the inefficient manner in which the data is stored, valuable

time is lost in attempting to distribute the design and manufacturing information to the shop

floor and to other locations throughout the factory. Considerable manufacturing time is also

lost during the development of the sheet metal part design and bending plan, since the

development of the part design and bending plan is primarily performed by the design

programmer and bending operator, and relies heavily on the individual's knowledge, skill

and experience.

In recent years, there have been developments and attempts to improve the

conventional sheet metal manufacturing process and to improve the efficiency of the overall

process. For example, the use and development of 2-D and 3-dimensional (3-D) modeling

in commercially available CAD/CAM systems has facilitated and improved the production

process and modeling of bent sheet metal components. The design programmer and

operator can now utilize both the 2-D and 3-D representations to better understand the

geometry of the part and more efficiently develop a part design and bending code sequence.

The ability to store and transfer data electronically has also improved the flow of

information from the design office to locations on the shop floor. With the advancement of

computers and data communication networks, it is no longer necessary to search through a

cabinet or file of old paper tapes or magnetic disks.

Despite such advancements, there is still a need to improve the organization and flow

of design and manufacturing information throughout the factory environment. For example,

conventional manufacturing systems do not logically associate both critical design and

manufacturing information associated with each customer's order so that it may be easily

accessed and retrieved from any area in the factory. Previous systems also fail to provide

the ability to search previous job information based on various criteria, such as the features and attributes of the sheet metal component. The ability to search and retrieve previous job

information based on, for example, an identical or similar part search, would greatly enhance

the overall production process and reduce the required manufacturing time for future jobs.

Past attempts also fail to facilitate the development of the

sheet metal part design and bending plan by the design programmer and shop floor operator.

While the introduction of 2-D and 3-D modeling systems have enabled the designer to have

a better understanding of the shape and geometry of the part, such systems have not reduced

the burdens placed on the design programmer and shop floor operator. For example, such

systems have not enabled the design programmer to easily convert an existing 2-D CAD

model into a 3-D representation. In addition, while 2-D and/or 3-D drawings of the

component may be provided to the shop floor operator to assist in the development of the

bending plan, the operator must still determine and develop the tooling requirements and

bending sequence by hand and/or experimentation.

SUMMARY OF THE INVENTION

In view of the foregoing, the present invention, through one or more of its

various aspects, embodiments and/or specific features or sub-components thereof, is

provided to bring about one or more objects and advantages, such as those specifically noted

below.

A general object of the present invention is to provide an apparatus and method for

managing and distributing design and manufacturing information throughout a factory in

order to facilitate the production of components, such as bent sheet metal components.

A further object of the present invention is to provide an apparatus and method that

prevents the loss or destruction of critical job information, and that enhances the efficiency and organization of stored expert knowledge at, for example, a progressive sheet metal

production facility.

Another object of the invention is to provide an apparatus and method for logically

storing both the design and manufacturing information for each customer's order, so that it

may be easily accessed and retrieved from any area in the factory.

Yet another object of the present invention is to provide an apparatus and method

for managing and distributing design and manufacturing information, wherein the job data

is stored at a central database or file server in a logical fashion so that it may be easily

searched and retrieved from any location throughout the factory. The job data may provide not only the design and manufacturing information associated with the job, but also the

actual bend code for executing the required bending operations.

Still another object of the present invention is to provide an apparatus and method

for searching previous job information, including design and manufacturing information,

based on various search criteria. The search criteria may include, for example, the basic

features and attributes of the sheet metal component to be manufactured, so that previous

job information relating to an identical or similar part can be utilized to reduce the overall

manufacturing time of future jobs.

Another object of the present invention is to replace the traditional paper job or work

sheet, associated with each customer's order, with an electronic job sheet that can be

instantaneously accessed from any location in the factory. The electronic job sheet may be

displayed at any location and include critical design and manufacturing information,

including the 2-D and/or 3-D model view of the component, the tooling selection, the

optimum bending sequence, the required staging information, and the bar code or

identifϊcationnumber associated with the job. The electronic job sheet may also include an audio and or video portion recorded by a bending operator to indicate, for example, any

special instructions or procedures that may be helpful when running the same job or a

similar job again in the future.

Another object of the invention is to shorten the time required to analyze a part

drawing by providing 2-D and 3-D computerized views of the sheet metal part. Various

viewing modes may be provided, including a solid 3-D viewing mode, a 3-D wire frame

viewing mode, a 2-D flat screen viewing mode, and an orthographic viewing mode.

Different viewing functions may also be provided, including zooming, panning, rotating and

auto-dimensioning, to facilitate analysis of the sheet metal part.

A further object of the invention is to provide an apparatus and method that facilitates the development of the sheet metal part design and bending plan by the design

programmer and shop floor operator. For example, it is an object of the present invention

to enable the design programmer to easily develop a 3-D representation of the component

from an existing 2-D model. It is also another object of the invention to provide a graphical

user interface to shorten the time required to develop the bending plan and programmed

bending code.

The present invention, therefore, is directed to an apparatus and method for

developing a bend model of a part to be produced in an intelligent production facility,

wherein the part includes a plurality of faces and at least one bendline. The apparatus may

comprise a receiving system for receiving initial part information relating to the part,

including data relating to a representation of the part within a first predetermined coordinate

space. A face detecting system may also be provided for detecting, within the first

predetermined coordinate space, the faces of the part based on the initial part information.

In addition, the apparatus may include a bendline identification system for identifying at least one bendline of the part based on the detected faces, and a system for generating

additional part information, relating to the part and including data relating to a representation

of the part within a second predetermined coordinate space, by performing a predetermined

operation on each of the faces detected by the face detecting system. The predetermined

operation may be performed based upon at least the initial part information and at least one

bendline identified by the bendline identification system.

The first predetermined coordinate space may comprise a 2-D coordinate space and

the second predetermined coordinate space may comprise a 3-D coordinate space, wherein

the predetermined operation comprises a folding operation performed on the faces detected by the face detecting system. The folding operation may include rotating and translating

each of the faces detected by the face detecting system relative to at least one bendline

identified by the bendline identification system. In addition, the initial part information may

further comprise a bend angle amount relating to at least one bendline of the part, whereby the folding operation is performed based on the bend angle amount.

According to another aspect of the invention, the first predetermined coordinate

space may comprise a 3-D coordinate space and the second predetermined coordinate space may comprise a 2-D coordinate space, wherein the predetermined operation comprises an

unfolding operation performed on the faces detected by the face detecting system. The

unfolding operation may include rotating and translating each of the faces detected by the

face detecting system relative to at least one bendline identified by the bendline

identification system. In addition, the initial part information may further comprise a bend

angle amount relating to at least one bendline of the part, whereby the unfolding operation

is performed based on the bend angle amount.

The part to be produced may comprise a sheet metal part, and the data relating to a representation of the part within the first predetermined coordinate space may include

coordinate data and/or vector data. Further, the faces of the part may comprise the base

surface(s) of the part as well as the folded surfaces of the part.

According to another aspect of the invention, the initial part information may

comprise data relating to a representation of the part in a 3-D coordinate space, and the data

may include thickness data of the part in 3-D coordinate space.

The apparatus for developing a bend model may also include an auto-trimming and

cleanup system for performing an auto-trimming and cleanup operation on the data relating

to the initial part information to prepare the data for the face detecting system and the

bendline identification system. The data may comprise part entity data representing, at least, line entities and bendline entities of the part, and the auto-trimming and cleanup system may

further comprise a system for detecting intersection points of the entities and for selectively

breaking the entities at detected intersection points, and a system for assigning the resultant

broken entities to have a common endpoint based on the detected intersection points. The

auto-trimming and cleanup system may also comprise a system for detecting open

intersection areas between adjacent entities and for selectively connecting the adjacent

entities by assigning a common endpoint to the adjacent entities.

Open intersection areas may be detected, by the system for detecting open

intersection areas, when the endpoints of the adjacent entities are determined to be within

a predetermined distance from one another.

According to yet another feature of the invention, the data relating to the initial part

information may comprise part entity data representing, at least, line entities of the part, and

the face detecting system may be adapted to perform a loop and entity analysis of the part

based on the part entity data to detect the faces of the part. The loop and entity analysis may initially be performed on an outside boundary of the part and then performed on inner

boundaries and areas of the part. The face detecting system may generate an initial linked

list of entities when performing the loop and entity analysis on the outside boundary of the

part, such that the initial linked list of entities define an outward loop and boundary of the

part. The face detecting system may further generate an additional linked list of entities

when performing the loop and entity analysis on the inside boundaries and areas of the part,

such that the additional linked list of entities define inner loops and boundaries of the part.

In addition, the face detecting system may further comprise a system for generating a loop

tree based on the outward loop defined by the initial linked list of entities and the inner loops

defined by the additional linked list of entities. Further, the face detecting system may

detect the faces of the part based on the loop tree and the sequence of boundaries defined by

the initial linked list entities and the additional linked list of entities.

The bendline identification system of the invention may comprise a system for

analyzing the initial linked list of entities and the additional linked list of entities to

determine common line entities between the faces detected by the face detecting system.

Bendlines may be identified based on the detection of one of the faces having only one common line entity with another one of the faces. In addition, the bendline identification

system may apply predetermined heuristics to identify bendlines of the part when the system

for determining common line entities detects that there are more than one common line

entity between the faces. The heuristics may comprise identifying bendlines of the part such

that a minimum number of total bendlines are identified for the part. The heuristics may

also comprise identifying bendlines of the part, when one of the faces have more than one

common line entity with another one of the faces, based on the common line entity that has the longest length. According to another feature of the invention, a system for receiving a deduction

amount may be provided to receive a deduction amount related to the part. A system for

compensating for bend deduction, when performing the predetermined operation on the

faces, based on the deduction amount may also be provided. The system for compensating

for bend deduction may increase a dimensional length of the faces by one half of the

deduction amount on each side of the bendline of the part when performing a folding

operation. The system for compensating for bend deduction may also be adapted to decrease

a dimensional length of the faces by one half of the deduction amount on each side of the

bendline of the part when performing an unfolding operation.

The method for developing a bend model may comprise the following steps: receiving initial part information relating to the part, the initial part information including

data relating to a representation of the part within a first predetermined coordinate space;

detecting, within the first predetermined coordinate space, the faces of the part based on the

initial part information; identifying at least one bendline of the part based on the faces

detected by the detecting; and generating additional part information, including data relating

to a representation of the part within a second predetermined coordinate space, by

performing a predetermined operation on each of the faces detected by the detecting, the

operation being performed based upon the initial part information and at least one bendline

identified by the identifying.

the second predetermined coordinate space may comprise a 3-D coordinate space. In

addition, the method may further comprise performing a folding operation on the faces

detected by the step of detecting. The folding operation may include rotating and translating

each of the faces relative to at least one bendline identified by the step of identifying. Further, the initial part information may comprise a bend angle amount relating to at least

one bendline of the part, whereby the folding operation is performed based on the bend angle

amount.

According to another feature of the invention, the first predetermined coordinate

space may comprise a 3-D coordinate space and the second predetermined coordinate space

may comprise a 2-D coordinate space. The method may also further comprise performing

an unfolding operation on the faces detected by the step of detecting. The unfolding

operation may include rotating and translating each of the faces relative to at least one

bendline identified by the step of identifying. Further, the initial part information may

comprise a bend angle amount relating to at least one bendline of the part, whereby the

unfolding operation is performed based on the bend angle amount.

The method may also comprise performing an auto-trimming and cleanup operation

on the data of the initial part information before detecting the faces and identifying the at

least one bendline. The data of the initial part information may comprise part entity data

representing, at least, line entities of the part, and the step of detecting of the faces may

comprise performing a loop and entity analysis of the part based on the part entity data to

detect the faces of the part. The loop and entity analysis may initially performed on an

outside boundary of the part and then performed on inner boundaries and areas of the part.

In addition, according to the invention, the step of identifying may further comprise

applying predetermined heuristics to identify bendlines of the part when more than one

common edge between the faces is detected. The heuristics may comprise identifying

bendlines of the part such that a minimum number of total bendlines are identified for the

part. The heuristics may also comprise identifying bendlines of the part, when one of the

faces have more than one common line entity with another one of the faces, based on the common line entity that has the longest length.

The present invention also encompasses a system for developing a bend model of a

part to be produced in a production facility, wherein the part includes a plurality of faces and

at least one bendline. The system comprises means for receiving initial part information

relating to the part, wherein the initial part information including data relating to a

representation of the part within a first predetermined coordinate space. Detecting means

may also be provided for detecting, within the first predetermined coordinate space, the

faces of the part based on the initial part information. The system may also include means

for identifying at least one bendline of the part based on the faces detected by the face

detecting means, and means for generating additional part information, including data

relating to a representation of the part within a second predetermined coordinate space, by

performing a predetermined operation on the faces detected by the detecting means. The

predetermined operation may be performed based on, at least in part, the bendline identified

by the bendline determining means of the system.

According to another aspect of the invention, a system and method is provided for

developing a bend model of a part to be produced in an intelligent production facility. The

system may comprise a receiving system for receiving initial part information relating to the

part, wherein the initial part information comprises respective representations of a plurality

of views the part in 2-D coordinate space, and each of the representations include part

thickness representations of the part. A cleanup operation system may also be provided to

perform a 2-D cleanup operation on the initial part information to eliminate any extraneous

information and to identify each of the representations. The system may also include a part

thickness elimination system for selectively eliminating the part thickness representations in each of the identified representations to provide modified representations of the views of the part in 2-D coordinate space with no thickness, and a system for developing a

representation of the part in 3-D coordinate space based on the modified representations of

the part in 2-D coordinate space with no thickness.

The initial part information may comprise part entity data representing, at least, line

entities of the part, and the cleanup operation system may include a break and trimming

system for detecting intersection points of the entities and for selectively breaking the

entities at detected intersection points. Further, the break and trimming system may assign

the resultant broken entities to have a common endpoint based on the detected intersection

points. The break and trimming system may further comprise a system for detecting open

entities by assigning a common endpoint to the adjacent entities.

Open intersection areas may be detected, by the system for detecting open intersection areas,

when the endpoints of the adjacent entities are determined to be within a predetermined

distance from one another.

In accordance with another aspect of the invention, the cleanup operation system

may comprise a system for developing a connectivity graph structure based on the initial

part information, such that the cleanup operation system eliminates extraneous information

based on the connectivity graph structure. The extraneous information that is eliminated

may comprise unconnected line entities, wherein the unconnected line entities relating to,

at least, dimension lines.

The initial part information may also comprise key words for identifying part entity

data and extraneous information relating to text. The cleanup operation system may

eliminate the extraneous information relating to text based on the key words included in the initial part information. According to yet another feature of the invention, the cleanup operation system may

comprise a system for detecting, based on the initial part information, the representations of

the top view, the front view, and the right side view of the part. As disclosed herein, the

plurality of views of the part may comprise a top view, a front view, and a right side view

of the part in 2-D coordinate space. In addition, the developing system may include a

system for performing a projection operation to develop the representation of the part in 3-D

coordinate space based on the representations of the part in 2-D coordinate space. The

projection operation may comprise detecting the relative depths of each of the plurality of

views and projecting each of the plurality of views into 3-D coordinate space.

The method for developing a bend model may comprise the steps of: receiving initial

part information relating to the part, the initial part information comprising respective

representations of a plurality of views the part in 2-D coordinate space, each of the

representations including part thickness representations of the part; performing a 2-D

cleanup operation on the initial part information to eliminate extraneous information and to

identify each of the representations;selectivelyeliminatingthe part thickness representations

in each of the identified representations to provide modified representations of the views of

the part in 2-D coordinate space with no thickness; and developing a representation of the

part in 3-D coordinate space based on the modified representations of the part in 2-D

coordinate space with no thickness.

entities of the part, and the step of performing may comprise detecting intersection points

of the entities and selectively breaking the entities at detected intersection points, with the

resultant broken entities being assigned to have a common endpoint based on the detected

intersection points. The step of performing may also comprise detecting open intersection areas between adjacent entities and selectively connecting the adjacent entities by assigning

a common endpoint to the adjacent entities. Open intersection areas may be detected when

the endpoints of the adjacent entities are within a predetermined distance from one another.

The method may also comprise the steps of developing a connectivity graph structure

based on the initial part information and eliminating extraneous information from the initial

part information based on the connectivity graph structure. The extraneous information may

comprise unconnected line entities, the unconnected line entities relating to. at least, dimension lines.

In addition, the initial part information may comprise key words for identifying part entity

data and extraneous information relating to text, wherein the step of performing includes the

step of eliminating the extraneous information relating to text based on the key words.

According to another feature of the invention, the step of selectively eliminating the

part thickness may comprise prompting a user to identify the part thickness representations

to be eliminated in each of the plurality of views and to identify a dimension of the part to

be retained in each of the plurality of views. The dimension of the part may comprise one

of an outside dimension or an inside dimension of the part. Further, the step of developing

may comprise performing a projection operation to develop the representation of the part in

3-D coordinate space based on the modified representations of the part in 2-D coordinate

space. The projection operation may include detecting the relative depths of each of the

plurality of views and projecting each of the plurality of views into 3-D coordinate space.

Further features and/or variations may be provided in addition to those noted above.

For example, the invention may be directed to various combinations and subcombinations

of the above-described features and/or combinations and subcombinations of several further features noted below in the detailed description. The above-listed and other objects, features and advantages of the present invention

will more be more fully set forth hereinafter.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention is further described in the detailed description which follows,

by reference to the noted plurality of drawings by way of non-limiting examples of preferred

embodiments of the present invention, in which like reference numerals represent similar

parts throughout the illustrations, and wherein:

Fig. IA is a block diagram illustration of a progressive sheet metal manufacturing

facility constructed according to an embodiment of the present invention;

Fig. IB is a block diagram illustration of a progressive sheet metal manufacturing

facility constructed according to another embodiment of the present invention;

Fig. 2 illustrates the respective data flow between the server module, database and

station modules, in accordance with an aspect of the present invention;

Fig. 3 is a flow chart of the general processes and operations that may be performed

by the server module, according to another aspect of the invention;

Fig.4 is a representative flow chart of the basic processes and operations that may

be performed by each of the station modules, in accordance with the teachings of the present invention;

Figs. 5 A and 5B are flowcharts that illustrate the logic flow of a similar part search

algorithm or process, according to an aspect of the present invention;

Figs. 6A, 6B, 6C, 6D, 6E, 6F and 6G illustrate, in accordance with an aspect of the

invention, a feature extraction operation for a four bend box with touched corners and for

a four bend box with open corners;

Figs. 7A, 7B and 7C illustrate, in accordance with another aspect of the present invention, a feature relation operation and process for identifying search keys for a part

having a four bend box, a bridge and another four bend box; Fig. 8 is a flow chart that illustrates the logic flow of the processes and operations

that may be performed to develop a 3-D model from a 2-D, single view drawing using a

folding algorithm;

Figs. 9A, 9B, 9C, 9D and 9E illustrate examples of an auto-trimming function and

cleanup function that may be performed to prepare a drawing for a face detection process;

Figs. 10A, 10B, IOC, 10D, 10E, 10F, 10G, and 10H illustrate the various processes

and operations that may be performed in a face detection process, in accordance with an

aspect of the present invention;

Figs. 1 IA and 1 IB illustrate the development of a final bend graph data structure

from the execution of a face detection process and bend line detection operation, according

to an aspect of the present invention;

Fig. 12 is a flow chart of the basic logic flow for developing a 2-D model based on

an original 3-D drawing (with no thickness) using an unfolding algorithm and other

processes, according to the teachings of the invention;

Fig. 13 is a flow chart of the basic logic flow for developing a 3-D model based on

an original 2-D, three view drawing using a 2-D clean-up operation, in accordance with an

aspect of the present invention;

Fig. 14A is a flow chart, according to an aspect of the invention, of the basic logic

flow of the processes and operations for performing a 2-D clean-up operation on a 2-D, three

view drawing;

Figs. 14B and 14C illustrate views and aspects of an exemplary 2-D, three view

drawing that may be processed by the 2-D clean-up operation of the present invention;

Fig. 14D illustrates a rotated view feature of the 2-D clean-up operation of the

present invention; Figs. 14E illustrates, in accordance with an aspect of the present invention, a

canonical form relating to the 2-D clean-up operation of the present invention;

Figs. 15 A and 15B illustrate an example of a 2-D, three view drawing with thickness

and a simplified 2-D, three view drawing model with no thickness that may be developed

using an eliminate thickness procedure, according to the teachings of the present invention;

Fig. 15C is an illustration of a cross thickness line and thickness arc of an exemplary

part, according to an aspect of the invention;

Fig. 16 is a flow chart of the logic flow of the various processes and operations that

may be implemented to develop a 3-D model with no thickness from a 3-D drawing with

thickness, in accordance with an aspect of the present invention;

Fig. 17 illustrates an exemplary data structure and access algorithm of the bend

model that may be utilized when implementing the present invention through, for example,

object oriented programming techniques;

Fig. 18 illustrates a block diagram of the structure of the bend model viewer, in

accordance with another aspect of the present invention;

Fig. 19 illustrates an exemplary solid view window display that may be provided as

output to a display screen;

Fig. 20 illustrates an exemplary wire frame view window display that may be

provided as output to a display screen;

Fig. 21 illustrates a 2-D flat screen image window display that may be provided as

output to a display screen;

Fig. 22 illustrates an orthographic view screen image that may be provided as output to a display screen;

Fig. 23 illustrates an example of the various dimension items that may be displayed in an automatic dimension mode of the present invention;

Figs. 24A, 24B and 24C illustrate a manner in which the flange length may be

defined for various different parts, according to an aspect of the invention;

Figs. 25A and 25B illustrate, in accordance with another aspect of the present

invention, adding an auxiliary flange length for two different types of parts;

Figs. 26A, 26B and 26C illustrate a manner in which the flange length may be

indicated for various parts that are displayed with thickness, in accordance with yet another

aspect of the invention;

Figs. 27A and 27B illustrate manners in which the flange length of parts with acute bend angles may displayed, in accordance with a tangent dimension method and an

intersection dimension method of the invention;

Fig. 28 is a flow chart of the logic flow of the processes and operations that may be

performed to develop a bending plan through the use of a graphical user interface, in

accordance with another aspect of the present invention;

Fig. 29A illustrates an example of a bend sequence input screen image that may be

displayed to a bending operator for developing a bending sequence;

Figs. 29B and 29C illustrates examples of selection a bend sequence and modifying

the insertion direction, in accordance with another aspect of the present invention;

Figs.29D and 29E illustrate further examples of a bend sequence input screen image

and a related screen display;

Fig. 30 illustrates, in accordance with an aspect of the present invention, a drag and

drop editing feature that may be provided to facilitate a bending operator in modifying and

editing a proposed bend sequence;

Fig. 31 illustrates an example of the various display menus and data tables that may be graphically displayed to aid a bending operator in selecting tooling;

Fig. 32 illustrates an exemplary tool set-up window that may be displayed to a

bending operator to facilitate the set-up of tooling in a proposed bending plan;

Fig. 33 A illustrates an example of a 3-D solid view window display with audio and

visual information attached through the use of pasted icons;

Fig. 33B illustrates another example of a display window that may be incoφorated

with icons for retrieving stored audio and video information, in accordance with an aspect

of the invention;

Fig. 34 illustrates an example of an image editing window that may be implemented

in accordance with the teachings of the present invention;

Figs. 35 A and 35B illustrate examples of a collision check function of the present

invention that may be implemented through a graphical user interface;

Figs.36A and 36B illustratea manipulation system of the invention for manipulating

the rotation and display of 3-D geometrical shapes by using, for example, a joystick;

Fig. 37 illustrates a manipulation system of the invention for manipulating the

zooming and display of 3-D geometrical shapes by using, for example, a joystick and zoom button;

Fig. 38 illustrates a manipulation system of the invention for manipulating the

panning and display of 3-D geometrical shapes by using, for example, a joystick and pan button;

Fig. 39 is an exemplary flow chart of the processes and operations that may be

performed in order to implement the 3-D navigation and manipulation system of the present

invention;

Fig.40 illustrates an example of mapping joystick movements to cursor movements, in accordance with an aspect of the invention;

Fig. 41 is an exemplary flow chart of the processes and operations that may be

performed to dynamically calculate the rotation axis of the rendered part;

Fig.42 illustrates an example of a main menu window display that may be provided

and displayed at, for example, a station module;

Fig. 43 illustrates an exemplary part information window display that may be

provided to permit a user to enter and modify part information;

Fig. 44 illustrates an exemplary bendline information window display that may be

provided to permit a user to enter and modify bendline information;

Fig. 45 illustrates an exemplary bend sequence window display of the present invention for viewing the intermediate bend stages of a sheet metal part;

Fig.46 illustrates an exemplary bend simulation window display of the invention for

simulating the intermediate bend stages of a sheet metal part;

Fig. 47 is an exemplary menu screen diagram and structure of the present invention

that may be provided and displayed to users for 2-D to 3-D conversions; and

Fig. 48 is an exemplary menu screen diagram and structure for a 2-D clean-up

operation of the present invention.

Fig. 49A illustrates an example of a 3-D representation of a part before one sided

open lines are removed, and Fig. 49B illustrates the part after the one sided open lines have

been removed from the 3-D representation, according to a 3-D clean-up process of the

invention that may be used when developing a 3-D model of a part from a 2-D, three view

drawing of the part;

Fig. 50A illustrates an exemplary 3-D representation of a part before the bendlines

have been identified, and Fig. 50B illustrates the part after the mold lines have been added, according to a 3-D clean-up process of the invention; and

Fig. 51 A illustrates an exemplary section of a part before cleaning the bendlines and

trimming the faces, and Fig. 5 IB shows the section of the part after cleaning and trimming

has been performed, according to a 3-D clean-up process of the invention.

BRIEF DESCRIPTION OF THE APPENDICES

In order to further facilitate the detailed description of the present invention,

reference is made to the noted plurality of appendices by way of non-limiting examples of

preferred embodiments of the present invention, in which sample source code and comments

are provided with respect to the various features, operations and functions of the invention,

and wherein:

Appendix A is an exemplary source code for executing a feature extraction operation

of the present invention when performing, for example, a similar part search;

Appendix B is an exemplary source code for effectuatinga similarity index operation

when performing, for example, a similar parts search of the invention;

Appendix C is an exemplary source code for performing a bendline detection

operation of the invention;

Appendix D is an exemplary source code for implementing a 2-D cleanup operation

of the present invention, which may be utilized when developing a 3-D model of a sheet

metal part based on an original 2-D, three view drawing;

Appendix E is an exemplary source code for implementing the various view modes

and functions of a bend model viewer of the present invention:

Appendices F, G, H and I are exemplary source code and comments relating to

executing and performing an auto dimensioning feature of the present invention;

Appendix J is an exemplary source code for implementing a part and entity visibility function of the bend model viewer of the invention;

Appendix K includes general comments relating to the implementation of the bend

model and the organization of the part structure, according to the various teachings of the

present invention; and

Appendix L includes exemplary source code for implementing a 3-D manipulation

and navigation system with dynamic calculation of the rotation axis of the rendered part.

DETAILED DESCRIPTION OF THE INVENTION

According to an aspect of the present invention, an apparatus and method are

provided for managing and distributing design and manufacturing information throughout

a factory, and for facilitating the production of components within the factory. The features of the present invention may be used in a wide variety of factory environments and settings

and, more particularly, the invention may be implemented in factory environments wherein

a series of production and manufacturing stages are effectuated at different locations. By

way of non-limiting embodiments and examples, the present invention will now be

described with reference to the production of bent sheet metal components at, for example,

a progressive sheet metal manufacturing facility.

Referring to Fig. 1 A, a progressive sheet metal manufacturing facility 38 is generally

illustrated in block diagram form, according to an embodiment of the present invention. As

shown in Fig. IA, the sheet metal manufacturing facility or factory 38 may include a

plurality of locations 10, 12, 14...20 that are dispersed throughout the factory. These

locations may comprise a design office 10, an assembly station 12, a shipping station 14, a

punching station 16, a bending station 18, and a welding station 20. Although the sheet

metal factory 38 in Fig. IA is depicted as having only six discrete locations, the factory may

of course include more than six discrete locations and may also include more than one location for each type of office or station illustrated in Fig. 1 A. For example, depending on

the size of and production capacity requirements for the facility 38, more than one punching

station 16, bending station 18, and/or welding station 20 may be provided. In addition, the

factory 38 may include more than one design office 10, assembly station 12 or shipping

station 14, and may also include other types of locations for facilitating the production and

manufacturing of components, such as bent sheet metal components.

Each of the locations 10, 12, 14...20 within the factory 38 may be adapted and

include equipment to execute one or more of the discrete production and manufacturing

stages or processes associated with the production and manufacturing of the components.

For example, the design office 10 may include an appropriate CAD/CAM system, to

facilitate the development of the sheet metal part design based on a customer's specification

The CAD/CAM system may comprise one or more personal computers, a display unit, a printer, and commercially available CAD/CAM software. By way of a non-limiting

example, the CAD/CAM system of the design office 10 may include AUTOCAD or

CADKE Y, or an Amada AP40 or AP60 CAD/CAM system available from Amada America,

Inc. (previously operating under the coφorate name of U.S. Amada Ltd.), Buena Park, California. In addition, other commercially available CAD systems may be used, such as

VELLUM, which is a Windows based CAD system available from Ashlar Incoφorated.

With the CAD/CAM software, the design programmer may develop a 2-D model and/or 3-D

model of the sheet metal part based on the drawings and data provided in the customer's

order. The design programmer may also generate control code based on the sheet metal part

design, in order to generate a part program for controlling, for example, CNC punch presses

and/or cutting machinery to punch or cut the sheet metal component from stock material.

Punching station 16 and bending station 18 may, each be provided with any combination of CNC and/or NC based machine tools. For example, punching station 16

may include one or more CNC and/or NC punch presses, such as COMA series and/or

PEGA series Amada turret punch presses or other commercially available CNC and/or NC

punch presses, and bending station 18 may include one or more CNC and/or NC press

brakes, such as RG series Amada press brakes or other commercially available multiple-axis,

gauging press brakes. Further, welding station 20 may be provided with appropriate welding

machinery in order to effectuate any required welding to the sheet metal component.

Punching station 16, bending station 18 and welding station 20 may be located at various

areas on the factory floor of the facility 38 and include machinery that is manually operated

by skilled operators (e.g., punch press operators, bending operators, etc.). Fully automated or robot assisted machinery, such as the Amada CELLROBO MINI and the Amada

PROMECAM, may also be provided at these locations. The required punching and bending

operations, and any necessary welding operations, may be performed at these stations during

the production process. As further shown in Fig. 1 A, the progressive sheet metal facility 38 may also include

assembly station 12 and shipping station 14. Assembly station 12 and shipping station 14

may include the necessary packaging, routing and/or transportation equipment to facilitate

the assembly and shipping of the manufactured components to the customer. The assembly

and shipping of the components may be performed or controlled manually by factory

personnel and also may be machine automated and/or machine assisted. In addition,

assembly station 12 and shipping station 14 may be physically located near the factory floor

(e.g., in close proximity to punching station 16, bending station 18 and/or welding station

20) or within a separate facility or area of the sheet metal factory 38.

In accordance with an aspect of the present invention, the management and distribution of critical design and manufacturing information is achieved by electronically

storing and distributing the design and manufacturing information. By replacing or at least

supplementing the traditional paper job set-up or work sheet with an electronic job sheet that

can be accessed instantaneously from any location in the factory, the present invention

improves the overall efficiency of the factory. In addition, through the various aspects and

features of the invention, the organization and accessibility of stored design and

manufacturing information is improved. Further, the ability to access and retrieve previous

job information relating to similar or identical sheet metal parts is enabled through the

various features of the invention. To this end, the various aspects of the present invention may be implemented and

effectuated by providing a communications network 26 that interconnects a server module

32 and a database 30 to each of the plurality of locations 10. 12, 14...20 within the sheet

metal facility 38. As further discussed below, each of the locations 10, 12, 12...20 may

include station modules that interface with communications network 26 and database 30.

Figs. IA, IB and 2 illustrate non-limiting examples of these features and implementation

of the invention.

As shown in Figs. IA and IB, communications network 26 may interconnect each

of the various locations 10, 12, 14...20 of the facility 38 with server module 32 and database

30. Communications network 26 may comprise any network capable of transmitting data

and information to and from the locations 10, 12, 14...20 and the server module 32 and

database 30. Such transmission may be achieved electronically, optically, by RF

transmission or by infrared transmission. By way of non-limiting example, communications

network 26 may be implemented by a Local Area Network (LAN), Ethernet or an equivalent

network structure. As further discussed below, each of the locations 10, 12, 14...20 may also include station modules having network terminating equipment (such as a computer,

minicomputeror workstation) and/or peripheral devices (such as a display monitor or screen,

printers, CD-ROMs, and/or modems) to transmit and receive information over

communications network 26. The network terminating equipment and peripheral devices

may include hardware and appropriate software or programmed logic for interfacing with

communicationsnetwork 26 and for providing the various features and aspects of the present

invention, as more fully discussed below. If a computer is provided at the factory location,

the computer may be a stand-alone, personal computer or a general puφose computer that

is part of an interface device of the equipment or machinery provided at the location. For

example, the computer may be an IBM compatible personal computer or may be a computer that is part of an interface/control system of the machinery, such as an Amada AMNC

system.

Server module 32 and database 30 are also connected to communications network

26. Server module 32 may comprise network terminating equipment, such as a personal

computer, minicomputeror mainframe, with suitable hardware and software for interfacing

with communicationsnetwork 26. Server module 32 may also include software or firmware

for implementing the various features of the invention, such as those described in greater

detail hereinafter. Further, according to an aspect of the present invention, server module

32 may also include database 30 for storing the design and manufacturing information

associated with each customer's order. Database 30 may be implemented by any

commercial available database with sufficient memory capacity for storing the design and

manufacturing information of the factory's customers and storing other data, tables and/or

programs. For example, database 30 may comprise a SCSI memory disk with 4 GB or more

of available memory space. The design and manufacturing information that is stored in database 30 may be accessed and distributed to the various locations 10, 12, 14...20 within

the sheet metal facility 38 via communications network 26. Various data formats, such as

Structured Query Language (SQL), may be used for accessing and storing data to database

30. In addition, information that is stored in database 30 may be backed-up and stored on

a wide variety of storage medium, such as magnetic tape, optical disks or floppy disks.

Server module 32 and database 30 may be connected to communications network 26 at a

separate area or location within the factory 38 (see, e.g., Fig. 1 A), or at a location that is

within or in close proximity to one of the predefined stations (e.g., within design office 10).

Although the embodiment of Fig. 1 A depicts database 30 as being part of server module 32

and interfacing with communications network 26 via the server module, database 30 may of course be physically located separately from server module 32 and connected to

communicationsnetwork 26 via a network database module 34, such as that shown in Fig.

IB.

By way of a non-limiting example, and in accordance with a preferred embodiment

of the present invention, server module 32 and each of the locations 10, 12, 14...20 may

comprise a personal computer, such as an IBM compatible computer with a 100-200 MHz

central processor unit (CPU), including a Pentium or an equivalent microprocessor, at least

32 MB of memory and a high resolution display screen, such as any commercially available

SVGA monitor with 800 x 600 resolution. Server module 32 and locations 10, 12, 14,...20

may also include a joystick or mouse device and a Sound Blaster or compatible sound and

game port adapter card for interfacing and controlling the display of information. Operating

system software may also be provided to support communications. For example, server

module 32 may be provided with Microsoft Windows New Technology (NT) or Windows

95 operating system software (both of which are available from Microsoft Coφoration, Redmond, WA), and each of the locations 10, 12, 14...20 may include Microsoft Windows

95 operating system software. In addition, server module 32 and locations 10, 12, 14...20

may be adapted to support multiple languages (such as English, Japanese, etc.) and full

support for an Object Linking and Embedding (OLE) server, such as an OLE2 server, may be provided.

Various database languages and management systems may also be used for creating,

maintaining and viewing information stored in database 30. A database language such as

Structured Query Language (SQL) may be used for defining, manipulating and controlling

data in database 30. For example, SQL Server (which is a retail product available from

Microsoft Coφoration) may be utilized to implement the present invention. In addition, the

invention may be provided with an Open Database Connectivity (ODBC) compatible driver

to facilitateaccessofinformationfrom database 30 over communicationsnetwork 26. More

information concerning OBDC may be found, for example, in the Microsoft Open Database

Connectivity Software Development Kit Programmers Reference manual.

Fig. 1 B illustrates, in block diagram form, a progressive sheet metal manufacturing

facility constructed according to another embodiment of the present invention. In the

embodiment of Fig. IB. the database 30 and server module 32 are provided separately, with

the database 30 being connected to communications network 26 via a network database

module 34. As discussed above, the present invention is not limited to this arrangement and

the database 30 and server module 32 may be provided together (as shown, e.g., in Fig. 1 A),

with the functionality of the network database module 34 for providing access to the

database being incoφorated in the server module. The embodiment of Fig. IB also

illustrates an example of the station module 36 that may be provided at each of the various

locations 10, 12, 14...20 throughout the sheet metal manufacturing facility 38. For puφoses of illustration, an exemplary station module 36 that may be located at bending station 18 is

provided in Fig. IB. Although not depicted in the example of Fig. IB, similar station

modules 36 may also be provided at the other locations within the facility 38.

As shown in Fig. 1 B, each of the modules (i.e., server module 32, network database

module 34, and station module 36) may be connected to communications network 26 via a

network interface card or port 42. The network interface card 26 may be vendor specific and

be selected based on the type of communications network that is selected. Each of the

modules 32, 34 and 36 may also include network software or programmed logic for

interfacing with the communicationsnetwork 26. The communicationsnetwork 26 may be

an Ethernet with any of a number of commercially available cable types, such as 10 Base/T

(twisted pair), 10 Base/2 (coax), or 10 Base/5 (thick cable), with the cable type being

selected based on the size of facility 38 and the amount or length of the cable required.

In Fig. IB, server module 32 may comprise a personal computer 40 with display

monitor or CRT 44 and input/output devices 46. which may include a keyboard, mouse

and/or joystick. The network interface card 42 may be plugged into an available expansion

slot or port of the personal computer 40. In addition, personal computer 40 may comprise

an IBM compatible computer with 100-200 Mhz operating speed and a Pentium or Pentium

Pro microprocessor. Personal computer 40 may also include, for example, 32 MB or more

of available main memory and 1.2 GB or more of available random access memory (RAM).

Display 44 may include a high resolution display screen, such as any commercially available

SVGA monitor with, for example, 800 x 600 resolution. To support the various graphics

and information that may be displayed on display 44, personal computer 40 may also

include any commercially available graphics card such as a PCI graphics card. Further,

computer 40 may include a Sound Blaster or compatible sound and game port adapter card and input output devices 46 may include a keyboard, joystick and/or mouse device.

In order to implement the various features of the invention, server module 32 may

be configured with software and various software packages. For example, server module

32 may be provided with operating system software, such as Microsoft Windows NT

(workstation version) or Windows 95. Further, in order to provide the server module

specific functionality and features of the invention (see, e.g., Fig. 3 ), server module 32 may

include software or programmed logic implemented routines. As discussed in greater detail

below, these routines may be developed using a high level programming language, such as

C++, and object oriented programming techniques. Server module 32 may also include or

interface with CAD or CAD/CAM software, such as VELLUM or Amada AP40 or AP60

software, to enter and/or develop original 2-D and 3-D drawings based on a customer's

specifications. For this reason, server module may be located in the design office 10 of the

manufacturing facility 38. In order to access data from database 30, server module 32 may

also include an OBDC driver, such as Microsoft ODBC driver, and may use SQL as a

standard for accessing data. An OLE server, such as OLE2 server, may also be provided to

link the data.

In the embodiment of Fig. IB, database 30 is provided separate from server module

32 and is connected to communications network 26 via network database module 34. As

indicated above, database 30 may comprise a SCSI disk with appropriate memory space

(e.g., 1 -4 GB), which may be selected based on the size of the factory 38 and the amount of

part information to be stored in the database. Network database module 34 may include a

personal computer 40, such as an IBM compatible computer with a Pentium microprocessor,

and an expansion slot fitted with network interface card 42 for interfacing with

communications network 26. Database 30 may be connected to personal computer 40 via a data bus and personal computer 40 may include standard display and input/output devices

(not shown in Fig. IB), such as a display monitor or CRT and a keyboard.

In order to facilitate access to database 30 based on SQL, personal computer 40 of

network database module 34 may be configured with a commercially available SQL server,

such as a Microsoft SQL server or Oracle SQL server. An OLE server, such as OLE2

server, may also be provided to link the data. Personal computer 40 may also be configured

with various operating software, such as DOS and Microsoft Windows NT (server version).

The embodiment of Fig. IB also includes an exemplary implementation of one

station module 36. In this embodiment, the station module 36 is implemented at bending

station 18. As shown in Fig. 1 B, the station module 36 may include similar hardware to that of the server module 32. That is, each station module (e.g., at the other stations shown in

Fig. 1 A) may comprise a computer 48 with display monitor or CRT 44 and input output

devices 46, which may include a joystick or mouse. The network interface card 42 may be

plugged into an available expansion slot or port of the computer 40. As discussed above,

the computer of the station module 36 may be a stand-alone, personal computer or a general

puφose computer that is part of an interface device of the equipment or machinery provided

at the location. For example, computer 48 may comprise a free-standing, personal computer

such as an IBM compatible computer with 100-200 Mhz operating speed and a Pentium or

Pentium Pro microprocessor, or computer 48 may be a computer that is part of or built into

an interface/control system of the machinery, such as an Amada AMNC system. Computer

48 may also include, for example, 32 MB or more of available main memory and 1.2 GB

or more of available random access memory (RAM). Display 44 may include a high

resolution display screen, such as any commercially available SVGA monitor with, for

example, 800 x 600 resolution. To support the various graphics and information that may be displayed on display 44, computer 48 may also include any commercially available

graphics card such as a PCI graphics card. Further, computer 48 may include a Sound

Blaster or compatible sound and game port adapter and to support, for example, a joystick

or mouse of the input/output devices 46.

In order to implement the various features of the invention, station module 36 may

also be configured with software and various software packages. For example, station

module 36 may be provided with operating system software, such as Microsoft Windows

95 or Windows NT (workstation version). Further, in order to provide the station module

specific functionality and features of the invention (see, e.g., Fig. 4 ), station module 36 may

C++, and object oriented programming techniques. In order to access and link data, station

module 36 may also include an OBDC driver, such as Microsoft ODBC driver, and an OLE

server, such as OLE2 server. Similar to server module 32, station module may use SQL as

a standard for accessing data from database 30.

If the station module 36 of bending station 18 is provided as a free-standing personal

computer, then software may be provided to

create bending code data (i.e., NC data) and to interface with the machinery 25 (e.g., a CNC

or NC controlled press brake). In the embodiment of Fig. IB, computer 36 is illustrated as

being implemented as a personal computer and is configured with software to interface with

bending machinery 25 via a standard RS-232-C wire interface. This interface may be

provided to permit the station module 36 to communicate with and send or receive bending

code data to the bending machinery 25 via the RS-232-C interface. The implementation of the interface is vendor specific and will depend on the data format and machine instruction set used for the bending machinery 25. All data that is sent from the station module 36 to

the bending machinery 25 should thus be formatted based on the machine instruction set that

is defined for the machinery. The computer 48 of station module 36 may also be provided

with any commercially available CNC or NC software for generating bending code data, in

order to simulate the functionality that is normally provided by a built-in computer of CNC

or NC systems (such as a Amada AMNC) for such machinery.

Fig. 2 illustrates an exemplary embodiment of the respective data flows between

server module 32, database 30 and the various locations of the sheet metal manufacturing

facility 38. For puφoses of illustration, and to better facilitate the description of the

respective data flow in the embodiment, server module 32 and database 30 (integrated with

network database module 34) are each shown in Fig. 2 as being separately and directly

connected to communications network 26, with the data flow between these elements being

carried out across the communicationsnetwork. Of course, as will be appreciated by those

skilled in the art, a wide variety of data flow arrangements may be provided between these elements; and, if database 30 is arranged to be directly connected to server module 32, then

the data and information can be directly transferred from the server module to the database

without use of communications network 26. In addition, for puφoses of facilitating the

description herein, the illustration of communications network 26 in Fig. 2 has been

simplified and only punching station 16 and bending station 18 are shown in the drawing.

Nonetheless, it will be appreciated that the data flow to and from locations 10, 12, 14...20

(as well as any other location or area that may be present in the factory) may be carried out

in a similar manner to that described for punching station 16 and bending station 18.

The design and manufacturing information associated with each customer's order may be organized and stored in database 30. When a customer's order is initially received, basic product and design information may be entered at server module 32 and then

transferred and stored to database 30. As discussed above, server module 32 may include

any suitable means for entering the data, such as a personal computer with a keyboard, etc.

If a personal computer is utilized at server module 32, software may be provided to generate

menu driven screens to facilitate the entry of the data by factory personnel. The data entry

program may be, for example, a Microsoft Windows based application with help and/or

menu screens, etc. By way of a non-limiting example, the data that is entered and/or

developed at server module 32 and transferred to database 30 may include part information,

bend model data, feature extraction data, and bend line information, as generally illustrated

in Fig. 2.

The part information may comprise, for example, a part or order reference number,

the customer's name, a brief description of the part, the batch size or quantity, and scheduled delivery date. The bend model data may include, for example, part geometry and

manufacturing data, such as the overall dimensions of the part (e.g., width, height, depth),

and part material information such as the material type (e.g., steel, stainless steel, or

aluminum), thickness and tensile strength. Further, feature extraction data may be manually

entered and/or automatically generated to identify the key features of the part and to

facilitate similar part searches and other searches of the database. The feature extraction

data may be stored in a separate data file in database 30, or may be stored with the bend

model data and other job information for each part. The feature extraction data may

comprise, for example, features of the part such as the number of surfaces or faces, the

number or types of bends present (e.g., a positive bend between two faces or a negative bend

between two faces), the relationships between the faces and or the number of holes or other types of openings in the part. As discussed more fully below, such data may be represented and organized in a feature based part matrix and/or a sequence of search keys (see, e.g., Figs.

5-7 below). Lastly, bend line information may be entered at server module 32 for storage

in database 30. The bend line information may comprise, for example, pertinent bend line

information for each bend in the part, including the bend angle, the bend length, the inside

radius (IR) of the bend, the amount of deduction, and the bend direction (e.g., front or back).

In order to transmit to and receive data from database 30 over communications

network 26, each of the locations 10, 12, 14...20 may comprise a station module (such as

station module 36 described above) that is connected to the communications network. In

Fig. 2, punching station 16 and bending station 18 are generally illustrated in block diagram

form with a station module. As discussed above, the station module may comprise, for

example, software or control logic and a stand-alone personal computer or a general puφose

computer that is part of the equipment or machinery provided at the location. For each

customer's order, the design and manufacturing information (including the part information,

bend line information, and bend model data) may be accessed and retrieved by entering, for example, a predetermined reference number or code. The reference number or code may be

entered manually (e.g., by keyboard or digital input pad) or by scanning a bar code with a

bar code reader or scanner provided at the station module. In addition, in accordance with

an aspect of the present invention, previous job data may be accessed and retrieved from

database 30 from any location 10, 12, 14...20 within the factory 38 by performing a similar

part search. As discussed more fully in the detailed description that follows, a similar part

search may be conducted based on the feature extraction data or search keys stored in

database 30 so that previous job information relating to identical or similar part(s) can be

retrieved and utilized to reduce the overall manufacturing time of future jobs.

The design and manufacturing information that is retrieved from database 30 may be used by the shop floor operators to develop and test the bending plan. For example, a

bending operator at bending station 18 may access and retrieve the part information, bend

line information and bend model data from database 30 in order to determine the necessary

tooling and the optimum bend sequence for the sheet metal part. In accordance with an

aspect of the present invention, an ODBC driver may be provided to permit each station

module to interface database 30 and display information stored in the database. In addition,

server module 32 or the network database module of database 30 may comprise a SQL

server to facilitate the access and retrieval of data stored in the database. Once the bending code has been programmed based on the final bending plan, the bending code along with the

bend sequence and tool setup information may be sent from the station module of bending

station 18 to database 30 over communications network 30, as generally shown in Fig. 2.

This information may then be stored along with the other design and manufacturing

information associated with that job.

Other information may also be stored in database 30. For example, the 2-D and/or

3-D image representation of the part may be stored with the bend model data for the part.

The 2-D or 3-D image representation may be developed at design station 10 or another

location with a CAD/CAM system and transferred to database 30 via the station module of

the design station (or another appropriate location) and through the communications network

26. Alternatively, the 2-D or 3-D image may be developed at server module 32, by utilizing

or interfacing with an appropriate CAD/CAM system or modeling software and performing

a series of functions or operations, as will be discussed more fully below.

Referring now to Figs.3 and 4, a detailed description of the processes and operations

that may be programmed and performed by server module 32 and the station modules of each of the locations 10, 12, 14...20 will be provided. Figs. 3 and 4 are flow charts of the basic logic flow that may be performed by server module 32 and the station modules of each

ofthe locations 10, 12, 14...20 withinthe sheet metal manufacturingfacility 38. While Fig.

4 is directed to the processes and operations that would typically be performed at, for

example, bending station 18, it will be appreciated that other processes and steps may be

performed depending upon the operations performed at each particular location within the

facility 38. The processes and operations discussed below may be implemented by software

and by using any one of a wide variety of programming languages and techniques. For

example, in accordance with an aspect ofthe present invention, the processes and operations

described below with reference to the accompanying drawings may be implemented by

using a high level programming language such as C++ and using object oriented

programming techniques. Further, by way of a non-limiting example, VISUAL C++ may

be utilized, which is a version of the C++ programming language written by Microsoft

Coφoration for Windows based applications.

Fig. 3 is a flow chart of the basic processes and operations performed by server

module 32, in accordance with an aspect ofthe invention. Fig. 3 illustrates the basic logic

flow of the processes and operations performed by the software or programmed logic of

server module 32. Server module 32 may include a Windows based application with tool

bar icons and help and/or menu screens to assist an operator or user in selecting and

executing the various processes and operations ofthe server module. The process begins

at step S.1 , when a customer's order is received at the sheet metal manufacturing facility 38.

The customer's order will normally include the necessary product and design information

so that the component may be manufactured by the factory 38. This information may

include, for example, the geometric dimensions ofthe part, the material required for the part, and other design information. Based on the information received from the customer, server module 32 may perform a search of previous job information stored in database 30, as

illustrated in step S.3. The job information stored in database 30 may be searched based on

a wide variety of search criteria. For example, information may be searched based on a

predetermined reference or job number or a similar part search may be performed based on

certain design features ofthe part, so that previous job information relating to an identical

or similar part can be retrieved and utilized for the current job. A more detailed description

of a similar parts search that may be utilized is provided below with reference to Figs. 5-7.

At step S.5, the results of the search of the database are analyzed to determine

whether the current customer's order relates to a new part, a part that is similar to a previous

job, or a repeat of a previous job. If an identical match is found (e.g., the same part or reference number is located) and the present customer's order is a complete repeat of a

previous job performed at the factory, then no further modifications to the job information

is necessary and the previous job information may be accessed from database 30 and used

to carry out the present customer's order, as shown at step S.1 1. The search of the database

may provide the part or reference number and or file name ofthe previous job so that the job

information may be accessed from the database by an operator at the server module 32 or

any of the station modules. If only the part or reference number is provided, then a

translation table may be provided so that the file name ofthe previous job information may

be determined and accessed based on the entry of the part reference or job number by an

operator. Thus, an operator at, for example, server module 32 may access the job

information and the 2-D and 3-D modeling information from database 30 to analyze the

geometry ofthe part and confirm that it is similar to that ofthe repeat order. If the order is

confirmed to be a repeat order, then a bending operator located at the station module of bending station 18 may also access the previous job information and utilize the manufacturing information, including the bending code data and tool setup information, to

bend and produce the part. The use of such stored expert knowledge thus enables repeat

orders to be manufactured more efficiently and without the need to reproduce previously

entered and developed job information.

If, however, it is determined at step S.5 that the current customer's order is similar

to a previous job or the same as a previous job but requires modification of, for example, the

job or reference number or batch size, etc., then at step S.7 the previous job data located by

the search may be retrieved from database 30, and edited and modified by an operator at

server module 32. An editing function may be provided to allow editing and modification

of previous job data to create new job data that may be stored in database 30 for the present

customer's order. The amount of editing required will depend upon the amount of similarity

that exists between the previous job and the current job. The amount of editing may

encompass simply modifying the reference or job number or batch size, and/or may involve

more extensive modifications such as editing the dimensions of the part and the defined

bend sequence. After the previous job information has been edited, the revised job information may then be stored in database 30 at step S.9. The revised job information may

be stored under a new reference or job number. In addition, various database management

functions (such as copy, delete, save, rename, etc.) may be provided to permit the previous

job information to be maintained in database 30 or to permit the previous job information

to be erased or overwritten upon entry of a special command.

If it is determined that there is no similar or identical match to the current job and,

thus, that the present customer's order relates to a new job, then logic flow proceeds to step

S.15, as shown in Fig. 3. Since, in this case, the current job relates to a new job it will be

necessary to independently develop and enter the design and manufacturing information. Menu and/or help screens may be provided by the server module 32 to assist the operator

in entering all of the necessary job information. In accordance with an aspect of the

invention, an operator at server module 32 may create a new file by first entering the basic

part information for the new job. The part information may comprise, for example, a

reference or job number, the customer's name, a brief description ofthe part, the required

batch size or quantity for the job, and the scheduled delivery date. The feature extraction

data or search keys may also be entered at step S.15, or this data may be automatically

developed or extracted concurrently with the development of the bend model data, as

described below. Other data or information may also be entered at step S.15, or entered after

or during the entry of the bend model data, such as the bend line information which may

comprise, for example, the bend angle, radius and length for each bend line in the part.

After step S.15, logic flow proceeds so that the bend model data may be developed and

entered at server module 32 by an operator, as shown in Fig. 3.

The development and entry of the bend model data may depend upon the original

drawings and information provided from the customer. The customer's order may include,

for example, a 2-D, single view flat drawing of the part to be manufactured and or a 2-D,

three view (e.g., including top, front and side views) drawing ofthe part. Occasionally, the

customer may also provide a 3-D, wire frame drawing of the part, with or without the

thickness of the material ofthe part being indicated in the drawing. In accordance with an

aspect ofthe present invention, the bend model data may include both the unfolded (i.e., the

2-D flat representation) and the folded (i.e., the 3-D representation) information for the part

to be manufactured. Thus, if only a 2-D flat drawing is provided by the customer, it will be

necessary to develop a 3-D drawing ofthe part by applying, for example, a folding algorithm

or process to the 2-D drawing. Alternatively, if only a 3-D drawing ofthe part is provided, then it will be necessary to develop a 2-D flat drawing by applying, for example, an

unfolding algorithm or process to the 3-D drawing. In accordance with another aspect ofthe

present invention, the 2-D and 3-D models that are saved in the bend model may be

developed and represented without the sheet material thickness (i.e., with no thickness).

This is possible due to the unique symmetry of all sheet metal parts. Providing and

representing the 2-D and 3-D drawings with no thickness provides modeling and simulation

views of the part that can be more easily inteφreted and understood by the design

programmer, the bending operator and other users. Removing the thickness information also shortens and improves the processing time required by the server module and station

modules when performing and executing the various features of the invention described

herein. A more detailed description of such features, as well as the folding and unfolding

algorithms that may be utilized in the present invention, is provided below with reference

to the accompanying drawings.

Fig. 3 shows the general processes and operations performed when developing the

bend model data. The various types of drawings that may be received or developed based

on the customer's order and that may be entered to develop the bend model data are

generally shows at steps S.19, S.23, S.27 and S.31. A tool icon bar and menu and/or help

screens may be provided by the server module 32 to assist the operator in selecting and

executing each of these steps. The processing of these drawings to develop the 2-D and 3-D

models of the part for the bend model will depend on what type of drawings are initially

provided. These drawings may be manually entered or developed at server module 32, or

they may be downloaded from a tape or disk. Server module 32 may, for example, interface

with a CAD/CAM system located at, for example, design office 10, or server module 32 may include a stand alone CAD/CAM system. Further, the 2-D and 3-D drawings may be saved as DXF or IGES files and imported to server module 32.

If a 2-D, single view flat drawing is provided, then processing to develop the bend

model may begin at step S.19, as shown in Fig. 3. At step S.19, the 2-D, flat drawing that

was received or developed may be entered at server module 32. Other bend model data,

such the overall dimensions of the part (e.g., width, height, depth), and part material

information may also be enter at step S.19. Thereafter, a folding algorithm or process may be utilized to develop a 3-D model (with no material thickness) based on the original 2-D

single view drawing, as generally shown at step S.21. An example of the processes and

operations that may be performed to develop a 3-D model from a 2-D, flat drawing is

provided below with reference to Figs. 8-1 1.

If a 3-D, wire frame drawing (with no material thickness) of the part is received or

developed, the drawing information may be entered at step S.27. In addition, other bend

model data, such the overall dimensions of the part (e.g., width, height, depth), and part

material information may be entered at step S.27. Thereafter, an unfolding algorithm or

process may be executed at server module 32 in order to develop a 2-D model of the part,

as shown at step S.29. An example ofthe processes and operations that may be performed

to develop a 2-D model from a 3-D drawing (with no thickness) is provided below with

reference to, for example, Fig. 12.

The 2-D and 3-D model representations of the part may be stored as part ofthe bend

model for that part. In addition, as noted above, during the development and entry of the 2-

D and 3-D models, other bend model data may be entered (such as the part material

information and other manufacturing information) so that it may be stored with the bend

model data in database 30. The various features and data structure arrangements that may

be implemented for organizing and storing the bend model data are discussed more fully below (see, for example, Figs. 17 and 18).

As shown in Fig. 3, if a simple 3-D drawing (with no material thickness) of the

component is not originally developed or received, additional processing may be necessary

in order to develop a 3-D model of the part (with no thickness), before executing the

necessary unfolding algorithm or processes to develop the final 2-D model. Steps S.23,

S.25, S.31 and S.33 generally show the additional processing and operations that may be

performed by server module 32 before executing an unfolding algorithm and developing the

2-D model at step S.29.

For example, if a 2-D, three-view drawing of the part is originally provided or

developed, then at step S.23 the drawing may be entered at or imported to server module 32.

Further, other bend model data, such the overall dimensions ofthe part (e.g., width, height,

depth), and part material information may also be enter at step S.23. Thereafter, at step

S.25, a simple 3-D, flat drawing ofthe part may be developed based on the 2-D, three-view

drawing that was entered. The developed 3-D drawing may then be used to develop the 2-D

model at step S.29, as shown in Fig. 3. An example of the processes and operations that

may be performed to develop a 3-D model from a 2-D, three view drawing is provided

below with reference to, for example, Fig. 13.

If, however, a 3-D drawing with material thickness is originally received or

developed, then the drawing information may be entered at step S.31 for further processing

before applying the unfolding algorithm. Other bend model data, such the overall

dimensions ofthe part (e.g., width, height, depth), and part material information may also

be enter at step S.31. Thereafter, at step S.33, an eliminate thickness procedure may be

executed to eliminate the thickness in the 3-D drawing. In accordance with an aspect of the invention, server module 32 may prompt the operator or user to indicate the thickness in the drawing and to indicate which surfaces (e.g., the outside or inside) should be retained when

executing the eliminate thickness procedure. An example of an eliminate thickness

procedure that may be utilized in the present invention is provided below with reference to,

for example, Figs. 15 A and 15B. After the thickness in the 3-D drawing has been eliminated

at step S.33, logic flow will proceed to step S.29, where the revised 3-D model with no

thickness may be utilized and an appropriate unfolding algorithm or process may be applied

to develop the final 2-D model. An example of an unfolding process and the various

processes and operations that may be performed to develop a 2-D model from a 3-D drawing

is provided below with reference to, for example, Fig. 12.

As shown in Fig. 3, after all of the relevant information has been developed and

entered, the part information, bend model information and other data associated with the

customer's order may be transferred from server module 32 and stored in database 30 at step

S.35. The data stored in database 30 may include feature extraction or search data that may

be utilized when performing database searches. As described below, the feature extraction

or search data may include data that is indicative of the basic or key features of the part

associated with each job, so that searches ofthe database may be performed to locate job

information and stored expert knowledge relating to the same or similar parts. The data and

information entered at server module 32 may be sent directly to database 30 or transferred

over communications network 26, as shown, for example, in Fig. 2. As indicated above, a

more detailed description of the various processes and operations that may be performed for

the various drawings when developing the bend model data will be provided below with

reference to the accompanying drawings.

Fig.4 is a flow chart of the basic processes and operations performed by each of the

station modules that may be provided at the locations 10, 12, 14...20 of the sheet metal manufacturing facility 38. For purposes of illustration, Fig. 4 provides an example of basic

logic flow of the processes and operations that may be performed by a station module

located at, for example, bending station 18. As will be appreciated by those skilled in the

art based on the teachings ofthe present invention, the logic flow illustrated in Fig. 4 may

of course be modified for each station module depending upon the nature of the operations

and processes to be performed at each ofthe locations. Further, as with server module 32,

the processes and operations ofthe station module described below may be implemented by

software or programmed logic. In addition, the station module may include a Windows

based application with tool bar icons or help and/or menu screens to facilitate an operator

or user in selecting and executing the various processes and operations ofthe station module.

Such help and/or menu screens may also be provided to facilitate the entry or transfer of data

at the station module.

As shown in Fig.4, after initializing the station module at step S.51 , an operator may input one or more database search criteria or key terms at step S.53. The search criteria may

be entered to locate previous job information or job information relating to a new or current

job that is stored in database 30. The operator may input, for example, a predetermined

reference number or code in order to retrieve particular job information from database 30.

For example, in accordance with an aspect of the present invention, a bar code may be

provided on a routing sheet or may be affixed to the punched stock material and scanned by

a bar code reader at the station module to access the information. Alternatively, the

reference code or number could be entered manually through a keyboard or digital input pad

at the station module. A translation table may be provided so that the file name of the

previousjob information may be determined based on the entry ofthe part reference or job

number by an operator. In addition, it is contemplated that search criteria or keys may be entered to perform a similar part search for previously stored job information. Such a search

may be performed based upon the various design features or feature extraction data ofthe

part. A description of a similar part search that may be implemented, in accordance with an

aspect ofthe present invention, is provided below with reference to Figs. 5-7.

After the search criteria has been entered at step S.53, the station module may

execute a search of the database 30 at step S.55 via communicationsnetwork 26 and network

database module 34. The results ofthe search may then be sent back to the station module

and analyzed at step S.57 in order to determine whether the operator or user has requested

information relating to a new job or a similar previous job, or whether the request relates to

the complete repeat of a previous job.

If an identical match is found (e.g., the same part or reference number is located) and

it is determined that a previous job is to be repeated, then the stored design and

manufacturing information relating to the job may be transferred from database 30 to the

station module, where it may be displayed for viewing by the operator, as generally shown at step S.59. The station module may include one or more menu display screens or

directories to permit the operator to select and display the various information retrieved from

database 30. The operator may review the displayed information and run various

simulations, such as a 3-D bending simulation at step S.61 , to view the various stages in the

bending sequence and to understand the geometry of the part for that j ob . The operator may

also review other information such as the required tooling and any other special instructions

or messages that may have been recorded with the job information. After confirming the job information, the operator can then set-up the bending or other required machinery and

operate the machinery to produce the specified sheet metal components. The job information that is retrieved from database 30 may include the final bending plan data,

including the bending code to control the machinery at, for example, bending station 18.

The set-up and actual operation of the machinery may thus be carried out by the operator,

as generally shown at step S.63 in Fig. 4.

If no identical or similar job information is located and it is determined that the

information relates to a new job (i.e., only preliminary job information has been entered at

the server module 32 and complete job information has not yet been developed), then the

partial part information and bend model data may be downloaded from database 30 and sent to the station module where it may be viewed by the operator at step S.77. Since the

information requested relates to a new job, it will be necessary for the operator to develop

and enter a bending plan, including the required tooling and bending sequence. Thus, at step

S.79, with the information provided at the station module, the bending operator may develop

and define the bending sequence and tooling selection for the new job. As will be discussed

in greater detail below, a graphical user interface (GUI) and other features may be provided

at the station module to facilitate the bending operator in developing the bending plan. The

GUI may be provided to help the operator develop a bending plan by, for example,

displaying tooling options, automatically checking for potential collisions between the part

and tool(s), and simulating each of the intermediate steps in a proposed bend sequence.

After developing and entering the bending plan at the server module, the operator may

program the bending sequence at step S.80 to generate the bending code (i.e., the CNC or

NC code for executing the bend sequence with the bending machinery). The bending code

may be directly entered at the server module or imported to the server module by interfacing

with, for example, a CNC or NC controller of the bending machinery. Thereafter, the operator may set-up and test the bending plan at the bending work station at step S.81.

When all ofthe necessary testing and any necessary modifications to the bending plan have

been completed, the final bending data may be entered and saved to database 30 at step S.83.

The final bending data may include the bend sequence and tool set-up information, as well

as the bending program. This information may be sent from the station module of, for

example, bending station 18 to database 30 so that it may be saved with the other design and

manufacturing information associated with the new job.

If it is determined at step S.57 in Fig. 4 that the information relates to a similar part

of a previous job or the same part of a previous job but having, for example, a different

reference or job number or batch size, etc., then logic flow may proceed to step S.65. At step S.65, the previous job information may be retrieved from database 30 and displayed at

the bending station 18. The bending operator or user may then view the data to determine

which changes to the data will be necessary for the similar part. Once again, the station

module may include a series of menu display screens or directories to enable the operator

to select which information to display and the manner in which the information is to be

displayed or modified. For example, at step S.69, the station module may provide a 3-D

bending simulation based on the retrieved information in order to facilitate the operator's

development of a bending plan for the similar part. After reviewing the previous job

information, the operator may modify the tooling and bending information, as well as the

bending program, at step S.70. Other job information, such as the dimensions of the part,

the reference number or batch size, may also be modified and edited at step S.70.

Thereafter, at step S.71 , actual tooling set-up and testing may be performed by the operator

on the shop floor in order to test the modified bending plan. Upon completion of testing and any further modifications to the bending plan, the operator may enter at step S.73 the final

bending data and store the same in database 30 under a new reference or job number. As

noted above, the previous job information may be maintained in database 30 along with the

other stored job files. Further, various database management functions may be provided for

storing, deleting, renaming, etc. the files stored in the database.

Referring now to Figs. 5-7, a detailed description of an example of a similar part

search function that may be implemented in accordance with the teachings ofthe invention

will be provided. In accordance with an aspect ofthe present invention, a similar part search

procedure may be provided that utilizes a feature based topology similarity search algorithm

to search and retrieve previous job information from database 30. The similar part search

may involve a search for identical and or similar parts based the design features and/or

manufacturing information relating to the part to be produced. Further, the similar part

search may be implemented through use of software or programmed logic residing within,

for example, server module 32 and/or the various station modules throughout the factory 38.

The similar part search may be executed from server module 32 or any of the station

modules of the locations 10, 12, 14...20 within the sheet metal bending factory 38. A high

level programming language, such as C++ or Microsoft's VISUAL C++ programming

language, and object oriented programming techniques may be utilized to implement the

various processes and operations ofthe similar part search.

Figs. 5 A and 5B illustrate the logic flow of a similar part search algorithm or process

that may be utilized. As shown in Fig. 5A, the relevant part model data file may be

accessed at step S.100. The part model may comprise, for example, the bend model data developed at a CAD system located at design office 10 and/or the data developed and

entered at server module 32. The part model data may include, for example, part topology

data representing the orientation, geometric relationships and relative position ofthe various

surfaces or faces and bendlines ofthe part. After the part model data has been retrieved, or

after the bend model data for a part has been manually entered, a feature extraction operation

may be performed at step S.102 to automatically derive feature extraction data for that part

based on the bend model and/or part topology data ofthe part.

According to an aspect of the present invention, feature extraction data may be

derived automatically by analyzing the various features of the sheet metal part. For example, various surface or face features and bend features may be analyzed to determine

the similarities between various parts. For instance, the various faces of a part may be

analyzed to determine whether adjacent faces have open or touching comers. Other features

such as the existence of parallel bends, serial bends, collinear bends or opposite bends may

be analyzed to determine and extract the distinct and unique features of each part.

Table 1 indicates various bend and face features that may be analyzed when

performing a similar part search. The extraction features that should be included in the

feature extraction operation include the positive bend and negative bend features, as well as

the touch comer and open comer features. In addition, the feature extraction operation

should also include at least feature analysis of parallel bends, serial bends, collinear bends,

different phase, collinear bends and thickness offset bends. TABLE 1

Feature Brief Description

PosBend positive bend between two faces

NegBend negative bend between two faces

P90B end 90 degree positive bend angle

N90Beπd 90 degree negative bend angle

MrPosBend multiple positive bendlines between two faces

MrNegBend multiple negative bendlines between two faces

ZBend Z bend heMBend Hemming bend in positive direction hem Bend Hemming bend in negative direction

TouchCnr two faces touched in corners with same bend direction

OpenCnr two faces open in comers with same bend direction

PrllBend two parallel bendlines with same bend angle direction and opposite bendline direction

SerlBend two parallel bendlines with same bend angle direction and same bendline direction cLnrBend collinear bendlines with same bend angle direction on one face

D CInrBend collinear bendlines with same bend angle and on different faces tHkOffBend thickness offset bendlines with same bend angle direction and on two neighboring faces touchCnr two faces touched in corners with opposite bend direction openCnr two faces open in corners with opposite bend direction prllBend two parallel bendlines with opposite bend angle direction and opposite bendline direction serlBend two parallel bendlines with opposite bend angle direction and same bendline direction clnrBend collinear bendlines with opposite bend angle direction on one face thkOfϊBend thickness offset bendlines with opposite bend angle direction on two neighboring faces

NoRelation no relation between two faces

The feature extraction operation performed at step S 102 may include a series of

operations including analyzing the bend model data and topology for each feature,

modifying the topologies, and developing feature based matrices from the topologies for further analysis. For puφoses of illustration, Figs. 6A-6G show an example of a feature

extraction operation for a part consisting of a four bend box with touched comers and a

part consisting of a four bend box with open comers. For puφoses of illustration, Figs.

6A-6G show the feature extraction based on the comer relationshipof adjacent faces. For

a closed, four bend box, with five faces (1-5) such as that shown in Fig. 6A. and for an

open, four bend box, with five faces (1-5) such as that shown in Fig. 6B, the same simple

face topology, such as that shown in Fig. 6C, may be provided to represent either part. This topology may be stored in and provided with the part or bend model data. The

simple face topology of Fig. 6C, however, only provides basic information regarding the

relationshipof the faces (1 -5) of the part and does not provide any information as to the

various features ofthe part, such as the comer relationship between adjacent faces or the

type of bends that are included. Accordingly, during the feature extraction operation, by

analyzing the part or bend model data and the related face topology stored therewith, the

basic face topology may be modified to contain additional information with respect to the

various features of the part.

For instance, by examining the part or bend model data for the closed, four bend

box of Fig. 6 A, the comer relationship of adjacent faces may be analyzed and a modified

face topology, such as that indicated in Fig.6D, may be developed to indicate the touched

co er status ofthe respective faces. Similarly, by examining the part or bend model data

ofthe open, four bend box of Fig. 6D, a modified face topology, such as that shown in

Fig. 6E, may be developed to indicate the open comer relationship between the various

adjacent faces in the part. As shown in Figs. 6D and 6E, special connectors may be added to the face topology to indicate the relationship (e.g., touched or open) between the comers of the faces. Other data may be added to the face topology data structure to

indicate other features (e.g., the type of bends present, etc.) and to developed a featured

based, face topology. After modifying the topology to include the feature based

information, a matrix may be developed so that the extracted information may be more

easily analyzed and compared. For example, based on the feature based face topology

of Fig. 6D, a matrix such as that shown in Fig. 6F may be developed to indicate the

various features of the closed, four bend box of Fig. 6A. Similarly, for the open, four

bend box of Fig. 6B, a matrix such as that shown in Fig. 6G may be developed based on

the feature based face topology shown for example in Fig. 6E. Other feature extraction

data may also be indicated in the matrix, such as the bend features ofthe part (e.g., a 90°

positive bend angle or a 90° negative bend angle, etc.).

As noted above, the features extraction operation of step S.102 may be performed

by analyzing the bend model data and topologies to determine if various features are

present in the part. In accordance with an aspect of the present invention, the feature

extraction operation may be performed on the bend model and topology data provided

for the part. This data includes all ofthe critical geometric and location data (e.g., in 2-D

space (X,Y) and/or 3-D space (X,Y,Z)) relating to the sheet metal part, including face

data, bendline data (e.g., bendline length and location, etc.), face-bendline relationship

data, bend angle data, and special features data (e.g., data relating to special bending such

as Z-bends and Hemming, etc.). The lines, bendlines and other entities may be defined

by endpoints and or vectors. For example, each 2D line may be specified by a set of 2D

endpoints (e.g., XI , Yl and X2,Y2) and each 3D line may be defined by a set of 3D endpoints (e.g., X1,Y1 ,Z1 and X2,Y2,Z2). Bendlines may be represented by vectors, which indicate 2D or 3D space location as well as direction ofthe bendline. Further, 2D

arcs may be specified by 2D space data (e.g., CenterX, CenterY, Radius, Begin Angle,

End Angle) and 3D arcs may be defined by 3D space data (e.g., CenterX, CenterY,

CenterZ, View Matrix, Radius, Begin Angle, End Angle). Part topology data may also

be provided to indicate the location ofthe various faces and bendlines ofthe part, as well

as their geometric relationships to one another. Each face may be defined by a collection

or linked data list of lines and or arcs.

To extract features ofthe part, the feature extraction operation may be performed

on the bend model and topology data to analyze and determine whether certain features are present in the part. This process may include analyzing the bend model and topology

data for the part based on the various characteristics and relationships associated with

each ofthe features to be extracted. By analyzing the bend model and topology data for

the existence of the characteristics and relationships for each feature to be analyzed, the

presence of features (such as a touched co er or open comer feature between faces, or

a parallel or serial bends feature) may be detected. Different processes may be provided

to detect the particular characteristics and relationships of each feature in the feature

extraction operation. Based on the similarity of characteristics and relationships among

the features to be analyzed, processes may be combined or developed to check for the

existence of more than one feature in the part.

By way of a non-limiting example, a process that may be performed during the

feature extraction operation of step S.102 in order to extract and detect comer features,

such as a touch co er feature of two faces having the same bend direction (i.e., a TouchCnr feature in Table 1), will be provided. The process described below may also be applied to detect other features, such as a touch comer feature of two faces having

with opposite bend direction (i.e., a touchCnr feature in Table 1 ) or open comer features

between two faces having the same or opposite bend direction (i.e., a OpenCnr or

openCnr feature in Table 1 ). The process may also be modified to detect other features

(e.g., parallel bends, serial bends, etc.). In addition, the data relating to each possible

combination of faces may be analyzed for the characteristics and relationships of each of

the features to be extracted.

For example, for the touched comer feature TouchCnr, the basic characteristics

or relationship to be detected include: two faces with a common face; the same bendline

directions; non-parallel bendline directions; and bendlines with a common vertex (or

vertices with a distance therebetween that is within a predefined range). For the touched comer feature touchCnr, similar characteristics or relationships should be detected;

however, instead of the faces having bendlines that are in the same direction, the faces

should have bendlines that are in the opposite direction (see. e.g., Table 1 ). The open comer features OpenCnr and openCnr may be similarly detected, however, for each

feature the presence of an open comer between the faces (e.g., the bendlines ofthe faces

are spaced apart by a distance that is greater than a predefined range) instead of a touched

comer relationship, and the detection ofthe bendlines having the same bendline direction

or the opposite direction (see, e.g., Table 1 and the definitions provided therein for

OpenCnr and openCnr) should be analyzed.

To detect the touch comer feature (e.g., the TouchCnr feature in Table 1 ), the

bend model and topology data for any two faces may first be analyzed to determine if the two faces are attached to a common face. This may be detected by looking at the bendline data for each ofthe faces and the bendline-face relationship data for each of the

bendlines to determine if a common face exists. If the two faces are attached to a

common face, then the bendline direction of each of the faces may be analyzed to see if

they have the same bendline direction (or the opposite bendline direction if detecting, for

example, the touchCnr feature). This may be determined by analyzing, for example, the

vector data indicating the bendline direction for each of the faces.

If it is determined that the two faces have a common face and have the same

bendline direction based on the bend model and topology data, then the data may be

checked to detect if the bendlines are parallel. Various methods may be used to detect whether the bendlines are parallel based on the bend model and topology data. For

example, the detection of parallel bendlines may be determined by taking the cross-

product ofthe vectors defining the bendline directions. Ifthe cross-productof the vectors equals zero (or is approximately zero), then it may be determined that the bendlines are

parallel. If the cross-product of the vectors does not equal zero (or is not approximately

zero), then the bendlines of the two faces are not parallel.

After determining that the two faces have a common face, the same bendline

direction and the bendlines are not parallel, then

the bend model data may be analyzed to determine the comer relationship between the

faces (e.g., touched or open). The comer relationship ofthe two faces may be determined

by detecting from the bend model data whether the bendlines ofthe faces have a common

vertex. If the bendlines have a common vertex, then the two faces have a touched comer

relationship with the same bendline direction (e.g., TouchCnr feature in Table 1). If the

bendlines have a common vertex, but it was determined that the bendlines of the two faces do not have the same direction, then it may be determined that the two faces instead

have a touched comer relationship with opposite bendline direction (e.g., touchCnr

feature in Table 1 ).

If the bendlines ofthe two faces do not have a common vertex, then it still may

be determined that the two faces have a touched co er relationship if the distance

between the vertices is within a predefined range. Often, a minimum amount of space

will be provided between adjacent faces ofthe part to provide clearance for passage of,

for example, the punch tool. This spacing is usually defined by the width of the tool at

the height ofthe flange. By way of example, a touched comer feature may be determined

to be present if the spacing between the vertices of the bendlines of the two faces is within 0-5 mm. If the spacing between the comer of the two faces is greater than the

predefined range, then it may determined that a open comer feature is present (e.g., the

OpenCnr or openCnr feature of Table 1).

The above described process may be performed for every possible combination

of faces in the part, to determine the comer feature of each ofthe faces. Other features

relating to the faces and bendlines ofthe part may be conducted in a similar fashion by

analyzing the part geometry and topology data. An exemplary code for performing the

feature extraction operation of step S.102 is provided in Appendix A. The code was

written in C++ programming language and includes the various processes for extracting

and detecting features such as those noted in Table 1. Comments are provided in the

code of Appendix A to facilitate the analysis ofthe logic and algorithms used therein.

In addition, the terminology for the various features in Table 1 is maintained in the sample code to aid understanding ofthe same. After detecting the various features ofthe part, the basic topology ofthe part may

be modified to include the extracted features. While it may be useful to provide feature

based topologies, such topologies can not be easily compared with one another. Instead,

the inventors of the present application have discovered that it is more efficient and easier

to compare feature extraction information when provided in the form of matrices.

Therefore, according to one ofthe features of the present invention, a feature based part

matrix (such as the representative matrix shown in Figs. 6F and 6G) may be created

based on the features detected during the feature extraction operation. The feature based

matrix for the part may then be compared with other predefined and stored matrices to

determine what basic shapes or features are included in the part.

A feature based matrix may be created and stored for each part after detecting and

extracting the various features for the part. As shown in Figs. 6F and 6G, the matrix may

be a two dimensional matrix that is symmetric and that has an order that is equal to the

number of faces in the part. The matrix may contain all of the detected feature

information for the part, with the various features between each of the faces being

provided in each of the locations of the matrix. The feature based part matrix may be

temporari ly stored in the memory of the server or station module and only used and

compared with the predefined matrices during execution of the similar part search.

Alternatively, the feature based part matrix may be permanently stored with the other job

information in database 30 and accessed from any location within the factory.

Referring back to Fig. 5A, after the feature extraction operation has been

performed, the resultant feature extraction data matrix may be compared with predefined feature extraction data matrices provided in a feature topology library. The feature topology library may be stored as a separate datafile in a database, such as database 30,

or the memory of the server module or the station module. The feature library may

consist of predefined matrices with feature extraction data corresponding to or defining

basic or fundamental part shapes (e.g., a four bend box, a bridge, etc.). Each of the

predefined feature based matrices, as well as the feature based part matrix, may be stored

as ASCII or text files. The comparison at step S.104 may be conducted to determine the

basic or fundamental shapes/features that are present in the sheet metal part, as illustrated

at step S.106. A stored lookup table may be provided to indicate which fundamental

shape corresponds to each ofthe predefined feature matrices. When a match is located,

the lookup table may be accessed at step S.106 to determine which fundamental shapes

are present. The matched matrices from the predefined library may be ofthe same order

as the feature based part matrix (in which case the part is determined to exactly

correspond and include only one fundamental shape), or may be sub-matrices of the part

matrix (in which case the part may include more than one fundamental shape).

Recursive programming techniques may be utilized to compare the feature based

part matrix with the matrices in the predefined library. By interchanging the indexes of

the matrices when comparing the information therein, the use of data assignments may

be avoided and the amount of required processing time reduced. The use of recursive

programming techniques and interchanging of indexes also facilitates the comparison of

matrices that have different orders and different base faces.

According to an aspect ofthe present invention, the comparison operation that is

performed at step S.104 may consist of a series of comparisons and may initially begin

based on the comparisons of matrices relating to more complicated shapes (e.g., those - 64 - shapes containing multiple bends or complex forming such as tabs) and then proceed

through less complicated shapes (e.g., those shapes containing fewer bends or less

complex bends or number of faces). This series of comparisons may be performed until

a predetermined number of fundamental shapes are located in the part. For example, the

comparison operation may be performed to extract the three most complicated features

or shapes within any particular part. In addition, this operation may be performed by first

conducting the series of comparisons on groups of matrices that relate to shapes that are

more common or frequently found in sheet metal parts, and then proceeding to less

common shapes. Various methods for comparing the part with the predefined library

may be performed to provide useful results.

For example, the series of comparisons may be first applied to a right angle group

of matrices that include fundamental shapes that include right angle bends, such as

rectangular and square shapes with multiple right angle bends and simple parts with right

angle bends. This group of matrices may be searched based upon a series of comparisons

extending from more complex matrices within the group (e.g., a matrix corresponding

to a four bend box with tabs) to less complex matrices within the group (e.g., a matrix

relating to a simple hat part). The series of the comparisons may then be applied to a

polygonal part group of matrices and then a special features group of matrices. The

polygonal part group may include matrices defining parts having more than five sides and

at least one bend angle that is greater than 90 degrees. The special features group of

matrices may include matrices within the predefined library that relate to parts with

special features or forming, such as Z-bends or Hemming. Once again, the series of

comparisons between the feature based matrix of the part and the predefined matrices within each ofthe groups may be performed based on decreasing levels of complexity.

Thereafter, other groups of predefined matrices may be compared, such as a multiple

features group of matrices that includes parts that have two or more features on a single

face ofthe part.

By comparing the part with the matrices in the predefined library in order of

complexity, and by applying the series of comparisons to groups of matrices based on

frequency of appearance and use, a more effective and efficient comparison ofthe library

may be conducted to determine the fundamental shapes in the part. In addition, an

overlap of detected features is prevented and only the more complex shapes are identified.

At step S.108, a feature relation operation may be performed to determine the

relationship between the fundamental features or shapes located in the part. The relation

between the features or shapes may be defined in terms of distance. The distance

between any two shapes may be determined based on the number of bendlines or faces

between the base face of each of the shapes. Altematively, the relationship between

features may be defined in terms ofthe physical distance or real dimension between the

features, by geometrically analyzing the part and the relative position and distance

between the base face of each ofthe features.

Assume, for purposes of illustration, that the three most complicated features or

shapes determined at step S.106 for the part consist of a four bend box, a bridge, and

another four bend box, as shown in Fig. 7A. A feature relation operation performed on

such a part may be conducted to determine, for example, the number of bendlines

between the base surface or face of each fundamental feature. As shown in Fig. 7B, the feature relation between the base (1) of the first four bend box and the base (2) of the

bridge is a spacing of two bendlines. Further, the relation between the base (1) of first

four bend box and the base (3) ofthe second four bend box is a spacing of four bendlines,

and the relation between the base (2) of the bridge and the base (3) of the second four

bend box is a spacing of two bendlines.

Various processes may be provided for determining the number of bendlines

between the base faces of the fundamental shapes of the part. For example, a matrix

analysis of the feature based part matrix and the predefined shape matrices may be

utilized to determine the feature relation in step S.108. First, the corresponding base

faces of each ofthe fundamental shapes may be located in the part matrix. This may be performed by correlating the base face ofthe predefined shape matrix with the face index

in the part matrix. As discussed above, the predefined shape matrices isolated during the

comparison operation may be sub-matrices of the part matrix. In order to locate the

corresponding base face for each fundamental shape in the part matrix, the location ofthe

shape matrix within the part matrix and the correlation between the indices of the

matrices may be analyzed. With the base face of each of the fundamental shapes being

predefined and located within the first column of the shape matrix, the corresponding

location and base face within the part matrix may be located.

After determining the base faces of each of the fundamental shapes in the feature

based part matrix, the distance between the base faces of each shape may be analyzed to

determine the feature relationships. This analysis may include a search process to identify the distance between any two base faces. By looking at the feature and bendline

information in the part matrix, the number of bendlines between any two base faces may be determined. If more than one path is possible between two faces, the minimum

distance may be used to define the feature relation at step S.108.

After performing the feature relation operation, logic flow continues to step

S.110. As shown in Fig. 5B, an identification of database search keys may be performed

at step S.110 in order to determine the search keys to be utilized in the similar part search

ofthe database. The search keys may consist of any number of combination of features

and feature relations identified for the part. In addition, any hierarchy of criteria may be

used for assembling the search keys. By way of a non-limiting example, the search keys

may be developed by the following criteria: (i) the first and second most complicated

features or shapes identified in the part; (ii) the distance or feature relation between the

two most complicated features; (iii) the third most complicated feature or shape identified

in the part; and (iv) the feature relation or distance between the first most complicated

feature and the third most complicated feature, and the distance or feature relation

between the second most complicated feature and the third most complicated feature

identified in the part. Fig. 7C illustrates the search keys that may be developed based on

the example of Fig. 7A.

In order to simplify the search ofthe database, the search keys may be represented

by a string of integers, with predetermined codes being assigned to the various

fundamental shapes defined in the topology library. For example, assume that the integer

code " 16" was assigned to a four bend box, and that the integer code "32" was assigned

to a bridge. In such a case, the search keys of the example in Fig. 7C would be

represented by a string of integers comprising "16, 16, 4, 32, 2, 2", wherein "4" and "2"

represent the various distances between the fundamental shapes or features. The representation of the search keys, however, is not limited to integer strings and any

combination of integers and/or character strings may be used to represent the search keys.

The search keys for each part may be stored with the job information (as a

separate file or in the same file) in a database, e.g., database 30. The search keys, which

are representative of the feature extraction data, may be manually entered or

automatically developed, as described above. Additional feature extraction data, such as

the feature based part matrix, may also be stored with the search keys. If the search keys

are stored in a separate data file, a lookup table may be provided for locating the part

information associated with each set of search keys. Altematively, the search keys may

be saved with a data field identifying the part information (e.g., by part or reference

number).

At step S.l 12, a cooperative search of the database is performed based on the

identified search keys. The cooperative search is a search using a cooperative database

search technique. The cooperative search technique not only locates parts with identical

search keys, but also parts having similar search keys. This enables the identification of

similar and identical parts in the database. When a search is performed based on a

particular part, the identified search keys for that part may be compared with the other

search key data in the database. The cooperative search performed at step S.112 may be

adapted to identify those items in the database which exactly match or are most similar

to a particular part defined by the search keys, by relaxing or modifying the sequence of

search keys. Various processes and methods may be employed for adapting the search

keys during the cooperative search.

For example, an initial search ofthe database is performed to identify parts having the exact sequence of search keys as that identified for the part to be searched. This is

performed by comparing the identified search keys with the search keys stored in the

database. After identifying the parts (if any) with the same search keys, subsequent

searches of the database may be performed based on different modified search key

sequences to locate other similar parts. Initially, the items or criteria within the search keys that are less critical or sensitive (such as the feature relation or distances) may be

modified and searched before modifying the more critical or sensitive search items (such

as the fundamental features or shapes located within the part). In addition, each of these

items may be modified in terms of their importance, with more weight or importance

given to those items relating to the first and second most complicated features or shapes

located in the part. For instance, a first subsequent search may be performed after

modifying the defined distances between the third most complicated feature and the first

and second most complicated features. The distance may be modified by increasing the

distance by a predetermined number of bendlines (e.g., 1 -3) or defining a predetermined

range for the distance based on the current value for the distance. Thereafter, the distance

between the first and second most complicated features or shapes may be altered to

provide another set of modify search keys for searching the database. After modifying

the feature relation or distance search keys for the part, the identified shapes may be

altered to derive additional modified search keys in the cooperative search. For example,

the search key item relating to the third most complicated feature or shape may be

changed to a related but less complex shape depending on the current feature or shape

(e.g., from a four bend box with tabs to a simple four bend box). In addition, the search

keys for the first and second most complicated features may be similarly altered to provided further modified search keys for the cooperative search.

The manner in which the distance and feature/shape related to the search keys are

modified during the cooperative search may be executed according to various methods

and techniques. As described above, the amount by which to vary the distance may

depend on the current value ofthe distance. The distance amount (e.g., 4 bendlines) may

be modified to a distance range (e.g., 3-5 bendlines) to expand and make the search more

cooperative. For the features or shapes, modification of the search keys may also be

performed to identify similar parts. The features or shapes may be modified through a

hierarchical structure of feature types. For example, the current feature type (e.g., a four

bend box) may be modified to a less complex feature type (e.g., a three bend box) that

is related and within the same feature type. The hierarchical structure by which the

features/shapes are modified may be predetermined and developed based on different

methodologies, such as type abstraction hierarchy (TAH). More information on TAH

and TAH generation are provided, for example, in CHU et al.. Wesley W., Cooperative

Ouerv Answering via Type Abstraction Hierarchy. CSD-900032, Department of

Computer Science, University of California, Los Angeles, (October 1990) and CHIANG,

Kuorong, Automatic Generation of Type Abstraction Hierarchies for Cooperative Ouerv

Answering, a dissertation submitted as part of the requirements for a Degree of

Philosophy in Computer Science, University of California, Los Angeles, (1995), the

disclosures of which are expressly incoφorated herein by reference in their entireties.

Other processes and steps may be performed during the cooperative search. For example, in addition to searching the database based on the identified search keys relating

to the features of the part, searching may also be performed based on search criteria relating to the manufacturing information for the part. For instance, additional search

keys may be utilized to compare, for example, the machine set-up required for each part.

The machine set-up information may include the type of machine or machinery required

to produced the part, the tool(s) and tool set-up used to produce the part, and/or the

backgaging setting(s) ofthe machinery. The additional search keys may be developed

based on the machine set-up information and/or other manufacturing information and be

used along with the identified search keys when performing the cooperative search ofthe

present invention. As a result, the parts that are identical or similar to the part to be

produced may be identified based on both the design and manufacturing features ofthe part.

In order to select the most similar parts, a selected parts search may be executed

at step S.l 14 to perform a more detailed comparison ofthe results from the cooperative

search and to select a predetermined number of parts that are the same or most similar to

the part searched. The selected parts search may involve the analysis of additional

information or characteristicsof each ofthe parts identified from the cooperative search.

This may involve analyzing various features ofthe located parts, such as the dimensions

ofthe part or the types of holes or openings in the part, which are not provided from the

search key data. This may also involve comparing the manufacturing information

relating to each of the located parts, such as the machine set-up required for each part.

As noted above, the machine set-up information may include the type of machine or

machinery required to produced the part, the tool(s) and tool set-up used to produce the

part, and or the backgaging setting(s) ofthe machinery. In order to perform the selected

parts search, the bend model and other job information for each part may be accessed from the database based on the search keys identified during the cooperative search. As

noted above, a lookup table or additional data field may be provided to provide the job

reference number or code associated with each set of search keys. After retrieving the

part information from the database, the additional information concerning each part (e.g.,

part dimension, material type, special forming, part holes or openings, etc.) may be

analyzed to determine which parts are most similar to the part searched. This process is

optional and may act as an additional

screening process for selecting and grouping those parts from the database that are most

similar to the part. By analyzing and matching this additional information or

characteristics of the part, the selected parts search may be performed to identify or select a predetermined number or set of most similar parts. For example, the selected

parts search may identify the five most similar parts based on the number of matching

search keys and matching additional part characteristics. The number of parts to be

selected from the selected parts search is not limited to five, and may be selected based

on the needs of the factory and the number of the parts actually stored in the database.

This number may also be selectively modified to provide more effective and useful

search results, and the user may be given the opportunity to modify this number to vary

the search set.

After performing the selected parts search, a similarity index may be calculated

at step S.1 16 to rank the parts (in terms of similarity of features and number of matching

search keys) identified in the selected parts search. The similarity index may be

calculated and provided as output at the server or station module at step S.1 18, so that the

user may select which job files are to be retrieved from the database and provided for viewing. The similarity index may provide a ranking ofthe selected parts (e.g., a ranking

of 1 through 5 with the job or reference number for each part) based on the level of

similarity of features between the selected parts and that ofthe searched part. For this

puφose, the feature based matrix for each ofthe parts may be compared to that ofthe

searched part. Comparing the feature based matrices may provide a better indication of

the similarity between the selected parts and the searched part. As noted above, a feature

based part matrix may be stored along with search keys for each part. However,

permanently storing the feature based part matrix for each previous job along with the

search keys may unnecessarily take up a large amount of memory space (particularly

when a large number of parts are stored on the database). As such, it is possible to only

store the search key data for each of the parts and to automatically generate the feature

based matrix for each ofthe selected parts when a similar part search is performed.

Accordingly, after the bend model and other job information has been retrieved

for each of the selected parts, a feature based matrix may be developed through the

feature extraction operation of the invention, as described above with respect to step

S.102. The feature based matrix for the searched part, which may be temporarily stored

during a similar part search, may then be compared with each of the developed feature

based matrices of the selected parts. Various methods and processes may be utilized to

compare the feature based matrices ofthe parts and to determine the similarity between

the parts. For example, for each feature based matrix of the selected parts, the locations

within the matrix may be compared with those ofthe searched part. Each location within

the matrices may be compared based on recursive programming techniques. The information in the matrices may be compared by determining the location of corresponding base faces in each matrix and interchanging the indexes of the matrices.

Since the selected parts may correspond to or have shapes that are sub-features of the

searched part, and since the indexes ofthe matrices may not be identical or numbered in

the same way, it will be necessary to locate comparable faces in the part matrices and to

switch the indices when comparing the information therein. In addition, where more than

one sub-feature is located in a searched part, it may also be necessary to introduce one

or more pseudo faces (i.e., face columns and rows in the matrix with no or blank

information) in order to provide matrices of the same order when comparing the

information in the matrices. When comparing the information in the matrices, different ranking schemes may

be used in order to determine the level of similarity between each of the selected parts

and the searched part. For example, a penalty based ranking scheme may be used

wherein a predetermined penalty level or amount is assigned for each non-matching

location within the matrix. After comparing all of the information in the matrices, the

total penalty level for each selected part may then be used to determine the similarity

index. The selected part with the lowest penalty level may be determined to be the most

similar part to the searched part. The other selected parts may also be ranked based on

the total penalty level associated with each part (e.g., the lower the penalty level the

higher the similarity index). In accordance with another aspect of the present invention, the penalty levels for

each non-matching location may be assigned based on the type of information located

therein. The penalty level may be an integer amount and may be varied based on the

criticality or importance ofthe non-matching information. For example, a higher penalty level or amount may be assigned for non-matching locations relating to different and

unrelated feature groups (e.g., a parallel bend feature versus a serial bend feature). In

contrast, for non-matching locations relating to different but similar feature groups (e.g.,

a touch comer feature with same bendline direction versus a touch comer feature with

opposite bendline direction). The penalty levels or amounts may be predetermined and

categorized based on the type of information and the type of difference present for non-

matching locations.

An exemplary code for performing the similarity index operation of step S.l 16

is provided in Appendix B. The code was written in C++ programming language and

includes the various processes and operations described above with respect to comparing the matrices and assigning penalty levels for non-matching location. As noted above, the

resultant total penalty levels for each selected part that was compared may be used to

derive and display the similarity index. The code listing in Appendix B includes

comments to facilitate understanding ofthe logic and structure ofthe exemplary program

code therein.

Referring now to Figs. 8- 16, a more detailed description ofthe various processes

and operations that may be performed for developing the bend model data and developing

the 2-D and 3-D models of the part based on various 2-D and 3-D drawings will be

provided in accordance with an aspect ofthe present invention. As discussed above, the

bend model data associated with each sheet metal component includes data relating to

both the 2-D and 3-D representations of the part. Based on the type of original drawings

provided or developed based on the customer's order, various folding and unfolding algorithms and other processes may be utilized to develop the 2-D and 3-D models. In particular, Figs. 8-1 1 show an example of the logic flow of the folding algorithm and

other processes that may be utilized for developing a 3-D model based on an original 2-

D, single view drawing ofthe part. Further, Fig. 12 shows an example ofthe basic logic

flow of the unfolding algorithm and other processes that may be used for developing a

2-D model based on an original 3-D drawing (with no thickness). Lastly, Figs. 13-15 and

Fig. 16 show examples ofthe logic flow ofthe various processes and operations that may

be implemented to develop a 3-D model with no thickness from, respectively, a 2-D,

three view drawing and a 3-D drawing with thickness. The resultant 3-D model (with no

thickness) that is developed from these processes and operations may then be utilized to

develop a 2-D model based on an unfolding algorithm or process, such as that disclosed

herein.

Fig. 8 illustrates the logic flow ofthe processes and operations for developing a

3-D model from a 2-D, single view drawing using a folding algorithm. The functions and

operations performed in the flow chart of Fig. 8 may be implemented with software or

programmed logic residing in, for example, server module 32. At step S.120, the 2-D,

single view flat drawing that was provided or originally developed based on the

customer's specifications is entered or imported to the server module 32. The 2-D flat

drawing may be developed and entered into the server module 32 through the use of CAD

software, or may be imported to the server module by interfacing with an appropriate

CAD or CAD/CAM system, such as VELLUM or CADKEY. The 2-D drawing may be

stored, for example, as a DXF or IGES file and may illustrate the punched and/or cut

stock material that is to be bent. The 2-D flat drawing may also indicate the location of the bending lines and the location of holes or other openings in the surfaces or faces of the sheet metal part. In order to prepare the 2-D drawing for later processing, an auto-

trimming and cleanup function may be performed by server module 32 at step S.122,

before a succeeding face detection process is performed at step S.124 and a bendline

detection operation is executed at step S.I 26.

The auto-trimming and cleanup function ofthe present invention is provided in

order to prepare the 2-D flat drawing for processing. The 2-D flat drawing is a 2-D

representation of the sheet metal part in its unfolded state and includes part entities, such

as lines and arcs, that make up and represent the geometry ofthe part, as well as indicate

the location of any openings or holes in the part. Normally, the entities of such 2-D flat

drawings are entered and developed using a CAD or CAD/CAM system. However, when

constructing the 2-D flat drawing, such entities are often improperly connected or

overlapped, and a single entity may be used to indicate the boundaries of more than one

face. Further, outside lines defining the boundary ofthe part may be disconnected at their

adjacent comers, making it difficult to detect the out dimensions of the part and each

face. Further, the 2-D flat drawing may include extraneous information, such as

dimensional information and text. Such irregularities make it difficult to properly

analyze the original 2-D drawing and to uniformly detect the faces and bendlines ofthe

part. By providing the auto-trimming and cleanup operation of the present invention,

each ofthe faces may be represented by a unique set of connected entities. As a result,

the 2-D flat drawing may be more easily and efficiently analyzed for subsequent

processing and eventual folding in order to develop the 3-D model representation.

As shown in Fig. 9A, an original 2-D drawing may not provide trimming between faces and a single line entity in the drawing may define the outer boundary or boundaries of more than one face. As discussed above, such an arrangement makes it difficult to

detect each of the faces. The auto-trimming function of the present invention may be

provided to analyze the end points and intersection points of each of the part entities

(such as lines, arcs and bendlines), in order to determine connectivity information and to

break such entities at their intersection points. Such a trimming function may include

setting the endpoints for each ofthe broken entities to the determined intersection point.

For example, trimming of the intersection point illustrated in Fig. 9A would result in

three meeting entities (two lines and one bendline), each having a common endpoint at

the intersection point. By providing such a trimming function, the faces of the part may

be more easily detected based on entity analysis and connectivity. A more detailed

description of a face detection operation that may be implemented is provided below with

reference to Figs. 10A-10G.

Various processes and operations may be utilized to detect the intersection points

of the entities of the 2-D drawing. Such processes and operations may be developed

based on the format and arrangement of the data in the 2-D drawing file. Typically, a 2-

D flat drawing will include geometrical data (defining the various part entities, etc.) and

non-geometrical (e.g., text, etc.) data. It is possible to distinguish between the

geometrical data from the non-geometrical data based on the keywords provided for each

line or sequence of data. Such keywords are set in accordance with the data format ofthe

2-D drawing. Common formats for 2-D and 3-D drawings include DXF and IGES

formats. By analyzing the geometrical data for each of the entities, the end points and

intersection points for such entities may be detected and, where appropriate, trimming

may be performed. As discussed above, the lines, bendlines and other entities may be defined by

endpoints and/or vectors. For example, for a 2-D flat drawing, each 2-D line may be

specified by a set of 2-D endpoints (e.g., XI, Yl and X2,Y2) and bendlines may be

represented by vectors, which indicate 2-D space location as well as direction of the

bendline. Further.2-D arcs may be specified by 2-D space data (e.g., CenterX, CenterY,

Radius, Begin Angle, End Angle). The geometrical data may also include attributes to

distinguish between the various types of line entities (e.g., arc, solid line, dashed line,

dot-dashed line. etc.). Typically, arc entities are used to indicate holes and openings in

a sheet metal part, and solid lines are used to indicate the boundaries and shape of the

part. Bendlines are usually represented by dashed lines, and the centerline ofthe part is

represented by a dot-dashed line.

The geometrical data from the original 2-D flat drawing may be analyzed to

determine the intersection points between each entity. Various data analysis techniques,

such as data assignment or recursion, may be used to analyze the geometrical data for

each entity ofthe 2-D drawing. Based on the endpoints and/or other 2-D space data for

each entity, simple geometrical analysis may be applied to determine whether lines and

other entities intersect. If two entities are determined to intersect, then each entity may

be broken at the determined intersection point and the resultant entities may have their

endpoints assigned to a common point defined by the intersection point.

The manner in which trimming is performed may be based on the type of entities

that are detected to intersect. For example, if two solid line entities are detected to intersect, then each line entity may be broken to provide four line entities that meet at the determined intersection point, as shown in Fig. 9B. Further, if a line entity and arc entity are determined to intersect, such as that shown in Fig. 9C, then each entity may be broken

to provide two line entities and arc entities that have common endpoints. Detection ofthe

intersection of other entities, however, may not result in trimming. For example, if any

entity is determined to intersect with a centerline (e.g., a dot-dashed entity), then no

breaking of the entities is required, since the centerline of any part does not define or

distinguish the faces or bendlines of the part and trimming is not required. In addition,

unconnected entities may be broken if the open intersection or area is within a

predetermined tolerance. For instance, if the endpoint of a potentially intersecting line

is within a predetermined tolerance or distance e (e.g., 0.0 - 0.01 mm, or 0.0 - 0.001

inches) of actually intersecting another entity, then the entities may be treated as

connecting and intersecting at that projected point; and the entities may be broken, as

shown, for example, in Fig. 9D.

After auto-trimming has been performed, the resultant data may then be processed

by a cleanup function to detect and correct non-connected entities; however, the present

invention is not limited to such processing; and, in order to reduce processing time, the

cleanup function may performed simultaneously with the auto-trimming function while

each ofthe entities are being analyzed. During cleanup, the geometrical data ofthe 2-D

drawing is analyzed to detect open intersections or areas between adjacent entities. As

with the auto-trimming function, the endpoints and other 2-D space data of each entity

may be analyzed to detect an open intersection areas between the entities. Simple

geometrical analysis may be applied to such data to determine whether the endpoints of

the entities are within a predetermined tolerance or distance ε (e.g., 0.0 - 0.01 mm, or 0.0 - 0.001 inches) of one another. If the endpoints of the entities are determined to have such an open intersection, then the entities may be connected and assigned a common

endpoint, such as that shown in Fig. 9E.

Once again, the manner in which the cleanup function is performed may be

controlled based on the type of entities that are detected to have an open intersection. If

two solid lines are detected to have an open intersection, then the endpoints ofthe lines

may be assigned a common endpoint (see, for example, Fig. 9E). However, if any entity

is determined to have an open intersection with a centerline of the part (e.g., a dot-dashed

entity), then the entities should not be connected nor assigned a common endpoint and

the centerline entity should be ignored. In addition, the cleanup function may include

additional processes or operations for deleting the non-geometrical data (text, etc.) from

the 2-D drawing. As noted above, the non-geometrical data may be distinguished from

the geometrical data based on the keywords provided with the 2-D drawing data. The

cleanup function may also incoφorate other cleanup functions, such as those described

in greater detail below with reference to the 2-D cleanup function ofthe invention (see,

for example, Figs. 14A- 14E).

After the auto-trimming and cleanup functions are performed at step S.122, a face

detection procedure may be performed on the processed 2-D drawing, at step S.124. In

accordance with an aspect of the present invention, the face detection procedure may

include face detection logic for detecting and defining the faces of the part based on

entity (line and arc) and loop analysis. Figs. 10A-10H illustrate an example of the

various processes and operations that may be performed in the face detection procedure.

Loop detection techniques may be used in the present invention to determine and detect the faces ofthe part. The face detection procedure may be implemented through software or programmed logic residing at, for example, server module 32.

According to an aspect ofthe present invention, a loop detection analysis of the

outside boundary ofthe part followed by analysisof the minimum or inside loops ofthe

part may be utilized to detect each of the faces. Due to the unique geometry of sheet

metal parts, the faces and openings ofthe part may be detected based on the analysis of

the sequence ofthe relative maximal (e.g., outside) and minimal (e.g., inside) loops. As

discussed below, loop analysis may be performed based on the connectivity ofthe line

and arc entities ofthe part. By performing loop analysis from the outside ofthe part and

proceeding towards the center of the part, the openings and faces of the part may be detected based on the boundaries defined between the loops in accordance with a cyclic

sequence (e.g., face material, opening, face material, opening, etc.).

Assume that a 2-D flat drawing, such as that shown in Fig. 1 OA, is provided with various line entities for each face as shown in the drawing. As noted above, loop and

entity analysis may be performed by starting from the outside boundary ofthe part. Any

entity on the outside boundary of the part may be used as an initial reference point. By

way of non-limiting example, the left most side line entity may be detected and used as

an initial reference point, as shown in Fig. 10B. The left most side line entity may be

detected by comparing the geometrical data of each ofthe entities in the 2-D drawing and

determining which entity has the lowest X-coordinate value. After the left most line

entity has been detected, an outward appearance ofthe part may be derived from a point

PI to detect the outside boundary of the part, as shown in Fig. IOC. Either endpoint of

the left most line entity may be used to define point PI . In the illustrated embodiment

of Fig. IOC, the upper endpoint (i.e., the endpoint with the greatest Y-coordinate value) has been used as point PI .

Conventional loop analysis techniques may be used to derive the outward

appearance or loop about the part. For example, lead line vectors may be projected from

the initial point P 1 and the endpoints ofthe connecting entities as the outward appearance

ofthe part is followed. As each entity is detected and traversed, a flag may be provided

to indicate that the entity has been picked (e.g., a flag in memory may be set to 1 to

indicate that it has been selected once). The loop path may be initiated in either direction

from the initial point P 1. For example, the lead line vector may be projected in the

counterclockwise direction (e.g., by projecting the lead line vector in the Y-coordinate

direction) from the point PI . The loop is completed when the loop path returns to the

initiating point (i.e., point PI).

As noted above, from the initial point PI a lead line vector may be projected in

the counterclockwise direction (e.g., by initiating the first lead line vector in the Y-

coordinate direction). Thereafter, in order to detect the first entity in the path ofthe loop,

the angle that each unselected entity about point PI forms with the lead line vector is

measured and analyzed based on a coordinate frame, with the entity that forms the

smallest angle with the lead line vector being selected. For the outside loop, each angle

may be measured based on the outside angle that the entity line forms with the lead line

vector. The entities about point PI may be determined based on which entities have an

endpoint that is common to point PI. The unselected status of each entity may be

determined by analyzing the flag associated with each entity. As shown in Fig. 1 OC, two

entity lines (one extending in the X-coordinate and one extending in the Y-coordinate) are provided about P 1 in the exemplary 2-D drawing illustrated therein. When analyzing these entities, the line entity that extends in the Y-coordinate would be selected since it

forms a smaller angle (i.e., 0 degrees) with the lead line vector than the angle (i.e., 270

degrees) ofthe other line entity.

The loop analysis would then continue to the other endpoint ofthe selected line

entity, which would have a flag set to indicate that it has been selected. At that endpoint,

another lead line vector would be projected and the unselected entities about that point

would be compared to determine which forms the smaller angle with the lead line vector.

Once again, the angle should be measured from the outside of the lead line vector and a

coordinate frame may be used to determine the angle amount. If an arc entity is

encountered, then the angle should be measured from the outside of the lead line vector

to a line that is tangent to the arc. Further, if only one entity is about the next endpoint

(such as at the comer positions of the part), then no comparison is necessary and that

entity is simply selected and included in the loop.

As the loop path proceeds about the outward appearance ofthe part, each selected

entity may be included in a linked list to indicate the connectivity of the entities within

the loop. When the path returns to the initial point P 1 , the cycle is complete and the loop

may be defined (L4) based on the outward appearance and the linked list of entities or

lines that indicate the outside boundary of the part. Each of the lines or entities within

the loop L4 may be connected at their end points. The direction ofthe loop L4 may be

tumed in the opposite direction (i.e., clockwise), as shown in Fig. 10D. in order to

indicate that it is an outside loop. The direction ofthe loop may be defined based on the

order in which the lines are linked in the loop L4; and, thus, the direction may be changed

by reversing the order of the link list. After the outward loop has been completed, inside loop analysis ofthe part may

be performed based on a similar process to that used for the outside loop analysis. In the

inside loop analysis, however, each ofthe unselected entities are compared based on the

inside angle that each entity forms with the lead line vector. Further, during the inside

loop analysis, where both entities about a point are indicated as being selected (e.g., when

comparing two outside line entities which border a face), the two entities may still be compared unless they have been selected twice (i.e., a flag setting of 2). When there is

an entity that has been selected at least once (e.g., an outside entity) and an unselected

entity (e.g., an inside entity), no comparison may be performed and the unselected entity

may be selected as part of the loop. Figs. 10E-10G illustrated exemplary inside loops

that may be performed to detect and define the faces ofthe part shown in Fig. 10A.

The inside loop analysis may begin at any ofthe outside entity endpoints or by

detecting an entity that has not been selected. For example, point P 1 may be selected to

initiate the inside loop analysis and may be used to project the lead line vector;

altematively, one ofthe inside line entities that was not selected during the outside loop

analysis may also be used as an initial point for analysis. As with the outside loop

analysis, the lead line vector may be extended in the counterclockwise direction (e.g., by

initiating the first lead line vector in the Y-coordinate direction). Each entity about point

P 1 is then compared to determine which entity forms the smallest angle with the lead line

vector. A coordinate frame may be used to determine the angle formed with the lead line

vector. As noted above, during inside loop analysis the entities are compared based on

the inside angle that each entity forms with the lead line vector instead of with the outside

angle. After the initial entity has been selected and included in the linked list for the loop, its flag may be increment by one and further analysis may be performed by

projecting the next lead line vector. The process continues until the loop returns to the

initial starting point, at which point the first inside loop is defined (e.g., Ll) by its

associated linked list of entities.

Further inside loop analysis may performed in a similar fashion by proceeding

inwardly of the part. Subsequent starting points may selected by determining which

entities have been only selected once. Entities with flags that have been selected twice

will indicate that it is an outside entity that has already been selected for the outside loop

(e.g., L4) and for at least one of the inner loops (e.g., L 1 ). Once again, as each entity is

selected, its associated flag may be incremented by one to indicate that it has been

included in the inside loop link list.

After all ofthe inside loops have been defined (e.g., after all of the entities have

been selected twice in the example of Fig. 10G), the resultant loops may be used to

construct a loop tree. Fig. 10H illustrates an exemplary loop tree that may be defined

based the detected loops L1-L4. The outside loop (L4) ofthe part may be defined as the

root ofthe tree, with each inside loop (L1-L3) that has a common entity with the outside

loop being defined as children of the root. The presence of common entities may be

detected based on analyzing and comparing the linked list of entities that define each

loop. If additional entities (e.g., holes or openings) are detected within the inside loops,

then these loops may be defined as children ofthe inside loops (i.e., grandchildren ofthe

root of the loop tree) within which they are located.

After the face detection procedure has been performed at step S.124, a bendline detection operation may be performed at step S.126. As shown for example in Fig. 11 A, when detecting and analyzing the loops of a part at step S. I 24. the face detection logic

of the invention may utilize the loop tree to define the face information and store the

detected faces as nodes in a bend graph data stmcture. The faces of the part may be

detected from the sequence ofthe outside and inside loops in the loop tree. As indicated

above, each ofthe loops may include a link list of entities or lines. These entities may

be used to define the boundaries of each face of the part. The bendline detection

operation of step S.126 may then be performed to determine the relationship between the

faces and bendlines of the part. The bendline detection operation of step S.126 may

include bendline detection logic for detecting all of the bendlines between the various faces ofthe part by searching for common edges or line entities between any two adjacent

faces. Further, for faces that connect at more than one area (e.g., when applying the

bendline detection algorithm to a 3-D model - see, e.g., Fig. 12 discussed below) a

number of heuristics may also be applied to detect and select the minimum number of

bendlines for the part. The detected bendlines then may be stored as connecting agents

between the face nodes to produce the final bend graph data stmcture, as shown for

example in Fig. 1 IB.

The bendline detection operation ofthe present invention may be implemented

through software or programmed logic residing at, for example, server module 32. The

purpose of the bendline detection operation is to detect and select the bendlines for the

part so that the part becomes connected with the minimum number of bendlines. The

bendline detection operation may be provided for both 2-D and 3-D versions of the part.

A discussion of an application of the bendline detection operation in connection with an

original 3-D model is provided below with reference to Fig. 12. As noted above, the detected bendlines may be stored as connecting agents between the

face nodes to produce the final bend graph data stmcture. This final bend graph data

stmcture may then be utilized to fold and construct the 3-D version ofthe part from the

2-D data model.

The original 2-D drawing that is provided as input at step S.120 in Fig. 8 may not

include bendline information or such bendline information may be ambiguous and not

uniquely or consistently defined. Thus, the bendline detection operation may be performed to detect the bendlines and their relation to the detected faces of the part.

During this process, the linked list of entities defining each ofthe faces may be analyzed

to determine adjacent edges or line entities that each face has with other faces ofthe part.

This may performed by analyzing all possible contacts between any given two faces. A

contact may be determined based on the presence of a common line entity (or entities that

are within a predetermined distance tolerance of one another) that has a length that is

greater than 0 (i.e., the line entity is not a point but a real line). The geometrical data in

the linked lists may be analyzed to determine the presence of such contact between any

two faces in the part.

If a particular face only has one common edge or area of contact with another

face, then the entity that is common to both faces may be defined as a bendline. For faces

that have common contact at more than one area (e.g., a 3-D model; however, this may

occur with 2-D models as well), a number of heuristics may be applied to detect and

select the minimum number of bendlines for the part. The heuristics that are applied

should be adapted so that the faces ofthe part become connected at the bendlines and so that no continuous loop of faces is formed (since such a bend sheet metal part is impossible to manufacture).

For example, one such heuristic that may be applied is to select the common area

that has the longest contact area as the bendline. Thus, where a face has more than one

common edge with other faces, this heuristic may be applied so that the common entity

having the longest length is selected as the bendline for the face. This heuristic is based

on the principal that is usually better to have a longer contact area when manufacturing

bent sheet metal parts. Another heuristic that may be applied relates to selecting among

different possible combinationsof bendlines (such as when determining the bendlinesfor

a 3-D model). According to this heuristic, when all possible common areas are detected

and various combinations of bendlinesmay be selected, the combination of bendlines that

produces the minimum number of bendlines is selected.

After the bendlines have been detected, the faces ofthe part and the determined

bendlinesmay be displayed to the operator for verification. If the operator is not satisfied

with the selection of bendlines for the part, the bendline detection operation may provide

a manual selection feature to permit the operator at the server module 32 to selectively

indicate the preferred bendlines for the sheet metal part. The operator may indicate to

keep or change a bendline by any suitable input means, such as a mouse or keyboard, etc.

The revised bendlines selected by the operator may then be used for developing the final

3-D (or 2-D) part.

Various processes and operations may be provided to implement the bendline detection operation ofthe present invention. An exemplary code for implementing the

bendline detection operation is provided in Appendix C attached hereto. The sample

code was written in C++ programming language and includes comments to facilitate the understanding of the logic flow therein. The sample code is an exemplary

implementation for the bendline detection operation that may be performed on a 2-D or

3-D model, and includes heuristics (such as those described above) for determining the

optimum selection of bendlines.

The detected face and bendline information may be utilized in the folding and

unfolding process of the invention. By performing a three dimensional rotation around

each bendline during folding or unfolding, the resultant 3-D or 2-D model may be

derived. This task may be accomplished by simply applying matrix transformation,

involving rotations and translations, to each of the faces and other entities of the part.

The features of various commercially available unfolding and folding software

applications may be utilized to implement these basic unfolding or folding steps ofthe

invention. For example, the Amada UNFOLD and FOLD system software may be

utilized to perform these basic operations. The Amada UNFOLD and FOLD system

software is available from Amada America, Inc. (previously operating under the coφorate name of U.S. Amada Ltd.), Buena Park, California. Information concerning

the Amada UNFOLD and FOLD system software may be found in the Amada UNFOLD

Manual for AUTOCAD (March 1994 Edition), the Amada UNFOLD Manual for

CADKEY (May 1994 Edition), and the Amada Windows UNFOLD Manual for

CADKEY (November 1995 Edition), the disclosures of which are expressly incoφorated

herein by reference in their entireties. Further discussion of the folding process to

develop the 3-D model from the 2-D model is provided below with reference to step

S.132.

Referring back to Fig. 8, after the bendline detection operation has been performed at step S.126, server module 32 may prompt the user for pertinent bend and

deduction information for subsequent use during the folding process. For example, at

step S.128, server module 32 may prompt the user to indicate the bend amount for each

bend line, including the bend angle and/or bend inside radius, as well as the bend

direction (e.g., front or back, etc.). At step S.130, the user may also be prompted by the

server module 32 to enter the V-width, material type, and/or the deduction amount. This

information may be utilized to compensate for bend deduction during the folding process.

Depending upon the thickness and type of material used for the sheet metal part, as well

as the angle ofthe bend and the V-width ofthe die to be utilized, the actual sheet metal

part will tend to stretch by a deduction amount during folding ofthe sheet metal part.

In order to compensate for this effect in the model, the deduction amount

information may be utilized so that the dimensions ofthe faces ofthe part are stretched

by half of the deduction amount on each side ofthe bending line when constructing the

3-D model through the folding process. In accordance with an aspect of the present

invention, this deduction amount may be entered directly by the user at the server module

32 (e.g., by keyboard, etc.). Altematively, a material table may be displayed to the

operator that includes the deduction amounts based on the material type and thickness of

the part. The material table may indicate the various deduction amounts based on

different bend angles and V-widths. The user may then automatically set the deduction

amount by selecting a desired V-width and bend angle from the material table displayed

at the server module 32 (e.g., by mouse or keyboard). The inside radius ofthe bend angle

may also be automatically set by the user through the material table when selecting a desired V-width. The deduction amount entered by the operator may be in (or converted after entry

by the operator to) a unit measure of length (e.g., mm) that is identical to that represented

by the part geometry data. During a folding process, the dimensional length of each of

the faces on either side of the bendline may be increased by one-half of the deduction

amount entered for that particular bendline. The dimensional length of the face that is

peφendicuiar to the bendline may be increased by extending the endpoints ofthe entities

defining the boundaries of the faces that are located on either side of the bendline. Such

deduction compensation may also be performed at each ofthe other bendlines of the part

based on the deduction amount provided by the operator for each bend.

In step S.132, a folding process is performed with deduction compensation to

develop the 3-D model based on the processed 2-D, flat drawing. As noted above, the

folding procedure may be carried out by conventional geometric modeling methods,

including the use of matrix transformation and using each of the respective bendlines

defined in the final bend graph data stmcture as a rotational axis. In addition, in order

to compensate for the effect of deduction, when folding and developing the 3-D model,

the faces ofthe part may be stretched by half of the deduction amount on each side ofthe

bendline to more accurately reflect the change in the dimensions ofthe faces when actual

bending ofthe sheet metal is performed.

For example, when performing the folding process at step S.132, the part

geometry and topology data (or bend graph stmcture) may be utilized, along with the

bend parameters (e.g., bend angle, inside radius, etc.). A transformation matrix may be

computed for each face, bendline, hole and forming in the part represented in 2-D space.

Conventional matrix transformation may be applied to the 2-D flat data to get the 3-D space data. The transformation generally involves a rotation followed by a translation.

As noted above, rotation is performed about each bendline axis in accordance with bend

angle amount. Translations are performed for shifting and moving the geometrical data

about space. Such translations may be determined based on the bending radius, bending

angle and deduction amount for each bend. During folding, deduction compensation is

performed so as to stretch or increase the dimensions of the faces by one-half of the

deduction amount on either side of the bendline, as described above. Such deduction

compensation will provide a 3-D representation of the part that more accurately reflects

the dimensions ofthe 2-D sheet metal part when it is folded by bending machinery.

For further information on geometrical modeling and transformations, see, for

example, MORTENSON, Michael M., Geometric Modeling. John Wiley & Sons, New York ( 1988) and FOLEY et al., James, The Systems Programming Series: Fundamentals

of Interactive Computer Graphics. Addison-Wesley Publishing Company, Reading,

Massachusetts (1983), the disclosures of which are expressly incoφorated herein by

reference in their entireties. Chapter 8 of MORTENSON provides a discussion of geometrical transformations, including translations and rotations (see, e.g., pp.345-354).

Further, FOLEY et al. at Chapter 7, pp. 245-265 provides information on geometrical

transformations, including matrix representation of 2-D and 3-D transformations.

Additional information of modeling and geometrical transformations may also be found

in MANTYLA.Martti. An Introductionto Solid Modeling. Computer Science Press, Inc.,

Rockville, Maryland (1988), the disclosure of which is expressly incoφorated herein by

reference in its entirety. Information on coordinate transformations may be found at pp. 365-367 of MANTYLA. Referringnow to Fig. 12, a description ofthe processes and operations that may

be performed for developing a 2-D model based on an original 3-D, flat drawing (with

no thickness) will be provided, in accordance with another aspect of the present

invention. Similar to the folding process described above with reference to Fig. 8, the

various processes and operations for unfolding a 3-D drawing and developing a resultant

2-D model may be implemented through the software and/or programmed logic at server

module 32. As shown in Fig. 12 at step S.140, the original 3-D, flat drawing which was

provided or developed based on the customer's specification may be entered or imported

into server module 32. The 3-D drawing may be stored as a DXF or IGES file and may be entered by interfacing with or utilizing a CAD or CAD/CAM system from server

module 32. After entering the 3-D drawing, an auto-trimming and cleanup operationmay

be performed at step S.142 by server module 32 to prepare the drawing for subsequent

face detection and other processing. As discussed above with reference to Figs. 9A-9E,

the auto-trimming and cleanup functions may break and connect entities and surfaces so

that the various faces of the part may be properly detected and defined.

The auto-trimming and cleanup operation described above with reference to Figs.

8 and 9 may similarly be applied to the geometric data o the 3-D drawing entered at step

S.140 of Fig. 12. Instead of analyzing the data in 2-D space (as was the case with the 2-D

flat drawing), each ofthe entities (e.g., lines, arcs, etc.) represented in the 3-D drawing

may be analyzed based on the 3-D coordinate and space information provided therein.

The intersection points and open intersection areas may be detected by analyzing each

entity individually and comparing it with other entities one at a time. Once again, basic

geometrical analysis ofthe endpoints and other attributes of the entities may be utilized to determine intersection points and open intersection areas within tolerance.

After performing auto-trimming and cleanup functions on the 3-D drawing, at

step S.144 a face detection operation may be performed to detect and define each ofthe

faces ofthe sheet metal part. Face detection for the 3-D drawing may be performed by

analyzing and detecting each ofthe faces in 2-D space and developing a loop tree, similar

to that described above. Face detection may be executed by starting at any predetermined

entity. For example, the left most entity (i.e., the entity with the lowest X-coordinate)

may be used as the initial entity. Thereafter, a plane may be defined by taking the initial

line entity and another connecting or adjacent line entity (e.g., any entity with a common

endpoint to the initial entity). A face detection operation may then be performed using

loop and entity analysis, such as that described above with respect to Figs. 10A- 1 OH. As each entity is detected within the defined 2-D plane, the various outside and inside loops

may be defined and the entities may be marked (e.g., by setting or incrementing a flag

ofthe selected entity) to indicate that they have been selected and included in a linked list

defining one ofthe loops in that plane.

Subsequent loop analysis may then be performed in the other 2-D planes that

comprise the 3-D drawing. In order to proceed with loop analysis ofthe other entities,

additional planes may be defined by searching for unmarked or unselected entities within

the 3-D drawing. Such planes may be defined between two unselected entities or an

unselected entity and a previously selected entity that was previously analyzed. In each

ofthe additional 2-D planes, further loop analysis may be performed to detect the inside

and outside loops. Once again, linked lists of connecting entities may be maintained and

the selected entities marked (e.g., by incrementing a flag associated with the selected entity) as each ofthe loop paths are defined.

After all of the entities have been detected, the resulting loops may be used to

develop a loop tree for each ofthe 2-D planes that were analyzed. As discussed above,

a loop tree may be provided to determine the faces and opening or holes in the sheet

metal part. For a 3-D drawing, a loop tree may be developed for each ofthe planes ofthe

sheet metal part. The loops detected within each plane may be grouped and analyzed to

develop each loop tree. The root of each tree may be defined as the outside loop detected

in the plane, with each inside loop of that plane that has a common entity with the outside

loop. If additional entities (e.g., holes or openings) are detected within the inside loops

of the plane, then these loops may be defined as children of the inside loops (i.e., the

grandchildren of the root ofthe loop tree) within which they are located. The generated

loop trees may then be used to detect all ofthe faces ofthe 3-D drawing. The detected

faces may then be stored as nodes in a bend graph data stmcture.

The resultant bend graph stmcture may then be supplemented with the connecting

bending line connecting agents after performing a bendline detection operation at step

S.146. The bendline detection operation and the development of the final bend graph

stmcture or part topology may be carried out in a similar manner to that described above

with reference to Figs. 1 1 A and 11 B.

As discussed above, an exemplary code for implementing the bendline detection

operation is provided in Appendix C attached hereto. The sample code is an exemplary implementation for the bendline detection operation that may be performed on a 2-D or 3-D model, and includes heuristics (such as those described above) for determining the

optimum selection of bendlines. The bendline detection operation may include a manual

selection feature to permit the operator at the server module 32 to selectively indicate the

preferred bendlines for the sheet metal part if not satisfied with the bendlines that were

detected. The operator may indicate to keep or change a bendline by any suitable input

means, such as a mouse or keyboard, etc. The revised bendlines selected by the operator

may then be used for developing the final 2-D part.

Before performing an unfolding process about the bending lines ofthe final bend

graph structure, the user may be prompted for the V-width, material type and or

deduction amount at step S.148. As discussed above, since metal tends to stretch when

it is folded, the dimensions of the 3-D part will be slightly larger than that ofthe 2-D flat

part. As such, during unfolding of the sheet metal part, the dimensions of the part should

be shrunk or reduced by the deduction amount based on the material type and V-width

selected. Accordingly, in accordance with an aspect ofthe present invention, a shrinking

process may be performed when unfolding the 3-D model to more accurately develop the

2-D model and the respective dimensions of its surfaces. As described above, the

deduction amount may be entered directly by the user or a material table may be

displayed in order to enable the user to automatically set the deduction amount by

selecting a desired V-width and bend angle.

The deduction amount entered by the operator may be in (or converted after entry

by the operator to) an unit measure of length (e.g., mm) that is identical to that

represented by the part geometry data. During an unfolding process, the dimensional

length of each of the faces on either side of the bendline may be decreased by one-half - 98 - of the deduction amount entered for that particular bendline. The dimensional length of

the face that is peφendicuiar to the bendline may be decreased by reducing the endpoints

ofthe entities defining the boundaries ofthe faces that are located on either side ofthe

bendline. Such deduction compensation may also be performed at each of the other

bendlines of the part based on the deduction amount provided by the operator for each

bend.

After entry of all of the necessary data, an unfolding process may be performed

at step S.150 to develop the 2-D model. Conventional methods may be used for

unfolding the 3-D bend model, including the use of matrix transformation with each of

the bending lines being used as a rotational axis. During the unfolding process, each of

the bend angles may be measured and the part may be unfolded by the bend angle amount to develop the flat 2-D model. In addition, based on the deduction amount entered, a

shrinking or reduction of the dimensions of the faces may be performed by half of the

deduction amount on each side of the bending line to more accurately simulate the

physical characteristics of the sheet metal material and the difference between the 3-D

and 2-D models.

When performing the unfolding process at step S.150, the part geometry and

topology data (or bend graph stmcture) may be utilized, along with the bend parameters

(e.g., bend angle, inside radius, etc.). A transformation matrix may be computed for each

face, bendline, hole and forming in the part represented in 3-D space. Conventional

matrix transformation may be applied to the 3-D data to get the 2-D space data. The

transformation generally involves a rotation followed by a translation. As noted above, rotation is performed about each bendline axis in accordance with bend angle amount. For unfolding, rotation is carried out in the reverse direction until there is a 180 degree

angle between the two faces (i.e., until the part is flat). Translations are performed for

shifting and moving the geometrical data about space. Such translations may be

determined based on the bending radius, bending angle and deduction amount for each

bend. During unfolding, deduction compensation is performed so as to shrink or

decrease the dimensions of the faces by one-half of the deduction amount on either side ofthe bendline, as described above. Such deduction compensation will provide a 2-D

representation of the part that more accurately reflects the dimensions ofthe sheet metal

part before it is folded during a bending operation.

Once again, further information on geometrical modeling and transformations

may be found in MORTENSON, FOLEY et al. and MANTYLA. As indicated above,

Chapter 8 of MORTENSON provides a discussion of geometrical transformations,

including translations and rotations (see, e.g., pp. 345-354). Further, FOLEY et al. at

Chapter 7, pp. 245-265 provides information on geometrical transformations, including

matrix representation of 2-D and 3-D transformations. In addition, information on

coordinate transformations may be found at pp. 365-367 of MANTYLA.

As discussed above with reference to Fig. 3, if a 2-D, three view drawing or a 3-

D, wire frame drawing with thickness is originally provided or developed based upon the

customer's order, then further processing will be required in order to develop a 3-D model

without thickness; and, thereafter, the developed 3-D model without thickness may be

used to create a 2-D model by applying an unfolding process or algorithm. Figs. 13-15 illustrate the various processes and operations that may be applied in order to develop a

3-D model based on an original 2-D, three view drawing. Further, Fig. 16 illustrates, in accordance with another aspect of the present invention, the additional processes and

operations that may be applied for developing a 3-D model without thickness from an

original 3-D, wire frame drawing with thickness. Once again, the various processes and

operations depicted in Figs. 13-16 may be implemented through the software and/or

programmed logic residing at, for example, server module 32.

Referring now to Fig. 13, a description of the logic flow of the operations and

processes that may be performed in order to develop a 3-D model (with no thickness)

based on an original 2-D. three view drawing will be provided, in accordance with the

teachings of the invention. Initially, the 2-D, three view drawing may be entered or

imported to server module 32 at step S.160. The original 2-D, three view drawing may

comprise various views ofthe part (e.g., a front view, a top view and a right side view -

see, e.g., Figs. 14B and 14C) and may be a CAD drawing such as a DXF or IGES file that

may be downloaded or imported to server module 32. Thereafter, at step S.162, a 2-D

cleanup operation may be performed by server module 32 in order to prepare the drawing

for subsequent processing into the 3-D model. The 2-D cleanup operation may be performed in order to eliminate extraneous and non-geometrical information, including

text, centerlinesand dimension lines, which do not represent the actual geometry of the

part. The 2-D cleanup operation may also be performed in order to connect all exterior

lines at, for example, their connecting ends and to break and trim any intersecting lines

or entities. Fig. 14A illustrates an example ofthe logic flow ofthe various processes that

may be carried out when the 2-D cleanup operation is performed by server module 32.

As shown in Fig. 14A, the 2-D drawing is first read from a data file or loaded at step S.180 by server module 32. Thereafter, at step S.182, server module 32 may analyze the respective entities and geometrical data in the 2-D drawing and break the various

entities in order to prepare the drawing for further processing. The break and trimming

function performed at step S.182 may be carried out in a similar manner to that described

above with respect to the auto-trimming and cleanup function of the present invention.

Thus, at step S.182, all of the geometrical data in the 2-D, three view drawing may be

analyzed in order to detect the intersection of entities and any open intersections that are

within tolerance. Any intersecting lines may be broken, with the resultant entities

meeting at a common endpoint defined by the intersection point. Further, for entities

having an open intersection area that is within a predetermined tolerance (e.g., 0.0 - 0.01

mm or 0.0 - 0.001 inches), such entities may be joined, in a similar manner to that

described above with respect to, for example, Fig. 9E.

At step S.184, the perimeter of the 2-D drawing sheet may be detected and any

exterior lines or data (such as border lines, coordinate grids and numbers, etc.) may be

eliminated. As shown in Fig. 14B, a 2-D, three view drawing will often be provided on

a drawing sheet. The drawing sheet may include extraneous and non-geometrical

information that is not necessary to process the views of the sheet metal part. As such,

during step S.184. this type of information may be detected and eliminated from the 2-D

drawing when developing the 3-D model utilizing the 2-D clean-up process of the

invention.

The 2-D drawing data may include key words or type fields to indicate the type

of data contained therein (e.g., geometrical or non-geometrical/text). Thus, these key

words or type fields, which are provided based on the data format of the drawing file,

may be used to eliminate various extraneous information, such as text and other non- geometrical data. However, further processing is usually necessary in order to properly

eliminate all of the unwanted drawing sheet data. Often, the border lines and other

outside information are saved as entities (e.g., lines, etc.) that can not easily be

distinguished based on the data key words or type fields. As such, according to an aspect

ofthe present invention, a connectivity graph stmcture may be developed when analyzing

the data of the 2-D drawing. The connectivity graph stmcture may indicate for each

entity a list of incident vertices and a list of connected entities. For each vertex, a list of

adjacent vertices and a list of entities on which it is incident may also be provided. With

this graph stmcture, which may be developed when performing the break and trimming

function of step S.182, it may be determined which entities are connected by matching

endpoints. As a result, extraneous data, such as as border lines, information boxes and

other non-geometrical data, may be eliminated since this data will typically not be

constmcted with or include connecting entities.

As noted above, the 2-D, three view drawing may include extraneous information,

such as dimension lines, arrow lines, centerlines, and text, which do not represent the

actual geometry of the part. These entities may be detected at step S.186 and deleted

from the 2-D data file in order to prepare the 2-D drawing for further processing. The

detection of these extraneous entities may be performed automatically by server module

32 (e.g., by detecting items in the 2-D data file that do not relate to the actual geometry

of the part). For example, by using the connectivity data graph stmcture, two-sided

opened entities (e.g., lines used for underlining text or to indicate a dimension or

centerline in the part) may be detected and eliminated. Other entities, such as arrows, may also be detected based on the presence of floating endpoints and other characteristics of such entities. In order to effectively eliminate all unnecessary data, the server module

32 may also provide a manual editing function to permit an operator to indicate (e.g., by

mouse or keyboard) which items in the 2-D drawing should be eliminated. Through the

assistance or confirmation of he operator, additional extraneous information may thus

also be removed from the drawing.

After step S.186, the various views in the 2-D drawing may be grouped and then

respectively defined at step S.188. In accordance with an aspect ofthe present invention

server module 32 may support predefined or standard views and orientations, such as a

top view, front view, right side view layout, such as that shown in Figs. 14C and 14D.

Other views and layouts, such as combinations of a top view, a front or back view, and a right or left view, may also be supported. As further described below, server module

32 may also support rotated views (see, for example, Fig. 14D) in order to process the

views in the 2-D drawing into the 3-D representation of the part. In any event, at least

two (and preferably three) different views of the part with thickness representations

should be provided so that a 3-D model ofthe part may be constmcted. By analyzing the

connectivity and grouping of the entities in the connectivity graph stmcture, server

module 32 may group and define the views based on the relative position and/or

coordinate location of each ofthe views.

By way of a non-limiting example, the definition ofthe views by server module

32 may be carried out in accordance with a predefined or customary arrangement or

layout for analyzing the views in the data file, and/or based upon detecting the orientation of the views and matching the various dimensions of the part in each of the respective

views in the drawing. A predefined or canonical form, such as that shown in Fig. 14E, may be used to determine and define each ofthe views according to possible view types.

Geometric comparisons ofthe various endpoints and relationships between the entities

defining each group may be performed in order to effectuate step S.188. The view

detection feature of server module 32 may label each of the views according to one of

plurality of possible view types (e.g., top view, front view, back view, left view, right

view). The detection of each of the views may be based on a predefined or canonical

view layout or form, and based on the detected relationships between each ofthe views

that are present.

Various processes and operations may be used at step S.188 to group and define

the views in the 2-D, three view drawing. For example, after accessing the processed 2-

D, three view drawing, the server module 32 may first identify the top view ofthe part in the drawing data. The top view may be detected based on the predefined or canonical

form, view layout (such as that in Fig. 14E). If three separate views are detected in either

a horizontal or vertical direction, then the center view may be defined as the top view.

Further, if three separate views are not detected, and only two separate views are detected

in a vertical direction, then the upper view may be defined as the top view. Once again,

the connectivity and grouping ofthe entities in the connectivity graph stmcture may be

utilized to detect each of the views. A stored lookup table or matrix representing the

predefined or canonical form may be used to compare the views of the 2-D drawing and

detect each of the views.

After detecting the top view from the 2-D, three view drawing data, the other

views of the part may be detected based on the relative position of each ofthe views to the detected top view. For example, based on the canonical view layout of Fig. 14E, if a view grouping is located above the top view, then the view may be defined as a back

view. If, however, a view grouping is located below the top view, then the view may be

defined as a front view ofthe part. Further, a right view and a left view may be detected

based on their relative position on the corresponding right hand side and left hand side,

respectively, ofthe top view. Thereafter, any remaining views, which do not conform

to the canonical form (such as Fig. 14E), may be detected based on their relative position

to the detected views (e.g., a detected back view or front view). For example, for the

layout B shown in Fig. 14D, the right view has been provided in a rotated position

relative to the top view. The right view in layout B, however, may still be detected based

on its relation to the detected front view. That is, undetected views that are present on

the right hand side or left hand side of a detected back or front view may be defined,

respectively, as a right view or a left view ofthe part.

Various predefined or canonical view layouts may be used to detect and define

views in the 2-D. three view drawing. The canoncial forms (such as that in Fig. 14C or

Fig. 14D) may be selected based on the numbers of view types that are to be supported

and/or based on the view layouts that are more prevalent or selected required by the

manufacturing facility. If any views are not detected, a warning signal may be provided

by the server module so that an operator may modify the 2-D, three view drawing data

in accordance with the preferred view layout, or take other appropriate action. In addition

to providing a predefined or canonical form for detecting the views in the 2-D drawing,

a predefined or canonical form (such as layout A in Fig. 14D) may also be provided for

processing the detected views and developing the 3-D model of the part. As such, a rotated view feature may be provide to properly group the detected views in accordance with the canonical form before further processing is performed.

As mentioned above, the 2-D cleanup operation may support and detect rotated

views which do not conform to the predefined or canonical form for detecting views in

a drawing. With the rotated view option, non-conforming views that have been detected

may be rotated or translated so that each of the views conform to the predefined or canonical view form for processing and developing the 3-D model ofthe part. Assuming

a canonical form such as that illustrated in Fig. 14E for detecting views of the part, each

ofthe views in layout B in Fig. 14D may be detected based on the relative position ofthe

views to the top view and the other detected views, as described above. If, for example, the layout A in Fig. 14D is to be used as a predefined or canonical view layout for

processing the various views in a 2-D drawing having a top view, front view and a right

view, then at step S.188 the right view in layout B may be rotated by 90 degrees to

provide a modified view layout for the part that is similar to layout A. By rotating the

right view in layout B by 90 degrees so that the right view of the part is provided on the

right side of the top view of the part, the views in the drawing may be processed in

accordance with the canonical form represented by layout A. A stored lookup table or

matrix representing the predefined or canonical form may be used to compare the views

ofthe 2-D drawing and to determine which views require rotation and translation.

In order to ensure that an accurate 3-D model ofthe part is developed from the

views in the 2-D drawing, the respective dimensions in each of the views should be

checked for consistency and matching. As further shown in Fig. 14A, at step S.190,

the boundaries ofthe views in the data file may be detected in order to confirm that all of the dimensions of the respective views are in scale with one another. If it is determinedthat the views do not match within a predetermined tolerance (e.g., 0.0 - 0.01

inches), then appropriate modification may be made at step S.190 to redimension any

particular view in order to ensure that all ofthe views are provided in the same scale. A

warning component may be provided in server module 32 to alert a user that the view

dimensions do not match so that the necessary modifications can be made to the 2-D

drawing data that is present.

Various operations and processes may be utilized to detect and verify the

consistency of the dimensions in the respective views of the part. For example, the

corresponding dimensions of each of the views may be compared to determine if they are

within a predetermined tolerance of one another. Such an analysis may include

comparing the line entities that define the boundaries of each view ofthe part. Assuming the canonical form in Fig. 14E, a top view may be detected as matching a right view or

a left view if the difference, for each view, between a maximal Y coordinate position and

a minimal Y coordinate position is within a predetermined tolerance (e.g., 0.0 - 0.01

inches). Further, the top view may be detected as matching a front view or a back view

if the difference, for each view, between a maximal X coordinate position and a minimal

X coordinate position is within a predetermined tolerance (e.g., 0.0 - 0.01 inches).

Moreover, the left view or right view may be determined as matching a front view or a

back view if the difference between a maximal X coordinate position and a minimal X

coordinate position compared to the difference between a maximal Y coordinate postion

and a minimal Y coordinate position is within a predetermined tolerance (e.g., 0.0 - 0.01

inches). Once again, a warning component or module may be provided in server module

32 to alert a user when the view dimensions or related face dimensions do not match so that the necessary modifications can be made to the 2-D drawing data that is present.

Finally, at step S.192, the inside loops, holes and shapes of the part may be

detected in accordance with the teachings ofthe face detection procedure ofthe present

invention. The various holes and shapes provided on the interior of the faces of each

view may be detected by looping through the various lines and boundaries of the part

from the exterior of the part towards the center. Loop and entity analysis may be

performed on each view ofthe part in the 2-D drawing. By analyzing each view from the

outside and working inwardly towards the center of the part, the detected loops will

define the boundaries and areas of the material and openings of the part based upon a cyclic sequence (e.g., material, opening, material, etc.). A loop tree, such as that in Fig.

10H, may be developed for each view to determine the location of the faces and any

openings within each face. Unconnected entities, such as floating arcs or lines, within

the faces of the part may also be detected and eliminated during step S.192.

An exemplary code for performing the 2-D clean-up operation of the present

invention is provided in Appendix D. The code was written in C++ programming

language and includes comments to facilitate the analysis of the logic and algorithms

used therein. The code includes the various processes and operations ofthe 2-D clean-up

mode, such as those discussed above with reference to Figs. 14A-14C.

Referring back to Fig. 13, after a 2-D clean-up operation has been performed,

logic flow will then continue to step S.164 where it may be determined whether the 2-D

drawing represents or includes the thickness of the material (i.e., whether the 2-D

drawing is with thickness). If the 2-D drawing is determined to contain the thickness

amount, then at step S.166, an eliminate thickness procedure may be performed by server module 32 in order to prepare the 2-D drawing for subsequent processing into a 3-D

model. The determination of the existence of thickness in the 2-D drawing may be

performed automatically by server module 32 based on the data ofthe drawing, or may

be performed by the server module through the assistance or response from the operator

(e-g-, the operator may be prompted to indicate whether thickness removal is necessary

or desired). The thickness ofthe part may be eliminated due to the unique symmetry of

all sheet metal parts. By eliminating the thickness ofthe part, the resultant sheet metal part with no thickness may be more easily analyzed by a sheet metal operator or designer.

Further, the inventors of the present application have found that by eliminating the

thickness of the 2-D, three view drawing, the time required to convert the 2-D drawing

and develop the 3-D model may be substantially reduced.

Since most 2-D, three view drawings include a material thickness amount, an

operator may often be confused by which bend lines should be selected in order to make

a 3-D model from the 2-D drawing. As a result, considerable time is wasted in selecting

the appropriate bend lines so that the 2-D, three view drawing may be converted into a

3-D model. An example of a three view, 2-D drawing with thickness is shown in Fig.

15 A. According to an aspect of the present invention, an eliminate thickness procedure

may be provided to display a simplified 2-D, three view drawing model which is

represented and processed with no material thickness, but retains the material thickness

amount and the inside or outside dimensions ofthe part in the bend model data. Fig. 15B

illustrates the simplified three view, 2-D drawing that may be viewed and displayed to

the operator at server module 32 after performing the eliminate thickness procedure.

When the eliminate thickness procedure is executed, the user may be prompted to specify the material thickness in the 2-D, three view display, and may be prompted to

specify which dimension (i.e., the outside dimension or inside dimension) is to be

retained in the display. The operator may indicate the thickness and surface to retain in

one of the views by use of, for example, a mouse. Based upon the data entered by the

user, server module 32 may modify the 2-D, three view drawing to eliminate the material

thickness indicated by the user and to retain the inside or outside dimension based on the

operator's selection.

In order to eliminate the thickness in the 2-D, three view drawing, the server

module 32 may analyze each of the three views based on the selections made by the

operator. The selected surface may be projected into each of the other views by

geometrical computations (e.g., by detecting the corresponding entities that lie in the

same X-coordinate or Y-coordinate projection as the selected entity line or surface) to detect the corresponding entities and lines in each of the views. The corresponding

entities may be marked and retained, with the nonmatching entities or surfaces being

eliminated or not displayed on the screen, such as that shown in Fig. 15B. Further, the

thickness dimension line indicated by the operator may be similarly projected into each

of the other views, with the matching thickness dimension lines or entities being

eliminated, as further shown in the example if Fig. 15B. As a result, each ofthe views

in the drawing may appropriately modified and then displayed to the user at the server

module 32. The resultant 2-D, three drawing with no thickness may also be used for

subsequent processing in order to develop the 3-D model of the part.

The thickness elimination procedure of the present invention may include a

manual thickness elimination mode to permit an operator to selectively indicate in each - i ll - view the thickness lines to be eliminated and the surface entities to be retained. A mouse

or other appropriate input device may be used by the operator to indicate which areas in

each ofthe displayed views are to be eliminated and which surfaces are to be retained.

Based on the data entered by the operator, the server module 32 may eliminate each line

entity selected by the operator from the 2-D, three view drawing in order to provide a

drawing with no thickness.

The present invention may also include a warning system or module to analyze

and detect whether all the thickness representations have been properly identified in the

2-D, three view drawing, and to alert the user when there are unmarked thickness

components and or when there are inconsistencies in the drawing data. For example, a

thickness warning component may be provided to highlight on the display screen

possible unmarked thickness segments, and a face warning component may be provided

to highlight on the screen possible mismatched faces when the face dimension does not

match the thickness mark in another view. A bendline warning component may also be

provided to highlight inconsistent bendlines and highlight mismatching thickness arcs.

An arc may be highlighted when at least one bendline projected on this arc is not bound

by two cross thickness lines. For instance, Fig. 15C illustrates a thickness arc that is

properly bounded by two or another non-zero even number of cross-thickness lines (i.e.,

a small line crossing the thickness in one ofthe views). Each bendline should also be

bound by two or another non-zero even number of cross-thickness lines. The analysis

of these entities ofthe part in each view may be based on performing a loop analysis on and analyzing the connectivity ofthe line and arc entities that make-up each view. An

open thickness line may be defined based on a thickness line that has at least one end point that is not connected to another thickness line or arc. A side that includes an open

thickness line may be defined as an open thickness side. A thickness line may be

highlighted if the open thickness side of an open thickness line does not match the

bounding box of a minimal loop. By providing such warnings relating to the processed

2-D, three view image to a user, the user may be alerted of inconsistenciesin the drawing

data, and the user will thus be able to modify and/or correct the drawing data before

further processing is performed to develop the 3-D model of the part. Inclusion of such

a warning system and user interaction also improves the accuracy of the representation

ofthe part by the 3-D model.

At step S.168 in Fig. 13, the processed 2-D, three view drawing with no material

thickness may then be converted and developed into a 3-D model. The conversion and

development of the 3-D model from the 2-D, three view drawing may be performed by

using well-known or established projection and/or extrusion methods. For example, in

order to develop the 3-D model from the three view, 2-D drawing, the depths of each of

the views may be detected and then each ofthe views projected in order to develop a 3-D model. The resultant 3-D model may then be used when developing the bend model data

and also converted into a single view, 2-D flat drawing by applying the above-described

unfolding algorithm. For more information geometric modeling techniques, see

MORTENSON, FOLEY et al. and MANTYLA. For additional information on

projection techniques to constmct 3-D models from 2-D drawings, see, for example,

WESLEY et al., W.A., Fleshing Out Projections. IBM J, Res. Develop., Vol. 25, No. 6,

p.934-954 (1981); AOMURA, Shigeru, Creating Solid Model with Machine Drawings.

The 6th Computational Mechanics Conference, JSME, No. 930-71, Japan, pp. 497-98 (1993); and AOMURA, Shige , Recent Trends and Future Prospect of Research and

Practical Use (Automatic Reconstructionof 3D Solid from Drawings). Toyo Engineering

Coφ., Japan, pp. 6-13 (1995); the disclosures of which are expressly incoφorated herein

by reference in their entireties.

When developing the 3-D model at step S.168, an additional clean-up process

may be included in order to further process and refine the resultant 3-D model. In

accordance with an aspect of the invention, a 3-D clean-up process may be provided to

compensate for ambiguities that are present in the 2-D, three view drawing ofthe part and

that create extraneous or superfluous information in the developed 3-D representation of

the part. As will be appreciated by those skilled in the art, a 2-D, three view

representation of a part includes ambiguities concerning the representations of various

features ofthe part in 3-D coordinate space. When developing the 3-D model from the

three view, 2-D drawing, extraneous and superfluous information may be generated as

a result of these ambiguities. As such, according to an aspect of the invention, the 3-D

clean-up process may include processes and operations for detecting and removing one

sided open lines and for detecting and cleaning bendlines and trimming faces. The 3-D

clean-up process may be performed automatically when developing the resultant 3-D

model ofthe part, or may be selectively performed based on input from an operator when

the developed 3-D model is determined to require further processing.

According to the 3-D clean-up process, by analyzing the developed 3-D drawing

data, every line or arc which is determined to be not connected to another entity at one

of its endpoints may be identified and defined as a one sided open line. Any entity that

is determined to be a one sided open line may be removed from the 3-D representation of the part. Once an open line is removed, it may cause another line or entity to be open.

As such, new one sided open lines are also identified and removed recursively until all

open lines or entities are removed. Fig. 49A illustrates an example of a 3-D

representation of a part before one sided open lines are removed, and Fig. 49B illustrates

the part after the one sided open lines have been removed from the 3-D representation.

As noted above, the 3-D clean-up process that may be performed at step S.168

may also include a process for detecting and cleaning bendlines. Bendlines may be

identified and cleaned (e.g., by adding mold lines) in order to facilitate the detection of

face information of the part in 3-D space. Based on the developed 3-D model data, each

bendline may be identified based on the detection of a pair of 3-D arcs (e.g., represented

by arc entities in the drawing data) which have the same normal defined by their centers. During this process, mold lines may be added to the bendlines that are identified. The

mold lines may be added by identifying corresponding endpoints in each pair of 3-D arcs,

and extending mold lines (e.g., represented by line entities) between the corresponding

endpoints of the 3-D arcs. Fig. 50A illustrates an exemplary 3-D representation of a part

before the bendlines have been identified, and Fig. 50B illustrates the part after the mold

lines (represented by dashed lines in the drawing) have been added.

After the bendlines have been identified and the mold lines have been added, the

3-D clean-up process may further process the 3-D representation ofthe part to further

clean all bendlines and trim the faces of the part. Due to frequent ambiguities in the

views of the 2-D, three view drawing data, superfluous sections of the faces may be

generated in the 3-D representation of the part. The 3-D clean-up process may identify these superfluous sections of the faces and trims the faces using sheet-metal domain knowledge (e.g., knowledge relating to what is unfoldable). Other extraneous

information, such as extra holes or openings, may also be identified and eliminated. As

a result, the superfluous sections ofthe part may be removed and the 3-D representation

may provide a more accurate representation of the sheet metal part. Fig. 51 A illustrates

an exemplary section of a part before cleaning the bendlines and trimming the faces, and

Fig. 5 IB shows the section ofthe part after cleaning and trimming has been performed.

Fig. 16 illustrates an example of the logic flow of the processes and operations

that may be performed in order to develop a 3-D drawing with no material thickness from

an original 3-D drawing with material thickness. At step S.200, the original 3-D drawing

with material thickness may be entered or imported to server module 32. The 3-D model

may be a 3-D, wire frame drawing with material thickness and may be a CAD drawing

file, such as a DXF or IGES file. After the 3-D drawing has been imported to the server

module 32, an eliminate thickness procedure may be performed at step S.204. The

eliminate thickness procedure at step S.204 on the 3-D model may be performed in a

similar manner to that provided in the Amada UNFOLD software system, described above. In order to eliminate the thickness in the 3-D model, the operator may first be

prompted to indicate the thickness and to select the surface to be retained. Based on the

operator's selection, the thickness is measured by analyzing the endpoints ofthe entity

line defining the thickness. Thereafter, the boundaries of the selected surface may be

traced, in a similar manner to that described above with respect to the loop and entity

analysis process, with the entities to be kept being marked (e.g., by setting or

incrementing a flag) and the corresponding thickness entities being eliminated. When tracing the entities ofthe 3-D part, the entities may be distinguished based on the length of the thickness entity selected by the user. Generally, all entities having the same length

ofthe thickness entity may not be selected and eliminated, with the other entities that are

not of the same length being marked and retained. Any remaining entities that are not

marked during the surface trace ofthe 3-D part may also be eliminated. Once again, the

server module 32 may provide a manual thickness removal mode, in which an operator

may manually indicate each entity in the 3-D part that is to be removed.

After step S.204, the resultant 3-D model with no material thickness may be

developed and or displayed to the operator at step S.206. An unfolding algorithm or

process may then be applied to the 3-D model with no material thickness to develop the single view, 2-D flat drawing for the bend model data, as described in greater detail

above.

As discussed above, the design and manufacturing information that is stored in

database 30 may include a bend model data file comprising part geometry and topology

as well as manufacturing data for the sheet metal component. Further, the software that

may be used to implement the various features ofthe invention may be developed using

a high level programming language such as C++ and by using object oriented

programmin techniques. Different object oriented techniques may be used, such as

Booch or OMT to implement the various features of the invention. If object oriented

programming is utilized, an object oriented data model may be utilized to represent the

sheet metal part and the bend model for the part may be implemented through a

completely self-contained class library. In accordance with an aspect of the present

invention, a description will now be provided of any exemplary data stmcture and access

algorithm for the bend model, based on object oriented programming techniques. Fig. 17 illustrates an exemplary data stmcture and access algorithm ofthe bend

model that may be utilized when implementing the present invention through object

oriented programming. Object oriented programming is a type or form of software

development that can model the real world by combining objects or modules that contain

data as well as instructions that work upon that data. In object oriented programming,

objects are software entities that may model something physical, like a sheet metal part,

or they may model something virtual, like business transactions. Objects may contain

one or more attributes (i.e., fields) that collectively define the state ofthe object, and may

contain an identity that distinguishes it from all other objects. In addition, objects may

include behavior that is defined by a set of methods (i.e., procedures) that can modify the

attributes or perform operations on the object based on the existence of certain

conditions.

According to an embodiment ofthe present invention, the sheet metal part may

be represented as an object oriented data model. As shown in Fig. 17. the bend model

for the sheet metal part may be defined as a completely self-contained class library. All

of the required data manipulation and functions for the sheet metal part (e.g., folding,

unfolding, etc.) may be captured as member functions of the class library. All of the

geometrical and topological data may be defined in objects that are grouped within the

bend model. The bend model class library may be a hierarchy of classes or objects with

a part class being the top level class in the hierarchy. The part class may include a part

object with various part attributes and may have various objects that define the part and

the actions that may be performed on or to the part.

Fig. 17 shows an example ofthe various objects that may be grouped in the bend model class library. For example, a part class 50 may be provided that includes various

attributes 52. The part attributes 52 may include various part information such as the part

number and/or name, the part material type and the thickness of the part. The attributes

52 may also include bend sequence information for indicating the order in which the

bends are to be performed and other manufacturing information such as tolerance

requirements for the various dimensions of the part. Part class 50 may also comprise

various objects, such as a faces object 54, a holes object 56, a formings object 58, and a

bendlines object 60, as shown in Fig. 17. Each of the objects 54, 56, 58 and 60 may

actually consist of a group of objects for each ofthe entities (e.g., faces, holes, formings,

and bendlines) represented therein. The faces object 54, holes object 56, formings object

58 and bendlines object 60 may each include geometry and dimension data, location and

coordinate data in both 2-D and 3-D space representations, and data relating to the edges

and surfaces of their respective entities (e.g., faces, holes, formings, and bendlines) of the

part. For example, the faces object 54 may include geometry and dimension data for each

of the faces, location space data of the faces in both 2-D and 3-D representation, and

edges and surface data for the edges and surfaces of the faces. In addition, the formings

object 58 may include data relating to special formings in the part, including geometry

and dimension data, 2-D and 3-D location space data, and edges and/or surfaces data.

As further shown in embodiment of Fig. 17, the part class 50 may also include a

topology object 62 and a bend properties object 64. The topology object 62 may include

the part topology data for the faces, holes, formings and bendlines of the part. The data

in the topology object 62 may indicate the stmcture and geometric relationships ofthe

various features of the part. The bend properties object 64 may also be provided and include information concerning special manufacturing constraints for one or more

features ofthe part. For example, bend properties information concerning how the sheet

metal part should be bent may be provided in the bend properties object 64. The bend

properties information may include specific manufacturing data for different bend

property types (e.g., simultaneous bend, colinear bend, Z-bend. etc.).

The bendlines object 60 may also include manufacturing specific data relating to

the bends to be performed. Thus, in addition to providing geometry and dimension data.

2-D and 3-D location space data, and edges data for each bendline, the bendlines object

60 may also include V-width data, bend pitch data, bend count data and/or orientation

data for each of the bendlines. Each of the bendlines may also include an associated

bending operation, as shown in Fig. 17. The bending operations may be implemented as

a group of objects with data and operations/instructions for performing bends at each

bendline. If provided as an object, each bending operation may include data and

instructions indicating how and what type of bending is be performed (e.g., conic bend,

Z-bending, Hemming, arc bending, etc.), as well as pertinent bend data such as the bend

angle, bend radius and/or bend deduction amount.

By implementing the bend model of the part through an object oriented data

model, all ofthe complex mathematical calculations, computational geometry and matrix

transformationsmay be built into a single class library. Special bending operations, such

as hemming, Z-bending and arc bending, may also be captured inside the class library.

Further, manufacturing information, such as the V-width, the bend deduction amount,

and the bend sequence, may be also captured within the class library. With the bend

model, simultaneous dual representation of both the 2-D flat model and 3-D model may be effectuated, as shown in Fig. 17. Further, bending operations may be performed in

accordance with the bend lines object 60 of the bend model. General comments

regarding the bend model and the part stmcture, as well as implementation ofthe same,

are provided in Appendix K attached hereto.

A bend model viewer may be provided to inteφret the bend model and display

visual images of the part in 2-D and/or 3-D space representation. Fig. 18 illustrates a

block diagram of the stmcture of the bend model viewer and its relation to the bend

model, in accordance with another aspect of the present invention. The bend model

viewer may be implemented through object oriented programming techniques and may

be a Windows based application that permits users at the station modules of the various

locations 10, 12. 14...20 in the facility 38 to display various views of the part based on the information provided in the bend model. The bend model viewer may comprise a set

of application library modules that are used to visualize the sheet metal part. Further, the

bend model viewer may be designed as a base view class ofthe Windows application so

that it can be used as a base view class for any Windows application. Most of the

standard operations to view the 2-D and 3-D models (e.g., zoom 92, rotate 96, pan 100,

dimension 102, etc.) may be implemented as member functions of the bend model

viewer. Geometric transformations and fundamental computer graphics techniques may

be applied to the bend model objects when performing viewing operations. In addition,

the bend model viewer may comprise view model attributes 88, that comprise, for

example, four major view modes including a solid view, a wire frame view, a 2-D flat

view and an orthographic view.

According to an aspect of the invention, the bend model class library 80 may includea set of procedures or functions that act upon sheet metal parts depending upon

the selected view (e.g., solid, wire, 2-D flat or orthographic view). The bend model

viewer view class 84 may comprise a series of standard operations, such as zoom 92,

rotation 96, pan 100 and dimension 102; and, depending upon the state ofthe bend model

viewer, the bend model viewer view class may call functions from the bend model class

library 80. As shown in Fig. 18, the various view model attributes or features 88 that

may be selected by a user may include a solid view, a wire frame view, a 2-D flat view

and an orthographic view. A brief description of these various view modes that may be

provided in the present invention is provided below with reference to Figs. 19-22.

Fundamental computer graphics and geometric modeling techniques, such as

geometric transformations and 3-D geometry techniques, may be utilized to implement

the various features of the bend model viewer and to provided the different viewing

modes and functions. Recent advancements and developments in computer-based 2-D and 3-D modeling and simulation, such as the availability of graphics libraries or

packages, may be applied to implement these features of the present invention. In

addition, a wide variety of publications and material are available regarding computer

graphics and modeling. See, for example, MORTENSON, FOLEY et al., and

MANTYLA, each of which is referenced above.

In order to provide the various viewing and modeling features of the present

invention, each station module and the server module may be provided with a high

resolution display screen, such as a SVGA screen with 800 x 600 resolution. A joystick

and game card may also be provided at the station module and server module to permit

the user to selectively modify and view the different 2-D and 3-D representations of the part. Software based graphics packages, such as OpenGL and RenderWare, may be used

to provide graphical computations. Such graphics librariesor packages may be Windows

based applications and can be used to render the various viewing modes. OpenGL, for

example, may be used to render the various 2-D wire frame views based on the part

geometry and topology data provided in the bend model. Further, Renderware may be

utilized to render the various 2-D and 3-D solid views of the sheet metal part based on

the part data provided in the bend model. For more information on OpenGL, see. for example, the OpenGL Reference Manual and the OpenGL Programming Guide. Release

1 , OpenGL Architecture Review Board, Addison-Wesley Publishing Company, Reading,

Massachusetts (1992). For information on RenderWare, see, for example, the

RenderWare API Reference Manual. V2.0, Criterion Software Ltd., United Kingdom

(1996).

In order to render the various views ofthe part, the bend model may be accessed

from the database 30 by, for example, the station module of the operator. The bend

model data may then be reformatted in accordance with the data format utilized by the

graphics library or package (e.g., OpenGL or RenderWare) that is utilized. Thereafter,

the graphics data may be processed in accordance with various programmed routines in

order to render the viewing mode (wire, solid, etc.) selected by the operator or perform

the viewing function (zoom, pan, etc.) executed by the user.

When a particular view mode is selected by an operator, the selected viewing

mode is detected along with the current zoom ratio or factor and orientation of the view.

This information is then used to make function calls to the graphics package to update the current display. Function calls to the graphics package may be made in accordance with the viewing mode to be rendered, as well as the zoom or other viewing function to

be performed. Based on these function calls, the graphics package provides the necessary

data so that the station module may render the view ofthe part to the operator. Based on

the modifications to the 2-D or 3-D representation by the user (e.g., by moving a joystick

or a mouse), additional function calls may be made to the graphics library to update the

rendered image.

To provide the wire frame views of the part, the line entity data ofthe part may

be provided to the graphics package to perform the necessary graphical computations.

With solid views, however, one or more polygons should be derived for each ofthe faces

and provided as input to the graphics package to render the view. Graphics packages

such as OpenGL and RenderWare will take as input polygonal data and fill in the areas

defined by the polygons to provide a solid image. Polygons may be derived from the

face and bendline information in the bend model and by determining the boundaries of

each face. Polygons should be created to present and define each face ofthe part. The

faces may then be connected, based on the part topology and other data in the bend

model, in order to render the overall sheet metal part. If a face contains an opening or

hole, then it will be necessary to define the face with several polygons that do not

encompass such openings. For orthographic views, data for each ofthe individual views

(which may be wire frame or solid) may be sent to the graphics package and the resultant

views combined on a single display screen, such as that further shown in Fig. 22.

An exemplary code for implementing the various viewing modes and functions

ofthe bend model view is provided in Appendix E. The sample code is written in C++

and includes comments relating to the processes and operations performed therein. The code in combination with an appropriate graphics package (such as OpenGL and

RenderWare) may not only be used to render the different views (e.g., 2-D and 3-D wire

frame or solid), but also provide the functionality of each of the viewing functions (e.g.,

zoom, rotate, pan, etc.). A brief description of the various viewing mode display screens

that may be rendered is provided below.

The solid view mode displays a solid rendered 3-D view of the part defined by

the bend model. Fig. 19 illustrates an exemplary solid view window that may be

provided as output to a display screen provided at any of the locations 10, 12, 14...20 within the sheet metal facility 38. In the solid view mode, the user or operator may be

provided with a plurality of viewing functions for manipulating 3-D space navigationand

3-D automatic dimensioning. The basic functions that may be provided for altering the

solid view of the part may include rotation, zooming, panning, and/or standard view

selections. The standard views that may be provided and selected by the user may

include the following: isometric, top, bottom, front, back, left and right. An automatic

or manual dimension operation may also be provided to display the critical dimensions

of the part based on the current viewing angle. An exemplary embodiment of the

dimension feature ofthe present invention is provided below with reference to Figs. 23-

27.

As shown in Fig. 19, the solid view window may be a Windows based

application; and, thus, multiple windows or sectional views ofthe part may be provided.

The multiple view windows may include a worm view that provides a very close-up

isolated view within a window, and a bird's eye view which provides a very far away

view of the part in an isolated window. The sectional view may provide a partial view of the object as selected by the user. In order to control the various viewing functions,

a joystick interface may be provided at the server module 32 and the station modules of

each of the locations 10, 12, 14...20. Operation of the joystick alone and/or in

combination with operation of certain keys on the keyboard (e.g., a shift key or control

key) may be performed by the user in order to carry out the various functions such as

rotating, panning and zooming. In addition, the displayed texture ofthe solid view ofthe

part may be selected in order to simulate the material specified for the part in the

database. For this purpose, a material texture library may be provided that comprises a

library of material textures, such as steel, stainless steel, aluminum, etc. The stored

material texture library may be accessed and applied by an operator when a solid view

is present, so that the surface ofthe displayed part will more closely simulate the actual

texture ofthe sheet metal part.

The wire frame view mode may provide a Windows based display of a wire frame view of the sheet metal part. An example of a wire frame window is shown in Fig. 20.

The key functions for providing 3-D space navigation and 3-D dimensioning in the wire

frame view may be similar to that described above with respect to the solid view. For

example, such functions as rotation, zooming, panning, and standard view selection may

be provided. Automatic dimensioning, multiple view windows and sectional view

options may also be provided in the wire frame view mode. In addition, a joystick and/or

keyboard interface may be provided to permit the user to select and activate the various

viewing functions.

The 2-D flat view mode may display an unfolded, 2-D flat view ofthe part in wire frame representation. An example of a 2-D flat view window is provided in Fig. 21. The 2-D flat view mode may comprise a plurality of viewing functions to permit a user to

change or alter the view in the window. For example, zooming and panning functions

may be provided to permit a user to selectively zoom and pan the 2-D flat wire frame

view. In addition, dimensioning and multiple windows viewing functions may be

provided in a similar manner to that described above with respect to the solid view mode.

A joystick and/or keyboard interface may be provided to permit the user to pan, zoom

and control other viewing functions. Any special forming or shapes that are provided in

the part may be displayed as a form or shape on the outermost boundary of the formed

area with a special forming indicator or description.

An orthographic view window, such as that illustrated in Fig. 22, may also be

provided as part ofthe bend model viewer. The orthographic view mode may display the

top, front, right and isometric views of the part in wire form representation. A hidden

line option may be provided to occlude blocked lines based on the viewing angle. The

hidden line option may be utilized to simplify each view window. Various view

functions may also be provided in the orthographic view mode to permit the user to

selectively manipulate and change the present view in the window. For example,

zooming and panning functions may be provided, as well as dimensioning and multiple

windows viewing functions. As described above, a multiple windows viewing feature

may be provided to permit the user to selectively display a worm view and/or a bird's eye

view of the orthographic views in multiple windows. A joystick and or keyboard

interface may be provided at each of the locations to permit the user to selectively

activate and manipulate each ofthe viewing functions in the orthographic view mode. In addition to rendering each ofthe various view displays noted above, the bend model viewer view class may be implemented with other features. For example, the bend

model viewer may include and maintain a selection set to indicate those items or entities

in the current view that are selected or highlighted by the operator. In accordance with

an aspect ofthe present invention, an operator may be permitted to select faces, bendlines

and other features ofthe rendered part in order to modify the data relating to the selected items or to perform certain operations of those items of the part. For instance, an

operator may be permitted to select a face of the displayed part and change the

dimensional data of the face along its width or length. The operator could also be

permitted to modify the various bend data associated with each bendline, such as the

bend angle or V-width.

The bend model viewer may maintain a list of entities or items (e.g., faces,

bendlines, edges of a face or bendline, etc.) that are selected by the user. The viewer may

update the list so that the current items that are currently selected by the operator are

always maintained in the selection list. Other portions of the software in the invention

may call the view class for the current list of selected entities when performing or

executing different routines (e.g., manual dimensioning, etc.).

In addition, the bend model viewer may also provide a visibility function, which

provides visibility information and coordinate information based on the current rendered

view. As discussed more fully below, the visibility function may provide information

as to whether a particular portion or entity ofthe part is currently visible on the screen

and may also provide coordinate information as to where a screen entity is presently

located. The visibility function of the bend model viewer may be called by a

dimensioning feature ofthe invention in order to determine what portions ofthe part are currently visible on the screen, so that only dimensional information of the portions of

the part that are visible on the screen are displayed to the viewer. A more detailed

discussion of the dimensioning and visibility functions of the invention is provided

below. In addition, an exemplary code for implementing the visibility function of the

bend model viewer is provided in Appendix J attached hereto.

Referring now to Figs. 23-27, an example of a dimensioning feature will be

described in accordance with an aspect of the present invention. As described above,

each of the viewing modes may include a dimensioning function which will

automatically display the dimensions ofthe part based on the current viewing angle. An

automatic dimensioning function may be provided so that the dimension of flanges or bend lines which cannot be seen at the current viewing angle will not be displayed to the

user. When the automatic dimension function or mode is activated, only the viewable

dimensions ofthe part will be displayed based upon the current viewing angle. Further,

in the automatic dimensioning mode, only predetermined dimensions (i.e., those

dimensions which are important to the bending operation) may be displayed based on the

state ofthe current viewing angle. A manual dimension mode may also be provided, to

permit a user to selectively indicate which dimension items are to be displayed. In the

manual dimensioning mode, only those dimension items that have been selected by the

user will be displayed based on the current viewing angle ofthe part. In both dimension

modes, the displayed dimension items may be erased or removed from the window

display when the part is being zoomed or panned.

Fig. 23 illustrates an example of the various dimension items that may be

displayed in an automatic dimension mode. The dimension items that are displayed in the automatic dimension mode consist of those items that are important to the bending

operation (e.g., flange length, bend line length, bend angle, etc.) and not extraneous

dimension items, such as the dimension of a punched hole or opening. The displayed

dimension items may include, for example, the width, depth and height ofthe sheet metal

part, as well as the flange lengths. In addition, the bend line length (L), the bend angle

(A), the inside radius (R) and the bend deduction (D) of each bend line may be displayed

either alone or together in a window or group information box. As noted above, only the

visible dimension items will be displayed based on the current viewing angle. Further,

all dimensions may be erased or removed from the display when the operator is rotating,

zooming or panning to change the viewing angle of the part, and the dimensions may be

redisplayed after each operation has been completed. The size and orientation of the

display information (including any text and reference arrows) may always be sized

relative to the screen size and not to the current zooming ratio or viewing angle.

However, in order to improve the readability of the dimension information, the color,

style, weight and/or font ofthe dimension information may be configurable to permit the

user to alter the same. As a result, an operator or designer may highlight critical

dimensions in a part by selecting a particular color, font size, etc. of the dimension

information. For example, the color, size or font of dimension text, or the color, line

weight, or style of a dimension reference, line or arrow may be highlighted or selectively

altered to indicate critical dimensions within the part. An operator may also be permitted

•to color, fill or style the window information boxes, or to color particular bend lines to

also highlight other critical dimensions within the part.

Various processes and operations may be utilized for implementing the dimensioning feature ofthe present invention. Further, as noted above, the bend model

viewer may be provided with a visibility function that may provide visibility information

to the dimensioning feature of the invention. Theses functions and features may be

implemented by software at, for example, server module 32 and/or each of the station

modules located throughout the factory. Exemplary code is provided in Appendices F-I

to implement the automatic dimensioning feature of the invention. In addition, sample

code for the visibility function of bend model viewer is provided in Appendix J. The

code in these appendices are written in C++ programming language and include

comments to facilitate an understanding of the logic flow ofthe processes and operations

performed therein.

The logic flow of the dimensioning feature of the invention may be generally

categorized into three phases. During the first phase, the bend model geometry and

topology data for the part is accessed from the database 30 and is used to compute all

dimensions of the part, as well as all possible ways that the dimensions can be displayed For each bendline and face of the part, all of the extreme points at which data can be

displayed is computed, and all ways in which the dimension lines and arrows can be

displayed are computed with respect to these points. Certain heuristics may be applied

when determining where the dimension data and other information may be displayed.

For example, as a general rule, it may be determined that all information may only be

displayed on the outside ofthe part. A heuristic such as this may be applied in order to

provide a more meaningful and less crowded display of information to the viewer.

The first phase described above may be executed whenever the dimensioning

feature of the invention is activated by an operator. Altematively, the computations of the first phase may only be done once when the part is initially viewed by the user. In

such a case, the computed data may be stored in memory for subsequent use and

modified when the dimensions or other geometry data ofthe part is modified or changed

by the user. In addition, all of the computations of the first phase may be performed

relative to the geometry of the part and not the view screen, so that the data can be reused

every time regardless ofthe current view or if the view is changed.

A second phase of the automatic dimensioning feature of the invention may be performed whenever the view ofthe part is updated. The main objective of the second

phase is to filter the data developed during the first phase based on what entities of the

part are visible in the changed view. During the second phase, all data that is not visible

in the current view is filtered out so that only the data computed in the first phase that is

presently visible remains. A function call to the bend model viewer may be performed in order to determine what points or portions of the part are presently visible. As noted

above, the bend model viewer may include a visibility function that maintains and

provides information on the visible portions ofthe part based on the present view that is

displayed. Based on the orientation of the part, the bend model viewer may determine

which faces and bendlines ofthe part (as well as what edges or portions of such faces and

bendlines) are visible and which are hidden on the screen.

As noted above, sample code for implementing the visibility function ofthe bend

model viewer is provided in Appendix J. In order to determine what points or portions

of the part are visible, the bend model viewer may determine and maintain the current

view orientation ofthe part and the current zoom aspect or ratio ofthe rendered part. The bend model viewer may use conventional perspective projection techniques (see, e.g., Chapter 12 of MORTENSON) to determine and maintain the current view orientation.

When determining the visibility of any point on the part, the visibility function may first

get the world coordinates (i.e., the coordinates in which the part is represented) of the

point. Then, the screen coordinates (i.e., the pixel location on the screen) corresponding

to the world coordinates are determined for that point based on the current view

orientation and zoom aspect or ratio. Thereafter, based on the screen coordinates it is

determined whether any entity or portion of the part is in front of the point of interest

from the perspective of the screen viewpoint. The hidden nature of a point on the part

may be determined based on whether some other entity or portion of the part is assigned

the same screen point as the point of interest. A function call to a graphics package or

library (such as OpenGl or RenderWare) may be made to determine whether more than

one point ofthe part is assigned to the same screen point. If something is assigned to the

same screen point, then it may be determined if the point ofthe part is behind it based on

the respective Z-buffer depths of the points. The Z-buffer depth is used by graphics

packages, such as OpenGL and RenderWare, to define the distance to each point from

the viewpoint or camera position. The Z-depth may be determined by making a function

call to the graphics package with the points of the part that are in interest.

The above-describedprocessedof the visibility function of the bend model viewer

may be executed whenever there is prompt to the bend model viewer from the auto-

dimensioning feature of the invention. Such processes may, thus, be performed

whenever the current view of the rendered part is modified or altered by the operator. As

discussed above, the bend model viewer may maintain and update the status of the

current view orientation and zoom ratio whenever there is a change made to the orientation of displayed image so as to accurately provide visibility information when

required.

After determining which data is visible, the auto-dimensioning function may

determine (e.g., based on the computations from the first phase) every possible way and

locations that the dimension data and other information may be displayed. A set of

heuristics may be applied to select the best way to display the data from the available

ways that the data may be displayed. For example, a first heuristic may require that the

area of the screen that is closer to the viewpoint of the viewer is preferred. A second

heuristic may define that the data is to be displayed in an area that is closer to the area

where the distance between the possible points defining the dimension is smallest. Other

heuristics may also be applied based on the relative position of other dimension data and

other information to prevent overlapping and crowding on the screen.

After determining the visible portions ofthe part and the best areas to display the information for the visible areas, a third phase ofthe auto-dimensioning function may be

executed to draw the various information on the display screen. For example, based on

the selection of areas to display the information, the dimension information may be

displayed on the screen for each ofthe visible dimensions of the part. Further, based on

which bendlines are visible, bendline information may also be displayed in information

boxes (such as that shown in Fig.23) in areas ofthe screen that do not overlap with other

part information. The part dimensions, including the width, height and depth ofthe part,

may also be displayed on the screen at a predetermined location (e.g., to the lower right

ofthe part) or a location that is closest to that part and that will not overlap or block other information. Figs. 24-27 illustrate the various methods and definitions that may be used when

displaying the dimension items in the dimensioning mode. In particular. Figs. 24A, 24B

and 24C illustrate the manner in which the flange length may be defined for various

different parts. In accordance with an aspect ofthe present invention, the flange length

may be defined as the farthest point on the flange from each bend line. If the farthest

point ofthe flange does not exist on the longest edge ofthe flange which is parallel to the

bend line, then the dimension of the longest flange may be added and displayed in the dimensioning mode. By way of non-limiting examples, Figs. 25A and 25B illustrate

adding an auxiliary flange length for two different types of parts.

When the thickness of a part is displayed, the flange length may be displayed as

an outside to outside dimension. For example, Figs. 26A, 26B and 26C illustrate the manner in which the flange length may be indicated for various parts that are displayed

with thickness. In addition, for parts with an acute angle bend, the flange length may be

indicated in a variety of ways. For example, as shown in Fig. 27A, the flange length may

be displayed based on a tangent dimension definition, in which the flange length is

measured from a tangent line extending from the acute angle bend. Altematively, an

intersection dimension method may be used, such as that shown in Fig. 27B to indicate

the flange length based on a point defined by the intersection of two lines extending from

both sides ofthe acute bend angle. An operator may be permitted to choose between the

tangent dimension or intersection dimension method for displaying the flange length,

and/or a particular dimension method (e.g., the tangent dimension method) may be

provided as a default setting.

In order to facilitate the development of the bending code sequence, a graphical user interface with various display functions may be provided to aid the development of

the bending plan by the operator. Figs. 28-32 illustrate the various processes and

operations that may be performed to develop the bending code sequence through the use

of a graphical user interface, in accordance with another aspect ofthe present invention.

Normally, the initial bend model data and other job information will be developed

by a design programmer through entering the critical geometric and manufacturing data

at server module 32. The resultant bend model file may then be stored in database 30.

Before the sheet metal part may be produced, it will be necessary for a bending operator

to develop a bending sequence for carrying out the required bending operations. The

bending operator must also decide what type of tooling is required and define the tooling

set-up for the bending machinery. This process of developing a bending plan may be

aided and made more efficient by use of a graphical user interface and the various

teachings of the present invention.

In order to develop the bending plan, a bending operator at, for example, bending

station 18, may access and download the bend model and other job information from

database 30. The bend model for the relevant part may be loaded or imported to the

station module on the shop floor at bending station 18 via the communications network

26. This step is generally shown at step S.220 in Fig. 28. Thereafter, at step S.224, the

bending operator may examine the shape and dimensions ofthe part by using the bend

model viewer. At this point, the bend operator may selectively zoom and pan the various

2-D and 3-D views of the part at a display screen located at the bending station. By activating the automatic dimension feature ofthe present invention, the bending operator

may also view the important bending dimensions for the part for carrying out the bending operations.

Once the operator understands the shape and dimensions of the part, the bending

operator may begin to develop the bending plan by selecting and displaying a bend

sequence input window, at step S.228. The bend sequence input window may provide

a graphical user interface to assist a bending operator to create, modify and delete a bend

sequence, and may also enable the operator to specify and enter various manufacturing

parameters (e.g., backgauge location, tooling, NC data, etc.) for each step in the bend

sequence. The bend sequence input window may contain a 2-D flat view image ofthe

part displayed on a portion of the screen (e.g., at the center of the screen or towards the

left hand side ofthe screen). The 2-D flat view image may include the various features

of the unfolded part, including the flanges, holes and openings of the part. As the

bending operator selects and indicates the bendlines and the bending sequence for each

bendline, a solid 2-D or 3-D image of the intermediate part shape at each bending step

may appear and be provided on a portion ofthe screen, such as on the right hand side of

the screen, as shown for example in Fig.29A. The images ofthe intermediate part shapes may be displayed in a sequence corresponding to the entered bend sequence, and may be

simultaneously displayed on the screen with the 2-D fl t view image of the part (such as

in the example of Fig. 29 A) or separately displayed on a different screen display.

Further, as each bendline is selected, the bendline may be highlighted and a bend

sequence number and an insert direction (e.g., represented by an arrow) may be displayed

on or near the bendline, as shown for example in Fig. 29B. The bend sequence number

for each bendline may be set automatically based on the sequence in which it has been selected, or the bend sequence number may be entered manually by an operator after each bendline has been selected. Other manufacturing information relating to the bendline,

such as the bend angle, bendline length and backgauge position, may also be entered

and/or displayed on the screen when each bendline is selected or highlighted, as shown

for example in Fig.29D and 29E. As shown in Figs.29D and 29E. dialog or information

boxes may be displayed on the screen to permit a bending operator to select, enter or

modify the manufacturing information and other parameters related to each bendline.

The dialog or information box may also permit a bending operator to highlight or select

a bendline, and hot function keys or fast switching keys may be displayed in the bend

sequence input window to permit the bending operator to select or enter tooling and view and modify NC data. For example, a bending operator may select a Tool function key

to switch from the bend sequence input window to a tool input display screen or display

screens to enter tooling information. An NC function control key (e.g., NC9 Ex) may

also be provided to permit an operator to view and or modify NC data related to the

bending operations to be performed.

Further, as shown in Figs. 29D and 29E, other function keys and controls may

also be provided related to defining and/or modifying the bendlines and the related

manufacturing information. For example, a Zoom All key may be provided to zoom in

and zoom out ofthe 2-D flat view image; a Backgauge key may be provided to select or

set a position for the backgauge; a Group and Ungroup control key may be displayed to

permit and control the bendlines that are to be bent together; and a control key (e.g., Ama

Bend) may be provided to define special bend operations. Other functions keys may also

be displayed to permit the bending operator to select, modify and/or delete the bend

sequence (e.g., Remove, Clear Fwd, Clear All, OK, Cancel). With the bend sequence input window, the bending operator is able to efficiently view and modify the bend

sequence and the various manufacturing information.

In addition, according to another aspect of the invention, cross-sectional views

of the part and/or a bending simulation of the part may be displayed on the screen for

each bending step in the bend sequence (see, e.g., Fig. 29E). The cross-sectional views

and bend simulations may be selectively displayed on the screen or displayed as each

bending line is selected by a bending operator. The cross-sectional views and bend

simulations may include representations of, for example, the upper and lower bending

tools (e.g., punch and die) and/or a backgauge position or setting, and may be displayed

simultaneously on the screen with the 2-D flat view image of the part or displayed

separately on a different screen display. The backgauge position may be determined

automatically based on the topology ofthe part or may be set or modified by the operator.

If the tooling information for the bendline has not been entered or set by the bending

operator, the cross-sectional view and/or bend simulation may not be displayed on the

screen, or only the representations of the intermediate part shape and a calculated or

defined backgauge position may be displayed. The bend simulations may comprise a

displayedsimulafionof any required flipping of the part, the handling and orientation of

the part, and/or the bending of the part that is performed at each bendline. It is also

possible to simultaneously display, on the display screen with the 2-D flat view image

ofthe part, a cross-sectional view ofthe part before the bending step and a cross-sectional

view of the part after the bending step has been formed (see, for example, Fig. 29E).

These cross-sectional views may be provided on the right hand side of the screen and

may include representations of the upper and lower bending tools and the backgauge for each bending step in the bend sequence. In addition, zoom control or function keys

(Zoo ln and ZoomOut) may also be displayed to permit an operator to control the zoom

ratio or aspect related to the before-bend and after-bend cross-sectional views.

Techniques and processes similar to that disclosed in Japanese Examined Patent

Application No. HEI 7-121418 (published on December 25, 1995 in the names of NIWA

et al.) and Japanese Unexamined Patent Application No. HEI 1-309728 (published on

December 14, 1989 in the names of NAGASAWA et al.), the disclosures of which are

expressly incoφorated herein by reference in their entireties, may be utilized to display

the cross-sectional views and bend simulations ofthe part.

In accordance with an aspect ofthe invention, software or programmed logic may

also be provided to automatically determine the insertion direction for the bend by

calculating the shorter or smaller side ofthe part relative to the selected bendline. Based

on a feature ofthe invention, each bendline may be used to divide the part into two sides.

The insert direction may be determined for each bendline based on the side of the part

that has a smaller or shorter length (e.g., the dimension of the side that is peφendicuiar

to the bendline) or based on the side that has a smaller overall area. If an operator is not

satisfied with the insertion direction that was selected, the operator may flip the insertion

direction, as illustrated in Fig. 29C. The operator may change or flip the insertion

direction by, for example, clicking a select button of the mouse or keypad when the

bendline is highlighted. The insertion direction information may include an arrow and/or

text to indicate the insertion direction of the flange defined by the bendline for bending

the part with a bending apparatus or machinery. The insertion direction information may

be displayed at or near the bendline (see, for example, Figs. 29B and 29C) or at or near the end of related flange (see, for example, Fig.29D). In addition, the insertion direction

information may be displayed when each bendline is selected, or may be selectively

displayed based on input received from a joystick device, mouse device or keyboard

device.

Thus, through the use of a graphical user interface, the bending operator can see

the various intermediate shapes and the forming of the final part based on the selected

bend sequence entered by the operator. Once again, the operator may enter and select

data on the screen through an appropriate input device, such as a joystick interface, a

mouse interface and or a keyboard interface. If the bend operator is not satisfied with the

proposed bend sequence, the bend operator may edit the bend sequence before finalizing

the bending plan, as generally shown at step S.232. The editing of the bend sequence

may be carried out in a variety of ways and methods. In particular, in accordance with

an aspect of the present invention, a drag and drop editing feature may be provided to

facilitate the bending operator in modifying and editing the bend sequence. As shown

in Fig. 30, the operator may edit a selected bend sequence by simply grabbing one ofthe

intermediate part shaped icons or displays provided on the left or right hand side ofthe

screen and dropping it to its desired location in the sequence. Thereafter, based on the

bend operator's modification to the bend sequence, the various intermediate part shapes

illustrated on the screen will be modified to indicate the intermediate bend stages in

accordance with the revised bend sequence. In addition, the bend sequence numbers on

the 2-D flat view image may be revised based on the bend operator's drag and drop

editing ofthe bend sequence.

After the bending sequence has been determined, the operator decides what type of tooling should be used by selecting tools from a stored library of tooling data, as

shown at step S.236. The pertinent tooling information may be displayed to the bending

operator on the shop floor and display menus may be provided to graphically aid the

bending operator in selecting tooling from the library. Once a particular tool has been

selected from the library, the pertinent data relating to the tool may be indicated on the

screen. Fig. 31 illustrates an example ofthe various display menus and data tables that

may be graphically displayed to the bending operator for manual tool selection. In the

example of Fig.31 , successive display menus or screen displays are graphically displayed

in order to aid the bending operator in picking a particular tool from the tool library. The

successively displayed screen displays may be displayed simultaneously on the display device (e.g., in overlapping or cascading fashion), or may be individually displayed with

the screen being cleared before each successive screen display is displayed. Once a

particular tool has been selected, the particular data for that tool may be provided in a

table and displayed to the operator. The data in the tooling library may be predefined and

stored (e.g., in database 30) during an initial set-up procedure ofthe software.

The manual tool selection feature of the present invention may permit a user to

select a tool type and the shape ofthe tool in each type. For example, various tool types

may be selected, including punch, die, die holder, and die rail. Each type may consist of

many shapes, and for each shape there may be many tools with different sizes and

dimensions. To select a tool, a user would first pick one tool type by selecting one icon

from the tool type icons that are displayed, such as that illustrated in Fig. 31. Thereafter,

the user would be provided with a menu of different shapes that are available for the selected tool. After analyzing the tool shapes, the user may select a tool shape by selecting one ofthe shape icons from the displayed shape icons for the selected tool (e.g.,

in Fig. 31 a goose neck shape punch has been selected). Finally, the user may select the

appropriate size and dimension for the tool shape that has been selected. As further

shown in Fig. 31 , a table may be displayed to the user to indicate the different sizes and

dimensions of tools available for the tool shape that was selected. By selecting an item

from the table, the selected tool may be displayed as an icon to replace the generic tool

type icon and to confirm the selection ofthe tool.

At step S.240, the bend operator may then set-up the various tool stages in the

press brake by aid of a graphical interface. Fig. 32 illustrates an exemplary tool set-up

window that may be provided to the bending operator to facilitate definition ofthe tool

set-up to be used in the bending plan. Various punch, die and rail data may be indicated

in the tool set-up window, as shown for example in Fig. 32. The tool and die information

for the sheet metal part may be entered by the operator. A joystick may be provided at

the bending operator's station module to permit the bending operator to indicate the

tooling location and to select the tool and dies from a list of available tools and dies. In

the tool set-up window, the left hand side of the screen may display the profile of the

current tool set-up and the right hand side of the screen may display the location of the

current set-up in the press brake. The current set-up location may be highlighted or

shaded, as shown in Fig. 32. Finally, once the bend operator is satisfied with the bend sequence, the bending

plan information including the tooling and bend sequence information, may be saved

with the bend model in the database 30, as generally shown at step S.242 in Fig. 28.

Actual testing ofthe bend sequence may then be performed with the press brake in order to verify the bend sequence selected by the bend operator. If required, any further

modifications to the tooling definitions or bend sequence may be performed by the

operator or designer at the station module.

Various other features of the invention may be provided to aid the bending

operator in the development of the bending plan. For example, in accordance with

another aspect ofthe present invention, a tooling expert may be provided to automatically

provide to the bending operator suggested tooling and bend sequences based on the part

geometry and other information stored in the bend model. The suggestions from the

tooling expert may be followed or revised by the bending operator after analyzing the

same. In addition, a more complex tooling expert system may be provided to make

tooling suggestions and bend sequence suggestions as to more complex operations based on the geometry ofthe part in the bend file and a profile analysis ofthe tooling to check

for potential collisions and interference. Such expert systems may be used and

implemented with either manual or robot assisted bending machinery. By way of non-

limiting examples, the present invention may be implemented with the features and

teachings disclosed in pending U.S. patent application Serial No. 08/386,369, entitled

"Intelligent System for Generating and Executing a Sheet Metal Bending Plan." in the

names of David A. BOURNE et al., and U.S. patent application Serial No. 08/338,115,

entitled "Method for Planning/Controlling Robot Motion," in the names of David A.

BOURNE et al.. the disclosures of which are expressly incoφorated herein by reference

in their entireties.

As described above, a graphical user interface and various functions may be provided to facilitate the bend operator when developing the bending plan for a sheet metal part. In accordance with another aspect ofthe present invention, additional features

may also be provided to aid in the design and manufacturing of the part. As described

more fully below, various multimedia features may be implemented into the present

invention, such as the storage of audio and visual information, to provide additional

assistance to the bending operator when developing the bending plan or executing a bend

sequence. Further, a collision check feature may be provided that automatically checks

for potential interference and collision between the tools and the parts at each of the

intermediate bending stages. This collision check feature may be provided to replace the

cumbersome and time consuming manual check of the tool profiles and spacing in the

part, which is customarily performed by bending operators when developing a bending

plan. These features and others will now be described with reference to the

accompanying drawings.

In accordance with an aspect ofthe present invention, a method may be provided

for storing audio and video information with the bend model data. Various audio and

video instructions can be recorded at the shop floor to provide special instmctions with

respect to, for example, the handling and bending of the sheet metal part. For this

puφose, a CCD camera or digital camera may be provided at each ofthe station modules

of the various locations 10, 12, 14...20, along with an audio microphone. Other

equipment, such as a video camera with an audio microphone, may be provided at the

station modules to permit the operator or user to record audio and video information. The various recording equipment may be connected to a station module computer at the shop

floor. By way of a non-limiting example, an Intel PROSHARE personal conferencing

CCD camera (available from Intel Coφoration) may be used for recording audio and video information. Other commercially available CCD cameras or digital cameras may

also be utilized to record such information.

The various audio and video information that is stored with the bend model data

may be accessed and retrieved by the user according to various methods and procedures.

For example, menu options may be displayed by the station module to permit the replay

of stored audio and video information. In addition, in accordance with a preferred

embodiment ofthe present invention, an operator may be provided the ability to attach

and associate the stored audio and video information with the various display screens and

views ofthe part by selecting and creating icons that are displayed in the view window.

This feature may be implemented through software and object oriented programming

techniques, whereby an icon object is created and stored within the bend model data

structure. The icon object may contain procedures for retrieving attached audio and

video information from memory based on certain conditions (e.g., the selection of the

icon by the operator by double clicking a mouse or indicating selection by use of a

joystick or other input means). With the icon feature of the present invention, the

operator may associate different audio and video messages or information with different

parts ofthe sheet metal part and to any display. By incoφorating the icon into the part

representation, the icons may be adapted to zoom, rotate and translate with the 2-D and/or

3-D model displays ofthe part as the view is changed on the screen.

Fig.33A illustrates an example of attaching audio and visual information through

the use of icons pasted to a 3-D solid model ofthe part. After the user has recorded the audio and visual information, the operator may paste an icon at any location within the 3-D model window. When the icon is subsequently selected by an operator or user, the stored audio and visual information may be played back and displayed in the window to

provide any special instmctions or messages concerning the part or the area of the part

to which the icon was placed. Other information, such as the simulation or recording of

the bending motion, may be associated with the part by placing icons in proximity to the

various bend lines of the part. The video information concerning the bending motion

may then be played back to the user when the icon is selected.

The operator or user may record both audio and video information, or only record

a simple audio message or a still or motion video signal, that may be selectively played back to the user. The icons that are attached to the window display may graphically

indicate the type of information that is stored (e.g., a microphone icon may be depicted

to indicate that audio information has been stored or a display monitor icon may be

provided to indicate that video information has been stored). Special icons may also be provided to indicate that both audio and visual information are associated with that icon

(e.g., an " A/V" symbol or a video camera icon that includes a microphone). A directory

of icons may be provided and displayed to allow the user to select among the various

icons when attaching audio and/or video information to the screen display or image.

Fig. 33B illustrates another example of a display window that may be

incoφorated with icons for retrieving stored audio and video information. The display

window depicted in Fig. 33B relates to the tool set-up screen image, such as that

describedabove with reference to Fig. 30. In the example of Fig. 33B, audio information

may be stored and retrieved by a microphone icon and separate video information may

be stored and retrieved by pasting a video icon to the display window. The audio and video information may relate to special instmctions or information regarding tool set-up or operation. Further, regardless of the type of window display that is currently active,

the operator may paste as many icons as required to the various areas in the window

display so that different audio and video information may be later retrieved.

In accordance with another aspect of the invention, an image editing window

feature may be provided to facilitate the operator in selecting stored images and applying

them to different screens. The imaging editing window feature may be provided as a

Windows based application that may be accessed at, for example, server module 32 or

any of the station modules provided throughout the manufacturing facility. Fig. 34

illustrates an example of the image editing window that may be implemented in

accordance with the teachings of the present invention. The images displayed in the

image editing window may include images shot by a digital camera or a CAM coder. The images that are displayed on the screen may be selectively chosen by the operator

(e.g., through a mouse or other appropriate data input means) and copied to different

screens so that they may be associated with particular model views of a part. The

operator may then paste the image or an icon to the model window (e.g., a 3-D solid model window of the part, such as that shown above with respect to Fig. 33B). The

images shown in Figs. 33 and 34 are photocopy reproductions of actual screen images;

the actual video images will themselves be clearer, dependent upon the resolution ofthe

camera and screen used. The images can include, e.g., a still or motion video image of

a bending operator discussing or illustratingspecial handling or other instmctions relating

to a bending operation, or may be a video image of a sheet metal bending operation. In

other words, any actual image which may be useful can be taken and later displayed. Thus, the actual images shown in Figs. 33-34 are for illustrative puφoses only. Referrin now to Figs. 35 A and 35B, an example of a collision check function of

the present invention will be provided. In accordance with an aspect of the present

invention, a collision check function may be provided to allow users to check for

potential collisions between the part and punch tool through use of a graphical user

interface. The collision check function may be a Windows based application that may

be accessed at any station module or location within the manufacturing facility. The

automatic collision check function of the present invention may be used by a bending

operator in place of the traditional and cumbersome manual profile check that is

customarily performed when developing the bending plan. Traditionally, when developing a bending plan for a sheet metal part, a bending

operator will first determine the bend sequence for the part. The bend sequence defines

the order and manner in which the sheet metal part is to be bent during production. After

the bend sequence has been determined, the bending operator will select and define the

tooling to be used to carry out each of the bending operations. During this process, the

profile of the tools selected and the intermediate shapes of the part will be analyzed to

ensure that there is no interference or collision(s) between the tools and the part when

carrying out each ofthe bending steps. If a collision or interference is detected, then the

type of tooling selected (or, if necessary, the bend sequence) will have to be modified so

that the bending operations may be carried out without interference or collision between

the tool(s) and the sheet metal part.

When detecting for potential collisions or interference, bending operators have

traditionally relied upon manual methods to analyze the clearance between the profile of

a tool and the bent parts or shapes of the sheet metal component. Typically, a model of the tool profile is constmcted and used by a bending operator. The tool profile model is

manually matched or overlaid with engineering or technical drawings (having the same

scale size as the tool profile model) ofthe various intermediate shapes ofthe sheet metal.

By fitting and matching the tool profile model with the drawings ofthe part, the bending

operator can determine whether there is sufficient space and clearance between the tool

and the part at each ofthe bending steps. This process, however, tends to be cumbersome

and time consuming.

The present invention overcomes the disadvantages of such traditional methods

by providing an automatic collision check function. The collision check function of the

present invention may be implemented through a graphical user interface to permit the

bending operator to check for collisions at each intermediate step within a defined

bending sequence. Figs.35A and 35B illustrate examples ofthe collision check function

implemented through a graphical user interface. When activated, the collision check

function will automatically check for collisions between each intermediate shape ofthe part within the bending sequence and the punch tool or tools defined for that sequence.

The intermediate shapes may be displayed on the screen (see, for example, Figs.35A and

35B) and if a collision is found, the step at which the collision is detected may be

highlighted on the screen. In addition, other display indications, such as text, may be

provided to indicate the number of collisions detected. In the examples of Figs.35 A and

35B, the collision information is provided in the upper right hand area of the display

window. In addition, the type of punch tool or tools for which the collision has been

checked may be displayed or indicated in the upper left hand area ofthe display window.

If a collision is detected for the punch tool selected by the operator, the intermediate shapes or stages at which the collision is detected may be highlighted on the

screen. In this case, an operator may select another punch tool for that particular bending

stage and the collision check function may be reexecuted to determine if a collision will

occur with the second choice for the punch tool. The operator may select a punch tool

for each bend and check for collisions with the collision check function. Drag and drop

editing may be provided to allow the user to change the bend sequence displayed in the

window display by dragging the intermediate bending shapes and dropping it to a desired

position within the proposed bend sequence. The bend sequence may then be modified

based on the drag and drop editing performed by the operator, in a similar manner to that

described above with reference to Fig. 32.

Various processes and operations may be used to implement the collision check

feature ofthe present invention. For example, in order to detect for a potential collision,

the geometry for the selected tool and the geometry for the part at each intermediate

shape may be accessed. The geometrical data relating to the part at each intermediate

step may be generated based on the bend sequence and the part dimensions and topology

data. Each flange ofthe part may be folded in accordance with the bend data (e.g., bend

angle, bendline location, deduction amount, etc.) in order to render the part at each

intermediate stage in the bending sequence. The above described folding process and

deduction compensation features of the invention may then be applied when generating

the geometrical data for the part at each intermediate stage. With the tool and part

geometry, the tool and part may be oriented with respect to one another by placing the

tip ofthe tool at the bendline ofthe part at each ofthe bending stages. A collision may

be detected by analyzing the geometrical data and boundaries ofthe tool and the part and determining whether there are common points or overlapping points in the tool and the

part. When a collision is detected at a particular bending step, that step may be

highlighted on the screen to indicate the detection of a collision to the user.

The tool data that is used to detect for collisions may be actively taken out of a

tool shape library based on the tooling selection(s) made by the user. Recalculation of

a collision at any intermediate bending step may be performed based on a different tool shape or modification of the bending sequence. By providing such features, and

displaying such information through a graphical user interface, such as that described

herein, the potential for collisions may more easily be determined and corrected by the

bending operator.

As described above, a joystick or mouse device may be provided at each ofthe

station modules and locations throughout the manufacturing facility in order to permit

the user to selectively activate and control various viewing functions (e.g., zoom, pan,

rotate, etc.) when viewing the rendered model of the sheet metal part. The joystick

device may be a multi-axisjoystick and include select and control buttons. The joystick

may be implemented through various commercially available joystick devices, including

a Microsoft Side Winder joystick, and may be plugged into a game port ofthe computer

of each ofthe station modules and/or other locations in the facility. The mouse may also

be implemented by any commercially available mouse support software, such as

Windows 95 or Windows NT, and any commercially available mouse device that is

plugged into a game or mouse port ofthe computer at each ofthe facility locations

By way of non-limiting examples, Figs. 36-41 illustrate various aspects of a system for manipulating 3-D geometrical shapes and renderings ofthe part by using a joystick or mouse device. The 3-D navigation system of the invention permits a user to

control various viewing functions, such as rotation, zooming and panning. In accordance

with an aspect ofthe present invention, the system may also use a dynamic rotation axis

that is calculated based on the current zoom view when viewing the 3-D model.

According to this aspect, the center of rotation may be dynamically changed and

calculated based on the current view and zoom ratio or factor so that the zoomed area of

the part will not disappear from the screen when the part is rotated with, for example, a

high zoom ratio or factor.

In accordance with an aspect ofthe present invention, the 3-D manipulation and

navigation system may be provided at the station modules and/or server module of the

facility. The processes and operations ofthe 3-D navigation system may be implemented

through software or programmed logic and by using any one of a wide variety of

programming languages and techniques. For example, the system may be implemented

by using a high level programming language such as C++ and using objected oriented

programmingtechniques. Further, by way of a non-limiting example, VISUAL C++ may be utilized, which is a version ofthe C++ programming language provided by Microsoft

Coφoration for Windows based applications. The viewing functions (e.g., zoom, rotate,

pan, etc.) may be defined and implemented as member functions of the view class of the

bend model viewer ofthe present invention described above (see, e.g., Fig. 18 and the

related disclosure provided above). Information conceming the current zoom factor and

position ofthe part (e.g., the position ofthe part in 3-D space) may also be accessed from

the bend model viewer to calculate the dynamic rotation axis and to provide the desired

viewing functions. Various hardware components and interfaces may also be provided in order to

implement the 3-D navigation system ofthe present invention. For example, the software

used to implement the system may be provided or reside in the computer or personal

computer of the station modules and server module. As discussed above, the computer

or personal computer may include a graphics card and a display screen or terminal, such

as a high resolution monitor, to display 3-D renderings ofthe sheet metal part to the user.

The computer or personal computer may also include a mouse or game port adapter card

to connect and interface with the mouse or joystick device. Commercially available

software may also be provided to inteφret the command signals received by the mouse or game adapter card from the user-operated mouse or joystick device.

Figs. 36A and 36B illustrate examples of rotation functions that may performed

with a multi-axisjoystick 112 to rotate, for example, a simple 3-D box shaped part. As

noted above, a joystick may be provided and connected to a computer or equipment

provided at the station modules and/or server module provided throughout the facility.

As shown in Figs. 36A and 36B, rotation of the part may be achieved by moving the

joystick 112 forwardly or rearwardly, and to the left and the right. The orientation or

direction ofthe rotation axis may be set based on the movement of the joystick 1 12 (or

mouse). For example, moving the joystick 112 forwardly or rearwardly may cause the

part to rotate in the clockwise or counter clockwise direction about a rotation axis defined

along the X-coordinate axis (see, e.g., Fig. 36A). Further, moving the joystick 112,

however, to the left or right may cause the part to rotate in the clockwise or counter

clockwise direction about a rotation axis defined along the Y-coordinate axis (see, e.g., Fig. 36B). When the zoom ratio or factor ofthe current view is low and the entire rendering

ofthe part is provided on the screen, the rotation axis may be defined as passing through

the geometric center or centroid ofthe part. As described above, the zoom factor and the

visibility of the part on the screen may be determined based on the visibility function

provided by the bend model viewer ofthe present invention. When it is determined that

the entire part is displayed in the screen (such as that in Figs. 36A and 36B), then

coordinate geometry techniques may be used to define and set the rotation axis to the

geometric center ofthe part. Rotation ofthe part may then be performed based on the

user-defined movement ofthe joystick device and through the rotate member, viewing

function of the bend model viewer of the present invention.

If, however, only part ofthe object is displayed on the screen and it is determined

that portions of the part are not visible (e.g., a high zoom factor or ratio has been

selected), the rotation axis should not be maintained at the geometric center or centroid

ofthe part, since to do so may cause the zoomed portion ofthe part to disappear from the

screen during rotation. Instead, according to the invention, when the zoom ratio is

increased, the rotation axis is dynamically recalculated such that the rotation axis passes

through the coordinate ofthe closest point to the viewpoint (or camera view) at the center

of the screen. By dynamically recalculating the rotation axis based on changes in the

zoom factor, the part may rotated about an axis that does not cause the visible portion of

the part to go out of view during rotation.

In order to perform zooming and panning of the 3-D model, additional control

buttons may be provided on a keypad that is provided separately or with the joystick or mouse device. For example, by pressing a zoom button 1 14, and moving the joystick 112 forwardly or rearwardly, the part may be zoomed in or out based on a predetermined rate,

as shown in Fig. 37. As discussed above, the rotation axis may be recalculated within

each zoom window to permit the user to view the zoomed portion of the part when

rotation is performed. In addition, panning of the 3-D shape may be controlled by the

user by pressing or activating a pan button 1 16 and moving the joystick 112, as shown

in Fig.38. As with the zoom button 1 14, the pan button 1 16 may be provided on a digital

input pad that is provided separately or with the joystick or mouse device at each ofthe

various locations ofthe facility.

In accordance with an exemplary embodiment of the invention, the various

processes and operations that may be provided to implement the 3-D navigation and

manipulation system will be described below with reference to Figs. 39-41. As indicated

above, the necessary processes and operations of the 3-D navigation system may be

implemented through a combination of software or programmed logic, and hardware

components and interfaces. Input signals from a user-controlled device, such as a

joystick or mouse device, may be inteφreted to determine the amount of movement and

reorientation of the rendered part that is desired. In accordance with the invention, the

rotation axis of the rendered part may be calculated dynamically, based on the current

view and zoom factor, in order to prevent the zoomed area ofthe part from disappearing

from the screen during rotation.

When updating the current view ofthe rendered part, input signals are received

from the user based on the manipulation ofthe joystick or mouse device, as generally

indicated at step S.301 in Fig. 39. Movement of the joystick or mouse device by the user in a particular direction, and/or in combination with the activation of special control buttons, may cause certain viewing functions (e.g., rotate, zoom, pan, etc.) and movement

of the rendered part in predetermined directions (e.g., clockwise or counterclockwise;

zoom-in or zoom-out; left or right; etc.) to be effectuated, as described for examples in

Figs.36-38. The received signals, whether they are from a joystick or mouse device, may

be mapped to cursor movement to define the amount of movement on the screen that is

desired by the user. If the user is not within one ofthe viewing function modes (e.g., the

user is selecting information on the screen or reviewing information in a dialog box or

window), then mapping of the received signals may not be required.

As will be appreciated by those skilled in the art, the signals that are received

from conventional joystick or mouse devices are based on different coordinate or

reference systems than that ofthe screen space, and must be translated in order to provide

meaningful information regarding cursor movement on the screen. Therefore, after

receiving the input signals from the user, the received signals may be mapped to cursor

movement, as indicated at step S.303, before calculating the rotation axis and updating

the current view of the rendered part.

Different methods and processes may be used to translate and map the input

signals from the user-controlled device to cursor movements in screen space.

Traditionally, movements of a mouse device have been translated and mapped to cursor

movements by commercially available software. For example, Windows 95 and

Windows NT include software routines for translating mouse movements to cursor

movements. As such, movements of a mouse device may be mapped to cursor movement

by using such commercially available software. If, however, the user is provided with

a joystick interface, then the joystick movements should also be translated and mapped to cursor movements in order to provide useful information. Various methods and

techniques may be used to map the joystick movements within the joystick virtual space

to cursor movements within the screen space. For example, the joystick movement

signals may be first processed and translated to mouse movements before they are finally

mapped to cursor movements. Altematively, the joystick signals may be directly mapped

to cursor movements as a function ofthe ratio ofthe screen space size to the size ofthe

joystick virtual space.

Fig. 40 illustrates an example of mapping joystick movements to cursor

movements in screen space, in accordance with an aspect ofthe present invention. As

indicated above, a joystick device may include its own virtual coordinate system or space 218. The joystick virtual space 218 includes an origin J 1 that corresponds to the position

at which the joystick is in a center or neutral position (i.e., a position at which the

joystick is not moved). When the joystick is moved to a new position (e.g., a current

position J2 as shown in Fig. 40), the joystick device generates signals to indicate the new

or current position within the virtual space of the joystick. Since the joystick virtual

space 218 is often larger (in terms of pixels) than the screen space 212, the virtual

coordinates and movements of the joystick must be translated to screen coordinates in

order to determine the desired cursor movements and, thus, movement ofthe part on the

screen.

Various methods and processes may be utilized to map and translate the virtual

coordinate movements of the joystick to screen coordinate movements. For example,

joystick movements may be mapped to screen cursor movements based on the ratio of

the screen space size to the joystick virtual space size. More particularly, when it is determined that a viewing function mode (e.g., zoom, rotate, pan, etc.) is active and the

joystick device has been manipulated by the user, the actual movement ofthe cursor from

a previous point Cl to current point C2 may be determined by the following equation:

current_point = previous_point + (scale factor x V); wherein, "current_point" is the

current point C2 of the cursor; "previous_point" is the previous point Cl ofthe cursor;

"scale_factor" is the ratio of the screen size to the joystick virtual space size (both in

pixels); and "V" is a vector representing the movement and direction of the joystick from

the joystick origin JI to the joystick current position J2. Thus, in order to map joystick

movements to cursor movements, a vector "V" indicating the direction and movement

of the joystick from the origin J 1 to the current position J2 may first be calculated based

on the signals received from the joystick device when it is manipulated by a user. After the vector "V" has been calculated, the joystick movement may be mapped to cursor

movement by using the vector "V" amount and the "scale f actor" amount in the equation

describedabove; that is, the new or current position C2 ofthe cursor may be calculated

by multiplying the vector "V" by the ratio of the screen size to the joystick space size

(i.e., the "scale factor"), and then adding the result of this computation to the previous

cursor position Cl .

Depending on the scale factor, it may be necessary to increase or decrease the rate

of scaling or movement by a predetermined or user selected adjustment factor. In such

a case, and depending on the preference ofthe user, the scale factor may be multiplied

by the adjustment factor when calculating the current point of the cursor in order to

increase or decrease the rate of scaling. For example, if the ratio of the screen size to the

joystick space size provides a scale factor of 1/64, then it may be preferable to increase the rate of scaling in order to provide a more satisfactory relationship between

movements of the joy stick and the rate of movement ofthe rendered part on the screen.

By way of a non-limiting example, with a scale factor of 1/64, an adjustment factor of 3

may be used when zooming or rotating the rendered part. Further, for a scale factor of

1/64, an adjustment factor of 6 may be used when panning of the rendered part is

performed. Of course, the rate of scaling may be modified based on the particular needs

ofthe user, and the adjustment factor may be predetermined or the user may be given the

option to adjust or select the adjustment factor to modify the rate of scaling. Further, as

indicated in the example discussed above, the adjustment factor may be set to the same

amount for each of the viewing functions or may be individually set to the same or different amounts for each ofthe viewing functions provided.

After the received signals have been appropriately mapped and translated, the

rotation axis ofthe part may be dynamically calculated, as generally shown at step S.305

in Fig. 39. Depending on the current view of the part, the rotation axis may be defined

to pass through the center ofthe part or to pass through another point so that the zoomed

area of the part will not disappear from the screen when the part is rotated with, for

example, a high zoom ratio or factor. Various methods and processes may be utilized to

dynamically recalculate the rotation axis ofthe part based on the current zoom view. In

accordance with another aspect ofthe invention, Fig. 41 illustrates an exemplary logic

flow and sequence of processes and steps that may be performed to calculate the rotation

axis whenever the view ofthe part has been modified by the user.

As shown in Fig. 41, the current zoom factor or ratio and the part position and

current view may be determined at steps S.311 and S.313. The zoom factor and orientation of the rendered part selected by the user will cause the entire part to be visible

on the screen (i.e., a full view) or cause only a portion ofthe part to be made visible on

the screen (i.e., a partial view). Thus, the current zoom factor and part orientation should

be determined in order to properly set the rotation axis of the rendered part. Various

methods and processes may be used in order to determine the current view ofthe part.

As described above, a visibility function may be provided with the bend model viewer

ofthe present invention to maintain and update the status ofthe current view orientation

and zoom ratio whenever there is a change made to the displayed image. A function call

may be made to the bend model viewer to determine what points or portions of the part

are presently visible. Whether all of the part is displayed on the screen may be

determined by comparing the view volume with the part's boundary base size.

If it is determined at step S.315 that a full view of the part is currently visible on

the screen, then the rotation axis may be set to pass through the center of the part at step

S.317. Setting the rotation axis to the center of the part when there is a full view is

possible, since the entire rendered part will be visible on the screen when it is rotated by

the user. With the entire part visible on the screen, the rotation axis may be defined so

as to pass through the geometric center or centroid of the part. Conventional coordinate

geometry techniques may be used to define and set the rotation axis to the geometric

center of the part. In addition, the direction of the rotation axis may be defined as a

vector that is peφendicuiar to the vector from the previous cursor position to the current

cursor position.

If it is determined at step S.315 that only a partial view of the part is currently

visible on the screen, then logic flow may continue to steps S.319-S.325 in order to calculate the rotation axis so that portions ofthe rendered part will not disappear from the

screen when the zoomed part is rotated by the user. As described above, when a high

zoom factor is selected by the user and only portions of the part are displayed on the

screen, the rotation axis should not be set to pass through the geometric center ofthe part,

since to do so may cause the zoomed portion(s) ofthe displayed part to disappear from

the screen during rotation. In order to prevent the displayed portion of the part from

being occluded or disappearing from the screen, the rotation axis should be set so as to

pass through the coordinate of the closest point to the viewpoint (i.e., camera) at the

center of the screen. In such a case, the direction ofthe rotation axis may be defined as a vector that is peφendicularto the vector from the previous cursor position to the current

cursor position.

Thus, at step S.319, the center ofthe screen is located and the object or portion

ofthe part at the center ofthe screen that is closest to the camera is selected. That is, the

portion of the rendered part that is located at the center of the screen and that is closest

to the camera or user viewpoint ofthe screen is picked.

If it is determined at step S.321 that there is an object at the camera (e.g., that there is a

solid portion ofthe part that is located at the center ofthe screen and that is closest to the

camera), then at step S.325 the rotation axis may set to pass through the picked point.

As discussed above, the direction ofthe rotation axis may be defined as a vector that is

peφendicuiar to the vector from the previous cursor position to the current cursor

position.

If it is determined at step S.321 that there is not an object at the camera (e.g., that

the part includes a hole or opening that is located at the center of the screen and that is closest to the camera), then logic flow may continue to step S.323. At step S.323, the

rotation axis may altematively be define to pass through the center ofthe screen (e.g., the

X and Y coordinate of the physical center of the screen) and at the Z-coordinate (i.e.,

depth) equal to the geometric center ofthe part. Thus, the rotation axis may be set to pass

through the X, Y, and Z coordinates discussed above, with the direction ofthe rotation

axis being defined as the vector that is peφendicuiar to the vector from the previous

cursor position to the current cursor position.

Referring back to Fig. 39, after the dynamic rotation axis has been calculated, the

selected viewing function (e.g., zoom, rotate, pan, etc.) may be called at step S.307. As

discussed above, the various viewing functions of the 3-D manipulation system may be

defined and implemented as member functions of the view class of the bend model

viewer (see, e.g., Fig. 18 and the related disclosure provided above). In such a case,

based on the viewing function selected by the user, a function call may be made to the

bend model viewer to update the current view ofthe rendered part, at step S.309. The

current view and orientation of the part may be updated based on the viewing function

selected by the user and the mapped cursor movements received from the user-controlled

input device (e.g., the mouse or joystick device). A graphics package, such as OpenGL

or RenderWare, may be provided to facilitate the update ofthe current view provided to

the user.

The logic flow and processes performed in the exemplary flowcharts of Figs. 39

and 41 may be implemented by software and by using a wide variety of programming

languages and techniques. For example, object oriented programming techniques and

C++ may be used to implement the noted processes and operations. An exemplary code for implementing the 3-D manipulation system ofthe present invention is provided in

Appendix L. The exemplary source code was written in C++ programming language and

includes the various processes and operations for calculating the dynamic rotation axis.

Comments are provided in the code of Appendix L to facilitate the analysis ofthe logic

and algorithms used therein.

Although the 3-D manipulation system described above has been described with

respect to the use of a joystick device and control buttons, the system may also be

implementedby any other particular type of input means, including a mouse or keyboard.

Further, in the above-described embodiments of Figs. 37-38, boundaries may be defined

to limit zooming or panning ofthe object in or out ofthe screen into infinity, since such continuous zooming or panning may cause the system to fail or crash.

In addition, various other features may be implemented in connection with the

joystick interface. For example, movement in any ofthe viewing functions may not be

effectuated unless the joystick is moved beyond a predetermined range or distance from

the joystick center position. Requiring such a threshold of movement of the joystick

before movement ofthe part is permitted prevents accidental movements ofthe rendered

part based on inadvertent handling or pushing of the joystick from the center position.

Other features may also be provided to improve the joystick interface and system

interaction with the user. For instance, continuous or incremental (e.g., step-wise)

movement in any one of the viewing functions (e.g., zoom, rotate, pan, etc.) may be

provided based on the single operation of the joystick by the user. Selection of the

continuous or incremental movements may also be provided based on the amount or

duration of movement of the joystick in any single direction. If desired, the rate of scaling or movement ofthe rendered part may also be increased based on the degree or

duration of movement of the joystick in any direction. Modification of the speed

adjustment factor described above may also be implemented by permitting the user to

manual insert the correction to the adjustment factor to increase or decrease the rate of

scaling.

Various other features and embodiments may also be implemented in the present

invention in order to aid in the design and manufacturing of components at the factory.

For example, a bar code system may be implemented to keep track of and access

information concerning each customer's order. A bar code with a predetermined

reference or job number may be assigned to each part ordered by a customer. This bar

code may be used for accessing and retrieving job information from database 30. A bar

code reader or scanner, such as a Barcode Anything SCAN CCD sensor from Zebra

Technologies VTI, Inc., Sandy, Utah, may be provided at each ofthe locations to permit

users to scan the bar code for a particular job in to the server module or station module

and to access and retrieve critical design and manufacturing information associated with

that part that is stored in database 30. The bar code reader may be plugged into the

computer of each of the station modules and/or the server module. The bar codes may

be formatted in accordance with any conventional bar code formats, such as UPC-A,

Codabar, Code 39, EAN/JAN-8 or Plessey, and the resultant bar code number may be

translated in accordance with a lookup table to find the corresponding job reference

number and/or file name in order to retrieve the job information from the database.

Altematively, the job number may be typed in or selected from a displayed directory at any ofthe stations located throughout the factory to instantaneously retrieve and display the job information at the user's location. The ability to instantaneously retrieve such

information is aided by the use of communications network 26 and the storage of the

design and information at a centrally located database, such as database 30.

In accordance with yet another aspect ofthe present invention, an apparatus and

method for scheduling and assigning jobs may be provided in the proposed system.

Conventionally, the scheduling and assignment of jobs throughout a manufacturing

facility is performed by a shop or factory foreman who determines the current set-up and

availability of machinery, as well as the status of current jobs. After gathering and

analyzing this information, the shop or factory foreman may develop a schedule and

distribute the assignments for the jobs (e.g., in the form of a job schedule sheet that is

distributed to the factory floor) that are to be performed at the various locations in the

factory. The scheduling and assignment of jobs is provided to ensure that each

customer's job is completed in a timely fashion and by the specified delivery date. The

conventional process of scheduling and assigning jobs is, however, often arduous and

usually performed manually by the factory foreman.

In accordance with an aspect ofthe invention, a job assignment and scheduling

system may be provided to assist a shop or factory foreman in establishing job schedules

for the factory. The system may take advantage of communications network 26 and the

bend model information that is stored in database 30 to automatically gather the

necessary information so that the shop foreman may more easily develop a job schedule.

The system may be implemented through software or programmed logic at the server

module or station modules located throughout the factory. By entering the various jobs

that need to be scheduled, the system software may analyze the design and part information and determine which machines are best suited for doing particular jobs. For

this puφose, the current status and set-up ofthe machinery in the factory may be defined

and stored in database 30 and accessed by the job scheduler software. Based on various

criteria, the software may suggest to the foreman, in the form of a display, which

machines are available to perform a particular job and which machines cannot perform

other jobs. In this regard, a table may be displayed that ranks the availability of machines

for particular jobs and that provides a proposed job schedule that may be implemented

or modified by the shop foreman.

The criteria that may be used to set and recommend job schedules may include

a wide variety of criteria including the current set-up of each machine in the factory, the

types of bends and tooling required for each job and the other types of jobs that need to

be performed within the same time frame or period. Information from the bend model

file for each part may also be utilized, such as the bend angle, the flange length, and type

of bending, in order to determine which machines can perform a particular job. A table

stored at, for example, database 30, may include critical information as to the current set¬

up and capabilities of each ofthe punching and bending machines on the shop floor.

Based on the proposed job schedule, the foreman may assign jobs to various

locations throughout the factory in order to optimize production and output capacity of the factory. The final job schedule or assignment may be entered electronically and

routed to each of the machines through communications network 26. A pilot lamp, such

as an LED, may be provided at each of the bending or machinery work stations to

indicate and confirm that a job has been assigned or transferred to that station. The job

assignment and schedule may be stored in a file ofthe server module that can be accessed instantaneously from any location within the factory.

In addition to the above features, other miscellaneous features may be

implemented in accordance with the teachings of the present invention. For example,

menu screens may be provided and displayed at the various station modules and locations

to facilitate the user in selecting the various display and function modes ofthe present

invention. For example, a main menu screen such as that shown in Fig. 42 may be

provided to a user when the station module is initialized. The main menu window

display may include icon images of each of the available window displays and viewing

modes provided by the station module. The main menu screen may pop-up anytime a

menu button (e.g., FI key) is selected. The user may select a window by moving a

highlighted block to the desired window icon and selecting the same. Such operations

may be performed through the use of a keyboard, mouse or joystick.

Other window screens may also be provided and displayed to the user to facilitate

the entry and display ofthe job information. For example, a part information window

may be displayed to permit a user to enter or modify the part information. An example

of a part information window display is provided in Fig. 43. The part information

window may include all of the pertinent part information (e.g., part number, material

type, dimensions, etc.) and may include a 2-D flat drawing and isometric view of the

sheet metal part. A bendline information window, such as that illustrated in Fig. 44, may

also be provided to allow a user to view the various bendline information, including the

bend sequence and deduction amount for each bendline. The bendline information

window may permit the user to enter or modify the bendline information for each bend,

and may include both a 2-D flat drawing and isometric view ofthe sheet metal part. Additional window display may also be provided to facilitate a bending operator's

analysis ofthe bending sequence. For example, a bend sequence window display and a

bend simulation window display may be provided to indicate the various bending stages

of the part and to simulate the part orientation during bending operations. A bend

sequence window, such as that shown in Fig. 45, may be selected from the main menu

screen and displayed to the user to indicate the intermediate shapes ofthe part (in static form) at each stage ofthe bending sequence. A bend simulation window (see, e.g., Fig.

46) may also be selected by the user and provide both static information ofthe bending

stages (in the form of part icons provided on the righthand side of the screen) and active

simulation (in the center ofthe display) ofthe orientation and bending performed at each stage in the bend sequence. By intermittently selecting the part icons on the screen, the

user may view an active simulation ofthe orientation of the part during bending at the

stage represented by the selected part icon. The part may be flipped, translated and

bent/rotated about the bendlines in order to actively simulate each bend sequence.

Each ofthe above-described window displays of Figs. 43-46 may be selected and

displayed to a user from the main menu window display of Fig. 42. In addition, a user

at any of the station modules may select the appropriate window icons in the main menu

window display to have the 2-D and/or 3-D representations of the part displayed in

accordance with the viewing modes (e.g., 2-D flat, wire-frame, solid, orthographic) of the

invention, as described in greater detail above with reference to Figs. 19-22. Various

menu windows may also be provided, for example, at the station modules to facilitate the operation of the features and functions of the present invention. Fig. 47 illustrates

exemplary menus that may be displayed for the 2-D to 3-D operations. In addition, Fig. 48 illustrates an exemplary menu structure for the 2-D cleanup operation ofthe invention

The present invention, however, is not limited to these menu arrangement, and other

menu screens and/or tool icon bars may be provided to facilitate a user's interaction with

the system.

Other features may also be implemented in the present invention. For example,

higher levels of automation may be provided to facilitate the development of the bending

plan. For example, bending and tooling expert systems may be provided to develop and

propose tooling set-up and bending sequences based on the part geometry and shape for

each job, such as that disclosed in pending, U.S. patent application Serial Nos.

08/386,369 and 08/338,1 15.

While the invention has been described with reference to several exemplary embodiments, it is understood that the words which have been used herein are words of

description and illustration, rather than words of limitations. Changes may be made

without departing from the scope and spirit of the invention and its various aspects.

Although the invention has been described herein with reference to particular means,

materials and embodiments, the invention is not intended to be limited to the particulars

disclosed herein; rather, the invention extends to all functionally equivalent structures,

methods and uses. APPENDIX A

///////////////////////////////////////////////////////////////////

// //

// Example of feature entity extraction from BendModel parts //

// //

/////////////////////////////////////////////////////////////////// int SPS_cal_part_matrix ( BM_PART *part, FENT **inp_sp, char *pname, int *inp_parran , int *inp_part_numface, int *inp_part_numbend, int *inp_part_maxface )

{

// local variables int parrank, part_numface, part_numbend, part_maxface ; int **imatrix = NULL ; int *nbl_face = NULL ;

BENDFACE *facelist = NULL ;

BENDLINE *bendlinelist = NULL ;

ARTIFICIAL_FACE *bendfacelist = NULL ;

int num_error = 0 ; // counter for the number of errors in the part int face_l, face_2, bend_l, bend_0, i ; long num_of_faces, num_of_bendlines, entity_id;

BM_FACE *face;

BM_BENDLINE *bendline;

BM_2D_BODY *t o_d_body;

BM_3D_BODY *three_d_body;

BM_BEND_OP *bend_op;

BM_TOPOLOGY_RECORD *topology_record;

// get name and number of faces and bendlines of the part part->get_name{ pna e ) ; double thickness = part->get_metal_thickness () ; num_of_faces = part->get_number_of_faces() ; num_of_bendlines = part->get_number_of_bendlines () ; APPENDIX A

if (num_of_faces == 0 | | num_of_bendlines == 0) return ( -1 ) ;

// create local working array facelist = new BENDFACE [num_of_faces] ; bendlinelist = new BENDLINE [num of bendlines] ;

// count number of faces defined. double maxfacearea = -1.0 ; part_maxface = 0 ,- part_numface = 0 ; face = part->get_face_list () ; for ( i = 0 ; face && i < num_of_faces ; i++, face = (BM_FACE *) (face->next () ) ) { // Count the defined faces.

// initialize the struc facelist [i faceid = 0 ; facelist [i facept = NULL ; facelist [i .twodpt = NULL ; facelist [i , hreedpt = NULL ; facelist [i ,topologyrecpt = NULL facelist [i ,numadjbody = 0 facelist [i .numadjbend = 0 facelist [i ,numadjhole = 0 facelist [i .numadjface = 0 facelist [i .face area = 0.

if (face == NULL) break ; two_d_body = face->get_3D_version() ; if (two_d_body == NULL) continue ;

// It is considered as a valid face, when its BM_2D_BODY exists. part_numface++ ; facelist [i] . faceid = part_numface ; facelist [i] .facept = face ; facelis ti] .twodpt = two_d_body ; facelist [i] .face_area = area_of_a_2D_body(two_d_body) ; if (maxfacearea < facelist [i] .face_area) { maxfacearea = facelis [i] .face_area ; part maxface = facelis fi] .faceid ; APPENDIX A

} three_d_body = two_d_body->get_3D_body() ; facelist [i] .threedpt = three_d_body ;

if (three_d_body == NULL) continue ; entity_id = three_d_body->get__name () ; facelist [i] .org_entity = entity_id ; topology_record = three_d_body->get_adj_list () ; facelist [i] .topologyrecpt = topology_record ;

if (topology_record == NULL) continue ; facelist [i] .nu adjbody = (int) topology_record->get_number_of_adjacent_bodies () ; facelist [i] .numadjface = (int) topology_record->get_number_of_adjacent_faces () ; facelist [i] .numadjhole = (int) topology_record->get_number_of_adjacent_holes () ; facelist [i] .numadjbend = (int) topology_record->get_number_of_adjacent_bendlines () ; } if (num_error > 0) { clean_up ( part_numface, imatrix, nbl_face, facelist, bendlinelist, bendfacelist) ; return ( num_error ) ;

} if (part_numface == 1) {

// this is a trivial case, where the part only has one flat face.

*inp_part_numface = part_numface ;

*inp_parrank = part_numface ;

*inp_part_numbend = 0 ;

*inp_part_maxface = 1 ;

*inp_sp = new FENT [2] ;

*inp_sp[2] = NullFent ; clean_up ( part_numface, imatrix, nbl_face, facelist, bendlinelist, bendfacelist) ; return ( 0 ) ;

} APPENDIX A

// varify all the valid face, the current requirements are: // 1) A face cannot be adjacent to another face. // 2) A face without an adjacent bendline is not allowed. // (Note: The single faced part has been processed.) // Also, creat a pointer array that links assigned part__face_id

// to the netry in the facelist. int *fid_pt = new int [part_numface] ; for ( i = 0 ; i < num_of_faces ; i++) { if (facelist [i] .faceid) { fid_pt [facelist [i] .faceid] = i ; if (facelist [i] .numadjface | | facelist [i] .numadjbend < 1) num_error++ ;

}

} if (fid_pt) delete [] fid_pt ; if (num_error > 0) { clean_up ( part_numface, imatrix, nbl_face, facelist, bendlinelist, bendfacelist) ,- return ( num__error ) ;

}

// count the number of bendlines that is defined. part_numbend = 0 ; bendline = part->get_bendline_list () ; for ( i = 0 ; bendline && i < num_of_bendlines ; i++, bendline = (BM_BENDLINE *) (bendline->next ()) )

{

// initialize the struct bendlinelist [i] .bendlineid = 0 ; bendlinelist [i] .bendlinept = NULL ; / / BM_BENDLINE pointer bendlinelist ti] .bendoppt = NULL ; / / BM_BEND_0P pointer bendlinelist ti] .twodpt = NULL ; / / BM_2D_BODY pointer bendlinelist [i] .threedpt = NULL ; / / BM_3D_B0DY pointer bendlinelist [i] .topologyrecpt = NULL ; / / APPENDIX A

BM_TOPOLOGY_RECORD pointer bendlinelist [i] .numadjbody = 0 bendlinelist [i] .numadjbend = 0 bendlinelist [i] .numadjhole = 0 bendlinelist [i] .numadjface = 0

if (bendline == NULL) break ; two_d_body = bendline->get_3D_version() ; if (two_d_body == NULL) continue ;

// It is considered as a valid bendline, when its BM_2D_BODY exists. part_numbend++ ; bendlinelist [i] .bendlineid = part_numbend ; bendlinelist [i] .bendlinept = bendline ; // BM_BENDLINE pointer bendlinelist ti] . twodpt = two_d_body // BM 2D BODY pointer bend__op = bendline->get_bend_op() ; bendlinelist ti] .bendoppt = bend_op ; // BM_BEND_0P pointer if (bend_op == NULL) num_error++ // Note: Bend operation must be defined for each // bendline, otherwise it is an error. three_d_body = two_d_body->get_3D_body () ; bendlinelist [i] .threedpt = three_d_body ; // BM_3D_B0DY pointer

if (three_d_body == NULL) continue ; entity_id = three_d_body->get_name () ; facelist [i] .org_entity = entity_id ; topology_record = three_d_body->get_adj_list ( ) ; bendlinelist [i] . topologyrecpt = topology_record ;

if (topology_record == NULL) continue ; bendlinelist [i] . numadj body = ( int ) topology_record- >get_number_of _ad j acent_bodies ( ) ; bendline list [i ] . nuraadj ace = ( int ) topology_record- >get_number_of_adjacent_f aces ( ) ; APPENDIX A

bendlinelist [i] .numadjhole = (int) topology_record- >get_number_of _adj acent_holes ( ) ; bendlinelist [i] .numadjbend = (int) topology_record->get_number__of_adjacent_bendlines ( ) ;

} if (num_error > 0) { clean_up ( part_numface, imatrix, nbl_face, facelist, bendlinelist, bendfacelist) ; return ( num_error ) ;

}

// varify all the valid bendlines, the current requirements are:

// 1) The total number of bendlines should not be less than

// the total number of faces minus 1.

// 2) A bendline without an adjacent face or bendline is not allowed.

// 3) A bendline with more than two adjacent faces and bendlines is not allowed.

// 4) The adjacent face or bendline of a bendline must be a valid face or

// bendline that is defined in the facelist or bendlinelist.

// 5) Two adjacent faces, facel and face2, will be defined for each bendline

// a) If the bendline has an adjacent faces, then the face's faceid is used

// as facel or face2.

// b) If the bendline has an adjacent bendline, then a bendline_only_face

// will be created inbetween these two bendlines. The faceid of the

// bendline_only_face will be used as facel or face2.

// c) If the bendline has only one adjacent face or adjacent bendline, then

// a bendline_only_face will be created for this bendline. The faceid of

// the bendline_only_face is face2.

// maxnewfaces is the maximum number of bendline_only_face APPENDIX A

need to be created

// without encounter error in the part. if (part_numbend > part_numface-l) num_error++ ; // condition 1 if (num_error > 0) { clean_up ( part_numface, imatrix, nbl_face, facelist, bendlinelist, bendfacelist) ; return ( num_error ) ;

} int maxnewfaces = part_numbend + 1 - part_numface ; if (maxnewfaces > 0) ( bendfacelist = new ARTIFICIAL_FACE [maxnewfaces] ; bendfacelist [0] . faceid = -part_numface ;

} for ( i = 0 ; i < num_of_bendlines ; i++) { if (bendlinelist [i] .bendlineid) { bend_0 = bendlinelist [i] .bendlineid ; int numadj = bendlinelist fi] .numadjface + bendlinelist [i] .numadjbend ; if (numadj < 1 | | numadj > 2) num_error++ ; // condition 2 & 3 else { if (bendlinelist [i] .numadjface > 0) { / / condition 4 - first face t h r e e _ d _ b o d y bendlinelist fi] .topologyrecpt->get_first_face () ; face_l = find_face_id( three_d_body, facelist, num_of_faces ) ; if (face_l <= 0) num_error++ ; else bendlinelist [i] .facel = face_l ;

} if (bendlinelist [i] .numadjface == 2) { / / condition 4 - second face t h r e e _ d _ b o d y bendlinelist [i] . topologyrecpt->get_next_f ace () ; face_l = f ind_f ace_id ( three_d_body, facelist, num_of_faces ) ; if (face 1 <= 0) num error ++ ; APPENDIX A

else bendlinelist [i] .face2 = face_l ;

} if (bendlinelist [i] .numadjbend > 0) { // condition 4 - first bendline t h r e e _ d _ b o d y bendlinelist [i] . topologyrecpt->get_f irst_bendline ( ) ; bend_l = find_bendline_id( three_d_body, bendlinelist, num_of_bendlines ) ; if (bend_l <= 0) num_error++ ; else { face_l = define_bendline_only_face ( bend_l, bend_0, bendfacelist, maxnewfaces ) ; if (face_l <= 0) num_error++ ; else { if (bendlinelist [i] .numadjface

> 0) bendlinelist ti] .face2 = face_l else bendlinelist[i] .facel face 1

}

} if (bendlinelist [i] .numadjbend == 2) { // condition 4 - second bendline t h r e e _ d _ b o d y = bendlinelist [i] . topologyrecpt->get_next_bendline () ; bend_l = find_bendline_id( three_d_body, bendlinelist, num_of_bendlines ) ; if (bend_l <= 0) num_error++ ; else { face_l = define_bendline_only_face ( bend_l, bend_0,

bendfacelist, maxnewfaces ) ; if (face_l <= 0) num_error++ ; else bendlinelist fi] .face2 = face 1 APPENDIX A

} } if (numadj == 1) { face_l = define_bendline_only_face ( bend 0, 0,

bendfacelist, maxnewfaces ) ; if (face_l <= 0) num_error++ ; else bendlinelist [i] . face2 = face_l ;

}

// now check whether there is any bendline only face been created

// increase the part_numface if there is. int numregfaces = part_numface ; int numnewfaces = 0 ; if (maxnewfaces > 0) { for ( i = 0 ; i < maxnewfaces ; i++ ) { if (bendfacelist [i] .faceid <= 0) { numnewfaces = i + 1 ; break ;

}

} part_numface += numnewfaces ;

}

// first create a integer topological matrix to record all the topological relations int j ; imatrix = new int * fpart_numface] ; for ( i = 0 ; i < part_numface ; i++ ) { imatrix[i] = new int [part_numface] ; APPENDIX A

for ( j = 0 ; j < part_numface ; j ++ ) imatrix ti] [j ] = 0

} for ( i = 0 ; i < num_of_bendlines ; i++ ) {

// save the bendline entry + 1 in imatrix if (bendlinelist [i] .bendlineid) ( face_l = bendlinelist [i] . facel ; face_2 = bendlinelist [i] . face2 ; imatrix [face_l-l] [face_2-l] = i+1 ; imatrix [face_2-l] [face_l-l] = i+1 ;

} }

// from imatrix to find the number of bendlines of each face, nbl_face [i] ,

// and to verify that each face has at least one bendline nbl_face = new int [part_numface] ; for ( i = 0 ; i < part_numface ; i++ ) { nbl_face [i] = 0 ; for ( j = 0 ; j < part_numface ; j ++ ) { if ( imatrix ti] [j ] ) nbl_f ace ti] ++ ;

} if ( ! nbl_face [i] ) num_error++ ;

} if (num_error > 0 ) { clean_up ( part_numface, imatrix, nbl_face, facelist, bendlinelist, bendfacelist) ; return ( num_error ) ;

}

// create the Cbpart' s topological matrix's input data FENT array

// and initialize it. part->get_name ( pname ) ; parrank = part_numface ; int spsize = parrank* (parrank + l)/2 + 1 ; // +1 is for End of FENT array indicator

FENT *sp = new FENT [ spsize ] ; for ( i = 0 ; i < spsize-1 ; i++ ) *(sp+i) = NoRelation ;

* (sp+spsize-1) = NullFent ;

10 APPENDIX A

// step 1: set up the positive or negative bend // The included FENT's are:

// * PosBend = ^• + ^■ ; // positive bend between two faces

// * NegBend = ' - ' ; // negative bend between two faces

// * P90Bend = 'A' ; // 90 deg positive bend angle

// * N90Bend = 'B' ; // 90 deg negative bend angle

// MrPosBend = '3' ; // multiple positive bendlines between two faces

// MrNegBend = '4' ; // multiple negative bendlines between two faces

// * marks what is currently implemented, for ( i = 0 ; i < num_of_bend lines ; i++ ) { if (bendlinelist fi] .bendlineid) { face_l = bendlinelist fi] .facel ; face_2 = bendlinelist [i] . face2 ; FENT btype = NegBend ;

BM_BEND_OP_REGULAR * bendopr egu 1 ar

(BM_BEND_OP_REGULAR * ) bendlinelist [i] . bendoppt ; if (bendopregular->get_bend_type () ) btype = PosBend

double angle = bendopregular->get_bend_angle () ; if (angle > PI) {

// angle > PI => reverse bend direction and reset the bend angle angle = 2*PI - angle ; if (btype == PosBend) btype = NegBend ; else btype = PosBend ;

} bendlinelist [i] .bendangle = angle ; bendlinelist [i] .bendtype = btype ; // set up 90 degree bend type if (angle == PI_over_2) { if (btype == PosBend) btype = P90Bend ;

11 APPENDIX A

else btype = N90Bend ;

}

// set_FENT (sp, face_l, face_2, parrank, btype) ;

}

// step 2: set up the corner relationships, which is the relation

// between two faces that are connected to a common face.

// The included FENT are:

// * TouchCnr = 'T' ; // two faces same bend dir

// * OpenCnr = '0' ; // two faces same bend dir

// * PrllBend = 'P' ; // two parallel bendline same bend angle dir opposite bendline dir

// * SerlBend = 'S' ; // two parallel bendline same bend angle dir same bendline dir

// * cLnrBend = 'L' ; // colinear bendline same bend angle dir on one face

// * DfClnrBend = 'D' ; // colinear bendline same bend angle on different faces

// * tHkOffBend = 'H' ; // tHickness offset bendline same bend angle dir on two neighboring face

// * touchCnr = 't' ; // two faces opposite bend dir

// * openCnr = 'o' ; // two faces opposite bend dir

// * prllBend = 'p' ; // two parallel bendline opposite bend angle dir opposite bendline dir

// * serlBend = ' s' ; // two parallel bendline opposite bend angle dir same bendline dir

// * clnrBend = ' 1 ' ; // colinear bendline opposite bend angle dir on one face

// thkOffBend = 'h' ; // tHickness offset bendline opposite bend angle dir on two neighboring face

// * marks what is currently implemented.

12 APPENDIX A

// Algorithm : for every face that has more than one bend line then

// a pair of any two bend lines will have a relationship. for ( i = 0 ; i < part_numface ; i++ ) { if (nbl_face[i] > 1) { int face_c = i + 1 ;

// create a list of faces that are connected to this face for ( j = 0 ; j < part_numface ; j ++ ) { if ( imatrix f i] [j ] ) { int bl_l = imatri [i] [j ] ; face_l = j + 1 ; for ( int k = j +1 ; k < part_numface ; k++

) { if ( imatrix [i] [k] ) { int bl_2 = imatrix [i ] [k] ; face_2 = k + 1 ;

// define the relation ship between the two bendlines set bendline_rel FENT ( sp, parran ,

facelist, face_c, face_l, face_2,

bendlinelist, bl_l, bl_2, outfile) ;

} } } } } }

// from imatrix to find the faces that may be of the thickness offset bendlines

// and verify these faces first based on the touch corner infomation then based on

// the bendlines' distance and parallel condition.

13 APPENDIX A

for ( i = 0 ; i < part_numface ; i++ ) { if ( nbl_face [i] < 2 ) continue ; // face i should have at least 2 bendlines for ( j = i+1 ; j < part_numface ; j++ ) { if ( ! imatrix [i] fj] | | nbl_face[j] < 2 ) continue ; // faces i and j must have a common bendline

// and face j should have at least 2 bendlines for ( int i2 = 0 ; i2 < part_numface ; i2++ ) { if ( !imatrixfi] fi2] | | i2 == j ) continue ; // faces i and i2 must have a common bendline

// and face i2 is different from j

// first requirement - bendlines imatrix[i] [j] and

// imatrixti] [i2] form a touch corner if ( get_FENT(sp, j+1, i2+l, parrank) !=

TouchCnr ) continue ; for ( int j2 = 0 ; j2 < part_numface ; 2++ )

{ if ( !imatrix[j] [J2] | | j2 == i ) continue

// second requirement - bendlines imatrix [i] [j] and

// imatrixtj] [j2] also form a touch corner if ( get_FENT(sp, i+1, j2+l, parrank) !=

TouchCnr ) continue

// comes here, we have obtained a candidate for the

// thickness offset bendlines, the two candidate

// bendlines are imatrix fi, i2] and imatrix[j] [j2] int bl_l = imatrixfi] [i2] ; int bl_2 = imatrix [j] [j2] ; check_thkoffbend ( sp, parrank, facelist, i+1, i2+l, j+1, 2+1, bendlinelist, bl_l, bl_2, thickness, outfile) ;

}

14 APPENDIX A

}

// DfClnrBend = 'D^■ ; // colinear bendline same bend angle on different faces

// Here is to find all the colinear bends that are colinear but not related to each other // with a common face, int num_undetermined = 0 ;

UNDETERMINED *undetermined = new UNDETERMINED [num_of_bendlines] ;

UNDETERMINED *undetpt = undetermined ; for ( i = 0 ; i < num_of_bendlines ; i++ ) { if (bendlinelist fi] .bendlineid) { int face_il = bendlinelist [i] .facel ,^■ int face_i2 = bendlinelist [i] . face2 ; for ( j = i+1 ; j < num_of_bendlines ; j++ ) { if (bendlinelist [j] .bendlineid) { int face_jl = bendlinelist [j] . facel ; int face_j2 = bendlinelist fj] . face2 ; if ( face_il == face_jl | | face_il == face_j2 I I face_i2 == face_jl | | face_i2 == face_j2 ) continue ; if ( bendlinelist [j] .bendtype != bendlinelist [j] .bendtype ) continue ; if ( bendlinelist [j] .bendangle != bendlinelist [j] .bendangle ) continue ;

// come here when the two bend lines have the same bend angle and type,

// and they do not share a common face.

// now examine whether they are colinear int bl_i = i + 1 ; int bl_j = j + 1 ; int unknown = check_DfClnrBend ( sp, parrank, facelist, face_il, face_i2, face_jl, face_j2,

15 APPENDIX A

bendlinelis , bl_i, bl_j, part_numf ac e , imatrix, outfile) ; if (unknown) { undetermined[num_undetermined] .bendline_i = bl_i ;

undetermined[num_undetermined] .bendline_j = bl_j ; num_undetermined++ ,- } } } } }

// Note: if num_undetermined is not zero, then there are confirmed DfClnrBend

// but with undetermined faces for specify this FENT.

// A tree structure of all the faces that records the connectivity between

// faces need to be constructed. And then each pair of undetermined

// bendlines need to be processed to determine which two faces of the four

// that connected with the two bendlines should be used to record this FENT.

// Currently, we will neglect these information and simply delete

// undetermined array, which will not be used any more. delete [] undetermined ;

// transfer all the data back *inp_part_numface = part_numface *inp_parrank = parrank ; *inp_part_numbend = part_numbend *inp_part_maxface = part_maxface *inp_sp = sp ;

16 APPENDIX A

num_error = fclose (ou file) ; clean_up ( part_numface, imatrix, nbl_face, facelist, bendlinelist, bendfacelist) return ( num error ) ;

17 APPENDIX B

/////////////////////////////////////////////////////////////////

// / // Example of similarity index feature. Compares two parts // // topological matrix and search for their best similarity // // index. The similarity index is the summation of all the // // penalties of the mismatched FENTs of the two part's FENT // // matrix (topological matrix) . The best similarity index // // is the one with the smallest value. // // // // Copyright 1996 AmadaSoft America, Inc. // // //

////////////////////////////////////////////////////////////////// int matrix_similarity_index ( FENT *newmatrix, int newmatrixdim, int *newmatchlist, FENT *oldmatrix, int oldmatrixdim, int *oldmatchlist, int nfacefixed, int requiredvalue, int *minimumvalue) { //input

/ / newmatrix - new part topological matrix to be matched on

// by the oldmatrix

// newmatrixdim - matrix dimension of newmatrix

// newmatchlist - the first nfacefixed entries should

// contain the entries (faces) of the new part

// matrix that has already been fixed for

// matching with the old part matrix.

// oldmatrix - old part matrix to match with the new part

// topological matrix

// oldmatrixdim - matrix dimension of oldmatrix

// oldmatchlist - the first nfacefixed entries should

// contain the entries (faces) of the old part

// matrix that has already been fixed for

// matching with the new part matrix.

// nfacefixed the number of entries in matchlist that

// has already been matched,

Enter 0 when

// there is no prematched entries.

// requiredvalu< - the known or interested upper bound on the

// minimum similarity index.

When this value APPENDIX B

// is inputed as a positive value, then a

// branch of the search tree in finding the

// minimum value of the similarity index may

// be cut off when the branch's current value

// becomes greater than this minimum value.

// set to 0 or a negative value when there is

// no restriction. In this case, a minimumvalue

// and its corresponding macthlists will always

// be found and returned. //output

// newmat enlist the first nfacefixed entries should // contain the entries (faces) of the new part

// matrix that has already been fixed for

// matching with the old part matrix.

// oldmatchlist the first nfacefixed entries should

// contain the entries (faces) of the old part

// matrix that has already been fixed for

// matching with the new part matrix.

// minimumvalue the minimum similarity index, which is the

// smallest possible value of summation of all

// the penalties of the mismatched FENTs of the

// two part's FENT matrix (topological matrix) .

// This value may not exist, if the original

// inputed value (which is the requirement) is

// too small.

// return - index on whether or not found the minimum

// valu .

// = 1, found the minium value and its match. APPENDIX B

// = 0, a new minium value cannot be reached.

// = -1, error in the input data newmatchlist .

// = -2, error in the input data oldmatchlist.

// -3, error, the minimum similarity index of

// the given matrices are larger than the

// inputed minimumvalue.

/////////////////////////////////////////////////////////////////

// create two integer pointer arrays to keep track the

// corresponding entries of the two matrices that matches. int matrixdim = newmatrixdim ; if (matrixdim < oldmatrixdim) matrixdim = oldmatrixdim ; int *pnewmatchlist = new int [ matrixdim ] ; int *poldmatchlist = new int [ matrixdim ] ; if ( poldmatchlist && pnewmatchlist ) { int *pl = pnewmatchlist; int *p2 = poldmatchlist; for ( int icol = 1; icol <= matrixdim; icol++,pl++,p2++) *pl = *p2 = icol;

} else cout << "Unable to allocate memory..." << endl ;

// if there are already fixed entries (nfacefixed > 0) , then reset

// the temporary working pointer arrays (newmatchlist & oldmatchlist)

// to contain those fixed face list, if (nfacefixed > 0) { for (int icol = 0; icol < nfacefixed; icol++) { int iface = * (newmatchlist+icol) ; for ( int jcol = icol; jcol < matrixdim; jcol++) { if (iface == * (pnewmatchlist+jcol) ) { if (jcol != icol) {

* ( p n e w m a t c h l i s t + j c o l ) * (pnewmatchlist+icol) ;

* (pnewmatchlist+icol) = iface ; break ;

}

// comes here only if the inputed face number "iface"

// from newmatchlist is wrong APPENDIX B return (-1) ;

} iface = * (oldmatchlist+icol) ; for ( jcol = icol; jcol < matrixdim; jcol++) { if (iface == * (poldmatchlist+jcol) ) { if (jcol != icol) {

* ( p o l d m a t c h l i s t + j c o l ) * (poldmatchlist+icol ) ;

* (poldmatchlist+icol) = iface ;

} break ;

}

// comes here only if the inputed face number

"iface"

// from oldmatchlist is wrong return (-2) ;

}

// convert the FENT matrix to the counter part of integer matrix

// at the same time, expand the smaller matrix to have the same dimension. int *pnewmatrix = new int [matrixdim*matrixdim] ; int *poldmatrix = new int tmatrixdim*matrixdim] ; convert_fent_to_int (newmatrix, newmatrixdim, pnewmatrix, matrixdim) ; convert_fent_to_int (oldmatrix, oldmatrixdim, poldmatrix, matrixdim) ;

// define required value int requiredv = requiredvalue ; if (requiredv <= 0) requiredv = matrixdim * 400 ;

// create the FENT counters and use calculate_partial_similarity_index

// to calculate the initial similarity index and set the initial FENT

// counts for each column of the fixed faces and that of all the

// unfixed faces.

FENTCOUNT *pcnew = new FENTCOUNT[matrixdim+1] _;

FENTCOUNT *pcold = new FENTCOUNT [matrixdim+1] ; int currentvalue = calculate_partial_similarity_index ( pnewmatrix, pnewmatchlist, pcnew, poldmatrix, poldmatchlist, pcold, matrixdim, nfacefixed ) ; if (currentvalue > requiredv) return (-3) ;

// reset the unfixed faces in pnewmatchlist to be in the

4 APPENDIX B sequence of

// its total weighting. int ncandidates = matrixdim - nfacefixed ; int *pcandidates = new int [ncandidates] ; int *pweights = new int [ncandidates] ; for ( int i = 0 ; i < ncandidates ; i++ ) { int facenew = * (pnewmatchlist+nfacefixed+i) ;

* (pcandidates+i) = facenew ; int weight = 0 ; int *pnewi = pnewmatrix + (facenew-1) *matrixdim ; for ( int j = 0 ; j < matrixdim ; j++, pnewi++ ) weight += fent_weight (*pnewi) ;

* (pweights+i) = weight ; sort_the_candidates (pcandidates, pweights, ncandidates) ; for ( i = 0 ; i < ncandidates ; i++ )

* ( p n e wma t c h l i s t + n f a c e f i xe d + i ) =

* (pcandidates+ncandidates-1-i) ; delete [] pcandidates ; delete [] pweights ;

// call the internal routine recursively to perform similarity index

// search.

♦minimumvalue = requiredv ; int recursive_level = 0; int *newminmatchlist = new int [matrixdim] ; int *oldminmatchlist = new int [matrixdim] ; int *returnlevelcounts = new int [matrixdim] ; int errorcode = matrix_similarity_index_loop

( pnewmatrix, pnewmatchlist, newminmatchlist, poldmatrix, poldmatchlist, oldminmatchlist, pcnew, pcold, &currentvalue, matrixdim, nfacefixed,

&re curs ive_l eve 1 , minimumvalue, returnlevelcounts) ;

// clean up delete pnewmatchlist ; delete poldmatchlist ; delete pnewmatrix ; delete poldmatrix ; delete pcnew ; delete pcold ; delete returnlevelcounts delete newminmatchlist ; delete oldminmatchlist ; return (errorcode) APPENDIX B

}

////////////////////////////////////////////////////////////////// // extract the list of entries from an existing part matrix that matches

// with a new part matrix. The dimension of the part matrix must

// be greater then or equal to the dimension of the topological matrix.

// Note: The first element of matchlist must be initialized to

// contain the face number of the part matrix for macthing

// with the first face in the topological matrix.

// If nmatched is greater than 0, then the first nmatched

// entries in matchlist should contain positive numbers to

// indicate the already matched entries of the part matrix.

////////////////////////////////////////////////////////////////// int matrix_similarity_index_loop

( int *pnewmatrix, int *pnewmatchlist, int *newminmatchlist, int *poldmatrix, int *poldmatchlist, int *oldminmatchlist,

FENTCOUNT *pcnew, FENTCOUNT *pcold, int *currentvalue, int matrixdim, int nfacefixed, int

*recursive_leve1, int *minimumvalue, int *returnlevelcounts) { // input

// pnewmatrix - the new part's topological matrix, where

// the FENT contents have been changed to

// their representing integers. // (integer array of matrixdim * matrixdim)

// pnewmatchlist the list of entries of the newmatrix that

// has been matched by the oldmatrix.

// (integer array of matrixdim) // newminmatchlist - the list of entries of the newmatrix that

// has been matched by the oldmatrix that

// provides the minimum value. // (integer array of matrixdim) // poldmatrix - the old parts' s topological matrix, where

// the FENT contents have been changed to

// their representing integers, // (int array of matrixdim * matrixdim) // poldmatchlist the list of entries of the APPENDIX B oldmatrix that

// has matched with the newmatrix.

// (integer array of matrixdim)

// oldminmatchlist - the list of entries of the oldmatrix that

// has matched with the oldmatrix that

// provides the minimum value.

// (integer array of matrixdim)

// matrixdim - the matrix dimension of both matrices

// nfacematched - the number of entries (faces) that has

// been fixed in matching the two matrices .

// recursive_level - this keep the number of times this procedure

// has been called recursivly.

This provides a

// way to detect the error before going into a

// infinite loop. (integer pointer)

// output

// return - error index,

// = 0 means there is no error,

// > 0 error code.

//

// verify the recursive level if ( *recursive_level > matrixdim ) { cout << "??? Error - the recursive level is too high. ???\n"; cout << " The matrix dimension is " << matrixdim << "An"; cout << " The recursive level is " << *recursive_level << ".\n"; return (9) ; }

// Step 1) Choose a row to be matched with in the new matrix.

// ( Currently, it uses whatever the next row in line. Therefore,

// nothing to do at this moment.

// This may need to be changed to be properly matching up with

// the algorithm that may be used in defining the sequence of the

// candidate list.

// Note: If facenew is not next to the nfacefixed position, APPENDIX B then

// it should be changed to that position. ) int facenew = * (pnewmatchlist+nfacefixed) ;

// Step 2) Create a list of candidate face list in the old matrix.

// and calculate the increases in the similarity index for

// each of the candidates.

// ( Currently we are using what ever the sequence that is inside

// the poldmatchlist.

// One may choose to use the faces that at least matches up with

// the face connectivity of the chosen face in the new matrix.

// Note: The sequence of faces in pcandidates does not need to

// be corresponding to that in poldmatchlist. ) int ncandidates = matrixdim - nfacefixed ; int *pcandidates = new int [ncandidates] ; int *pincreases = new int [ncandidates] ; for ( int i = 0 ; i < ncandidates ; i++ ) { int faceold = * (poldmatchlist+nfacefixed+i) ; * (pcandidates+i) = faceold ; int increase = increase_on_similarity_index ( pnewmatrix, pnewmatchlist, facenew, poldmatrix, poldmatchlist, faceold, pcnew, pcold, currentvalue, matrixdim, nfacefixed, minimumvalue) ; * (pincreases+i) = increase ; }

// Step 3) Sort the candidate face based on the increased values

// the candidates with the lower increase will be tried first. sort_the_candidates (pcandidates, pincreases, ncandidates) ;

// Step 4) change the FENT counters of the newmatrix for fixing facenew int errorcode = change_counts_for_fix_one_face

( pnewmatrix, pnewmatchlist, facenew, pcnew, matrixdim, nfacefixed) ; if (errorcode != 0) return (errorcode) ;

// Step 5) Loop thru the candidate face and based on the increased value

// determines whether or not to continue to the lower level match.

8 APPENDIX B for ( int icandi = 0 ; icandi < ncandidates ; icandi++ ) ( // get the candidate face number of the old matrix and // its corresponding amount of increase on similarity index. int faceold = * (pcandidates+icandi) ; int increase = * (pincreases+icandi) ;

// Now check whether it is any need to continue the matching

// If the current value plus the increase has already exceed the

// minimum value, then there is no need to continue the matching.

// add the returnlevelcount and go to the next candidates. if (*currentvalue + increase >= *minimumvalue) { returnlevelcounts [nfacefixed] += 1 ;

} else if (nfacefixed+1 == matrixdim) {

// A new minimum similarity index has been found, update

// the minimum value

♦minimumvalue = *currentvalue + increase ; for ( i = 0 ; i < matrixdim ; i++ ) { newminmatchlist [i] = pnewmatchlist [i] ; oldminmatchlist [i] = poldmatchlist fi] ; }

} else {

// It is necessary to go down another level in this recursive

// call to define the similarity index.

// change the FENT counters of the oldmatrix for fixing faceold errorcode = change_counts_for_fix_one_face ( poldmatrix, poldmatchlist, faceold, pcold, matrixdim, nfacefixed) if (errorcode != 0) return (errorcode) ;

// call recursively ^♦currentvalue += increase ; *recursive_level += 1 ; errorcode = matrix_similarity_index_loop ⁽ pnewmatrix, pnewmatchlist, newminmatchlist, poldmatrix, poldmatchlist, oldminmatchlist, APPENDIX B pcnew, pcold, currentvalue, matrixdim, nfacefixed+1, recursive_level, minimumvalue, returnlevelcounts) ; if (errorcode != 0) return (errorcode) ; ♦recursive_level -= 1 ; ♦currentvalue -= increase ;

// change the FENT counters of the oldmatrix for unfixing faceold change_counts_for_unfix_one_face ( poldmatrix, poldmatchlist, pcold, matrixdim, nfacefixed+1) ;

} }

// change the FENT counters of the newmatrix for unfixing facenew change_counts_for_unfix_one_face ( pnewmatrix, pnewmatchlist, pcnew, matrixdim, nfacefixed+1) ;

//clean up delete f] pcandidates ; delete U pincreases ; return (0) ;

////////////////////////////////////////////////////////////////// // updates counters for fixing one more face.

////////////////////////////////////////////////////////////////// int change_counts_for_fix_one_face

( int ♦pnewmatrix, int ^♦pnewmatchlist, int facenew,

FENTCOUNT ♦pcnew, int matrixdim, int nfacefixed) { // input

// pnewmatrix - the new part's topological matrix, where

// the FENT contents have been changed to

// their representing integers.

// (integer array of matrixdim ^♦ matrixdim)

// pnewmatchlist the list of entries of the newmatrix that

// has been matched by the oldmatrix.

// (integer array of matrixdim)

10 APPENDIX B

// facenew - the face that is to be fixed. // pcnew - the FENTCOUNT of all the faces.

// matrixdim the matrix dimension of both matrices

// nfacematched - the number of entries (faces) that has

// been fixed in matching the two matrices .

// output

// pnewmatchlist the updated list of entries of the

// newmatrix matched by the oldmatrix,

// with the facenew entry is moved to the

// nfacefixed+1 location.

// pcnew - the updated FENTCOUNT of all the faces

// switch the to be fixed face to the location of nfacefixed+1 // in the pnewmatchlist int ifound = 0 ; for ( int i = nfacefixed ; i < matrixdim ; i++ ) { if (♦ (pnewmatchlist+i) == facenew) {

♦ (pnewmatchlist+i) = ^♦ (pnewmatchlist+nfacefixed) ;

♦ (pnewmatchlist+nfacefixed) = facenew ; ifound++ ;

}

} if ( ifound != 1 ) { cout << "Fatal error from change counts for_fix one_face

/n' return ( 91 )

}

// define the pointer to the FENT on the to be fixed face int *pnewro = pnewmatrix + (facenew-1) ♦matrixdim ;

// first change the counters for the previously fixed faces for ( i = 0 ; i < nfacefixed ; i++ ) { int newcol = ^♦ (pnewmatchlist+i) ; int pnewv = ^♦ (pnewrow + newcol - 1) ; pcne [newcol] .count [pnewv] -- ; int igroup = fent_group (pnewv) ; pcnew[newcol] .gcount [igroup] -- ; if ( pcne [newcol] .count [pnewv] < 0 \ \ pcne [newcol] .gcount [igroup] < 0 ) { c o u t < < " F a t a l e r r o r f r o m change_counts_for_fix_one_face /n" ; return ( 92 ) ;

11 APPENDIX B

}

// second change the counters for the unfixed faces

// use count_specified_fents to initialize the FENTCOUNT of

// the newly seleted face and count the numbers of fents of

// the undetermined columns on the to be fixed rows int listdim = matrixdim - nfacefixed ; pcnew[facenew] = count_specified_fents ( pnewrow, pnewmatchlist+nfacefixed, matrixdim, listdim ) ;

// decrease the FENTCOUNT of the newly seleted face from that

// in remaining unfixed faces for ( i = 0; i < NumlntFent; i++ ) pcnew[0] .count [i] -= pcnew[facenew] .count [i] ; for ( i = 0; i < NumlntFentGroup; i++ ) pcne [0] .gcount [i] -= pcnew[facenew] .gcount [i] ; return (0) ;

////////////////////////////////////////////////////////////////// // This function updates the counters for release a fixed face. // The released face is the face currently on location nfacefixed.

////////////////////////////////////////////////////////////////// void change_counts_for_unfix_one_face

( int ♦pnewmatrix, int ^♦pnewmatchlist,

FENTCOUNT ♦pcnew, int matrixdim, int nfacefixed) { // input

/ / pnewmatrix - the new part's topological matrix, where

// the FENT contents have been changed to

// their representing integers.

// (integer array of matrixdim ^♦ matrixdim)

// pnewmatchlist the list of entries of the newmatrix that

// has been matched by the oldmatrix.

// (integer array of matrixdim)

// pcnew - the FENTCOUNT of all the faces.

// matrixdim the matrix dimension of both matrices

// nfacematched - the number of entries (faces) that has

12 APPENDIX B

// been fixed in matching the two matrices.

// output

// pcnew - the updated FENTCOUNT of all the faces

// get the to be unfixed face number and

// define the pointer to the FENT on the to be fixed face int facenew = ^♦ (pnewmatchlist+nfacefixed-1) ; int ♦pnewrow = pnewmatrix + (facenew-1) ^♦matrixdim ;

// first change the counters for the previously fixed faces for ( int i = 0 ; i < nfacefixed-1 ; i++ ) { int newcol = ^♦ (pnewmatchlist+i) ; int pnewv = ^♦ (pnewrow + newcol - 1) ; pcnew[newcol] .count [pnewv] ++ ; int igroup = fent_group (pnewv) ; pcnew[newcol] .gcount [igroup] ++ ;

}

// second change the counters for the unfixed faces by // adding the FENTCOUNT of the to be released face to that // of the remaining unfixed faces for ( i = 0; i < NumlntFent; i++ ) pcnew[0] .count [i] += pcnew[facenew] .count [i] ; for ( i = 0; i < NumlntFentGroup; i++ ) pcnew[0] .gcount [i] += pcnew[facenew] .gcount [i] ;

////////////////////////////////////////////////////////////////// // This function counts the number of FENTs in the given array of // intfent.

////////////////////////////////////////////////////////////////// FENTCOUNT count_fents ( int ^♦pintfentarray, int arraydim) { // input

// pintfentarray - the pointer to the intfent array

// arraydim - the dimension of the intfent array

// output

// return - FENTCOUNT, the FENT count of the

// input array

// define an FENT count and initialize it static FENTCOUNT fentc ; for ⁽ int j = 0; j < NumlntFent; j++ ) fentc.count [j] = 0

13 APPENDIX B for ( j = 0; j < NumlntFentGroup; j++ ) fentc.gcount [j] = ^ ;

// Count the numbers of fents in an array of intfent for ( int *i = pintfentarray ; i < pintfentarray+arraydim ;

1 + + ⁾ { fentc.count [*i] ++ ; fentc.gcount [fent_group(^i) ] ++ ; } return (fentc) ;

////////////////////////////////////////////////////////////////// // This function counts the number of FENTs in a specified set of

// an intfent array.

////////////////////////////////////////////////////////////////// FENTCOUNT count_specified_fents ( int ^♦pintfentarray, int ♦locationlist, int arraydim, int listdim) { // input

// pintfentarray - the pointer to the intfent array // locationlist - the locations in the intfent array

II that are to be included in counting

II arraydim - the dimension of the intfent array

II listdim - the dimension of the location list

II output

II return - FENTCOUNT, the FENT count of the

II selected elements in the input

// define an FENT count and initialize it static FENTCOUNT fentc ; for ( int j = 0; j < NumlntFent; j++ ) fentc.count [j] = 0 ; for ( j = 0; j < NumlntFentGroup; j++ ) fentc.gcount [j] = 0 ;

// Count the numbers of fents in an array of intfent for ( int ⁺i = locationlist ; i < locationlist+listdim ,- i++

) { int intfent = ^♦ (pintfentarray+ (^i) -1) ; fentc.count [intfent] ++ ; fentc .gcount [fent group (intfent) ] ++ ; } return (fentc) ;

14 APPENDIX B

// an intfent matrix.

// Note: The intfent matrix is symmetric. Only half of the matrix

// are included in the counting.

//////////////////////////////////////////////////////////////////

FENTCOUNT count_specified_fents_matrix ( int ♦pintfentmatrix, int ♦locationlist, int matrixdim, int listdim)

{ // input

// pintfentmatrix - the pointer to the intfent matrix // locationlist - the locations in the intfent matrix // that are to be included in counting

// matrixdim - the dimension of the intfent matrix // listdim - the dimension of the location list

// output

// return - FENTCOUNT, the FENT count of the

// selected elements in the input matrix

// define an FENT count and initialize it static FENTCOUNT fentc ; for ( int j = 0; j < NumlntFent; j++ ) fentc.count [j] = 0 ; for { j = 0; j < NumlntFentGroup; j++ ) fentc .gcount [j] = 0 ;

// Count the numbers of fents in an matrix of intfent for ( int i = 0 ; i < listdim ; i++ ) { int facenum = ^♦ (locationlist+i) ; int ^♦pintfentrow = pintfentmatrix + (facenum 1) ^♦matrixdim ;

// Note: only half of the symmetric matrix is counted => // the k is atarted from i (diagonal included) . for ( int k = i ; k < listdim ; k++ ) { int intfent = ^♦ (pintfentrow+ (^♦ (locationlist+k) ) -1) ; fentc.count [intfent] ++ ; fentc.gcount [fent_group (intfent) ] ++ ;

} return (fentc) ;

}

////////////////////////////////////////////////////////////////// // This function returns the penalty on a mismatched pair of FENT.

15 APPENDIX B

////////////////////////////////////////////////////////////////// int penalty_of_fent_pair ( int intfentl, int intfent2) { // input

// intfentl - the intfent of part 1

// intfent2 - the intfent of part 2

// output

// return - the penalty for the mismatched intfents.

// No penalty, if they are the same if (intfentl == intfent2) return (0) ;

// add the penalty for mismatching the individual intfent int indexvalue = 0; indexvalue += fent_weight (intfentl) ; indexvalue += fent_weight (intfent2) ;

// add the penalty for mismatching their fent group int fentgroupl = fent_group (intfentl) ; int fentgroup2 = fent_group(intfent2) ; if (fentgroupl 1= fentgroup2) { indexvalue += fent_group_weight (fentgroupl) ; indexvalue += fent group weight (fentgroup2) ;

} return (indexvalue)

}

////////////////////////////////////////////////////////////////// // This function returns the minimum penalty for the least possible // mismatches of two FENTCOUNT.

////////////////////////////////////////////////////////////////// int penalty_of_FENTCOUNT_jpair ( FENTCOUNT const fcfentcountl,

F E N T C O U N T c o n s t

&fentcount2 ) { / / input

// fentcount1 - the FENTCOUNT of part 1

// fentcount2 - the FENTCOUNT of part 2

// output

// return - the minimum penalty on the mismatches

// of two FENTCOUNTS.

// Now loop thru the FENTCOUNT,

// currently method uses the penalty of the minimum possible

// individual mismatches as the penalty plus the penalty of

// the minimum possible mismatched groups int indexvalue = 0; for ( int i = 1; i < NumlntFent; i++ )

16 APPENDIX B if ( fentcountl .count [i] != fentcount2.count [i] ) indexvalue += fent_weight (i) ^♦ abs ( fentcountl . count [i] f ent count 2.count [i] ) ; for ( i = 1; i < NumlntFentGroup; i++ ) if ( f entcountl. gcount [i] != fentcount2.gcount ti] ) indexvalue += fent_group_weight (i) ^♦ abs ( fentcountl . gcount [i] fentcount2.gcount [i] ) ; return (indexvalue) ; }

////////////////////////////////////////////////////////////////// int change_from_seperate_fent_sets ( int ntnew, int ntold, int nsnew, int nsold

)

{ // input

// ntnew - number of FENT in the whole set of the new part .

// ntold - number of FENT in the whole set of the old part .

// nsnew - number of FENT in the sub-set of the new part.

// nsold - number of FENT in the sub-set of the old part.

// output

// return - the change in the number of counts of the mismatch

// due to the separation of the sub-set from the

// whole set. int difforg = abs ( ntnew-ntold ) ; int diffsub = abs ( nsnew-nsold ) ; int diffnew = abs ( ntnew-nsnew-ntold+nsold ) ,- int change = diffsub + diffnew - difforg ; return ( change ) ;

/////////////////////////////////////////////////////// / / This function returns the minimum penalty for the least possible // mismatches of two FENTCOUNT.

////////////////////////////////////////////////////////////////// int increase_from_separate_a_pair_of_fent (

FENTCOUNT const &pcnew_total, FENTCOUNT c on s t &pcold_total, int intfentnew_separ, int intfentold_separ )

17 APPENDIX B

// input

// pcnew_total - the total FENTCOUNT of the new part

// pcold_total - the total FENTCOUNT of the old part

// intfentnew_separ - to be seperated FENT of the new part

// intfentold_separ - to be seperated FENT of the old part

// output

// return - the increase in the penalty due to the

// seperation of the FENTs. static int ntnew, ntold, igroup ; static int increase ; if (intfentnew_separ == intfentold_separ) { ntnew = pcnew_total.count [intfentnew_separ] ; ntold = pcold_total .count [intfentnew_separ] ; increase = fent_weight (intfentnew_separ) ^♦ change_from_seperate_fent_sets (ntnew, ntold, 1, l ) igroup = fent_group (intfentnew_separ) ; ntnew = pcnew_total.gcount [igroup] ; ntold = pcold_total.gcount [igroup] ; increase += fent_group_weight (igroup) ^♦ change_from_seperate_fent_sets (ntnew, ntold, 1,

1 ) else ( ntnew = pcnew_total.count [intfentnew_separ] ; ntold = pcold_total.count [intfentnew_separ] ; increase = fent_weight (intfentnew_separ) ^♦ change_from_seperate_fent_sets (ntnew, ntold, 1,

0 ) ntnew = pcnew_total.count [intfentold_separ] ; ntold = pcold_total.count [intfentold_separ] ; increase += fent_weight (intfentold_separ) ^♦ change_from_seperate_fent_sets(ntnew, ntold, 0, l ⁾ ; i f ( f e n t _ g r o u p ( i n t f e n t n e w _ s e p a r ) = = fent_group (intf entold_separ) ) { igroup = fent_group (intf ntnew_separ) ; ntnew = pcnew_total .gcount [igroup] ; ntold = pcold_total.gcount [igroup] ; increase += fent_group_weight (igroup) ^♦ change_from_seperate_fent_sets (ntnew, ntold, 1, 1 ) ;

} else { igroup = fent_group(intfentnew_separ) ;

18 APPENDIX B ntnew = pcnew_total.gcount [igroup] ; ntold = pcold_total.gcount [igroup] ; increase += fent_group_weight (igroup) ^♦ change_from_seperate_fent_sets (ntnew, ntold, 1, 0 ) ; igroup = fent_group (intfentold_separ) ; ntnew = pcnew_total.gcount [igroup] ; ntold = pcold_total .gcount [igroup] ; increase += fent_group_weight (igroup) ^♦ change_from_seperate_fent_sets (ntnew, ntold, 0, 1 ) ; }

} return (increase) ;

}

////////////////////////////////////////////////////////////////// int increase_from_separate_a_pair_of_fentcount (

FENTCOUNT const &pcnew_total, FENTCOUNT c on s t &pcold_total,

FENTCOUNT const &pcnew_separ, FENTCOUNT c on s t &pcold_separ ) { // input

// pcnew_total - the total FENTCOUNT of the new part

// pcold_total - the total FENTCOUNT of the old part

// pcnew_separ - to be seperated FENTCOUNT of the new part

// pcold_separ - to be seperated FENTCOUNT of the old part

// output

// return - the increase in the penalty due to the

// seperation of the FENTCOUNTs.

// Now loop thru the FENTCOUNT,

// currently method uses the penalty of the minimum possible

// individual mismatches as the penalty plus the penalty of

// the minimum possible mismatched groups int change, increase ; increase = 0; for ( int i = l; i < NumlntFent; i++ ) { change = change_from_seperate_fent_sets ( p c n e w _ t o t a l . c o u n t [ i ] , ρcold_total . count [i] , pcnew_separ . count [i] , pcold_separ . count [i] ' ;

19 APPENDIX B if ( change != 0 ) increase += fent weight (i) ^♦ change ; } for ( i = 1; i < NumlntFentGroup; i++ ) { change = change_from_seperate_fent_sets ( p c n e w _ t o t a l . g c o u n t [ i ] , pcold_total . gcount [i] , p c n e w _ s e p a r . g c o u n t [ i ] , pcold_separ .gcount [i] ) ; if ( change != 0 ) increase += fent_group_weight (i) ^♦ change ; } return (increase) ;

////////////////////////////////////////////////////////////////// // This function calculates the similar index for the given set of

// faces up to the specified fixed faces.

// Note: only half of the matrix are included in the calculation.

////////////////////////////////////////////////////////////////// int direct_calculate_similarity_index

( int ^♦pnewmatrix, int ^♦pnewmatchlist, int ♦poldmatrix, int ^♦poldmatchlist, int matrixdim)

{

// loop thru the faces and calculate the similarity index int indexvalue = 0 ; for (int iface = 0; iface < matrixdim; iface++) { int facenew = ^♦ (pnewmatchlist+iface) ; int faceold = ^♦ (poldmatchlist+iface) ; int ^♦pnewrow = pnewmatrix + (facenew - 1) ^♦matrixdim ; int ^♦poldrow = poldmatrix + (faceold - 1) ^♦matrixdim ;

II first from the mismatches of the fixed faces.

II Note: the diagonal terms of the matrix are always iNoRelation.

// therefore, they are skip in the loop. // also due to the symmetry, only half of the matrix are

// included in the calculation of the similarity index for (int icol = iface; icol < matrixdim; icol++) ( int newcol = ^♦ (pnewmatchlist+icol) int oldcol = ^♦(poldmatchlist+icol) int pnewv = ^♦ (pnewrow + newcol - 1) int poldv = ^♦(poldrow + oldcol - 1) if ( pnewv != poldv ) indexvalue penalty_of fent_pair (pnewv,poldv) ;

20 APPENDIX B

} return (indexvalue) ;

////////////////////////////////////////////////////////////////// // This function calculates the similar index for the given set of // faces up to the specified fixed faces.

////////////////////////////////////////////////////////////////// int calculate_partial_similarity_index

( int ♦pnewmatrix, int ^♦pnewmatchlist, FENTCOUNT ♦pcnew, int ♦poldmatrix, int ^♦poldmatchlist, FENTCOUNT ^♦pcold, int matrixdim, int nfacefixed)

// loop thru the faces and calculate the similarity index int indexvalue = 0 ; for (int iface = 0 ; iface < nfacefixed; iface++) { int facenew = ^♦ (pnewmatchlist+iface) ; int faceold = ^♦ (poldmatchlist+iface) ; int *pne row = pnewmatrix + (facenew - 1) ♦matrixdim ; int ^♦poldrow = poldmatrix + (faceold - 1) ^♦matrixdim ;

// first from the mismatches of the fixed faces. // Note: the diagonal terms of the matrix are always iNoRelation.

// included in the calculation of the similarity index for (int icol = iface ; icol < nfacefixed; icol++) { int newcol = ^♦(pnewmatchlist+icol) ; int oldcol = ^♦(poldmatchlist+icol) ; int pnewv = ^♦ (pnewrow + newcol - 1) ; int poldv = ♦(poldrow + oldcol - 1) ; if ( pnewv != poldv ) indexvalue += penalty_of fent_pair(pnewv,poldv) ;

// use count_specified_fents to initialize the FENTCOUNT of

// the facenew and faceold and to count the numbers of fents of

// the columns of the unfixed faces int listdim = matrixdim - nfacefixed ; pcnew[facenew] = count_specified_fents ( pnewrow, pnewmatchlist+nfacefixed, matrixdim, listdim ) pcold[faceold] = count_specified_fents ( poldrow, poldmatchlist+nfacefixed, matrixdim, listdim )

21 APPENDIX B

// Now loop thru the FENTCOUNT of the facenew and faceold // and calculate the penalty of their mismatches. indexvalue += penalty_of_FENTCOUNT_pair ( pcnew[facenew] , pcold[faceold] ) ;

// use count_specified_fents_matrix to get the FENTCOUNT

// of the unfixed faces of both matrices int listdim = matrixdim - nfacefixed ; pcne [0] = count_specified_fents_matrix ( pnewmatrix, pnewmatchlist+nfacefixed, matrixdim, listdim ) ; pcold[0] = count_specified_fents_matrix ( poldmatrix, poldmatchlist+nfacefixed, matrixdim, listdim ) ;

// Finally calculate the penalty of the FENTCOUNT // of the unfixed faces of the two matrices, indexvalue += penalty_of_FENTCOUNT_pair ( pcnew[0] , pcold[0]

) ;

// Return the amount of penalty on the two partially fixed // part matrices as its minimum possible similarity index, return (indexvalue) ;

////////////////////////////////////////////////////////////////// // This function calculates the increase in the similar index for // the mismatches of the given facenew and faceold.

////////////////////////////////////////////////////////////////// int increase_on_similarity_index

( int ♦pnewmatrix, int ^♦pnewmatchlist, int facenew, int ♦poldmatrix, int ^♦poldmatchlist, int faceold, FENTCOUNT ♦pcnew, FENTCOUNT ^♦pcold, int ^♦currentvalue, int matrixdim, int nfacefixed, int ^♦minimumvalue)

{

// loop thru the faces to see how much increase is in the

// current value. int increase = 0 ; int ^♦pnewrow = pnewmatrix + (facenew-1) ^♦matrixdim ; int ^♦poldrow = poldmatrix + (faceold-1) ^♦matrixdim ;

// first loop thru the previously fixed faces and calculate // the increase for the mismatches between the chosen columns for ( int i = 0; i < nfacefixed; i++ ) { int newcol = ^♦ (pnewmatchlist+i) ; int oldcol = ^♦ (poldmatchlist+i) ;

22 APPENDIX B int pnewv = ^♦ (pnewrow + newcol - 1) ; int poldv = ♦ (poldrow + oldcol - 1) ; if ( pnewv ! = poldv ) {

FENTCOUNT pcnewcol = pcnew[newcol] ;

FENTCOUNT pcoldcol = pcold[oldcol] ; increase += increase_from_separate_a_pair_of_fent ( pcnewcol, pcoldcol, pnewv, poldv ) ;

}

// use count_specified_fents to initialize the FENTCOUNT of

// the newly seleted face and count the numbers of fents of

// the undetermined columns on the to be fixed rows int listdim = matrixdim - nfacefixed ; pcnew[facenew] = count_specified_fents ( pnewrow, pnewmatchlist+nfacefixed, matrixdim, listdim ) ; pcold[faceold] = count_specified_fents ( poldrow, poldmatchlist+nfacefixed, matrixdim, listdim ) ; increase += increase_from_separate_a_pair_of_fentcount ( pcnew[0] , pcold [0] , pcnew [facenew] , pcold[faceold]

// Return the amount of increase in the similarity index for // the matching of the chosen two faces, facenew in new part // and the faceold in the old part . return (increase) ;

23 APPENDIX C

1/ II

II Example of bendline detection with comments. This module // // contains BM_PART: :auto_bend() implementation. //

// //

111111 U 11111111111111111111111111 ! 11111 n I n 111 ! 111111 ! 111 ! 11111

/ *

*********** ********** ********** ********** *********** ********* This is a major high-level Bend Model function for part design and construction. Its purpose is to create bendlines between faces so that the part becomes connected (if possible) .

This function was created to facilitate part creation. Normally, a third-party CAD program is used to draw a part. Since Bend Model does not have any control over the CAD system, for our purposes this drawing is just a set of edges. Therefore this drawing has to be analyzed in order to detect the structure of the part. After that we can create faces and bendlines of the part. However, the input drawing is often ambigious - bendlines are not uniquely defined. In that case we use a number of heuristics to create bendlines. This heuristics define preference criterias for picking one outcome when many are possible.

Since this problem is common to many applications this function is part of the Bend Model. Moreover, implementing this function in Bend Model will simplify the face detection software that has to be part of the CAD system. In this function we assume that edges of faces are sorted according to the left-hand side rule in a way that is consistent with the plane normal (which defines the plane orientation) . Basically, this function assumes that every plane is correct by itself, when viewed in isolation. However, this function does not require that the orientation of adjacent planes be correct with respect to the bendlines between them (there is a Bend Model function for fixing that later) .

This function can be used to generate bendlines for both 3D-version of the part and the flat version of the part. However, the part can have either the flat or 3D version, but not both. In other words, either all 3D-version of 3D-bodies must be NULL (in which case we have a flat version) , or all flat versions must be NULL. */ APPENDIX C

*********** ********** ********** ********** ********** **** ***** Algorithm/Process :

The objective of the algorithm is the create a minimum number of bendlines (because we want to change the part as little as possible) so that the entire part becomes connected (ie. all faces all connected via bendlines - we don't want any stand-alone faces hanging around), given the restriction that we don't create loops in the topology graph of the part.

It is not required that the input part contain no bendlines. AutoBend takes into account existing bendlines and adds new bendlines to the part. On the top level, this algorithm performs a version of the maximum spanning tree algorithm.

At any point in time, it has a current connected component and a set of fringe faces (that are part of the connected component) . These fringe faces have contacts with faces outside of the connected component. During one iteration, one new face that is outside of the connected component will be added to the connected component . Among the faces outside the connected component that are in contact with a fringe face, AutoBend picks one that has the largest contact .

This heuristic of using the largest contact is based on the fact that in sheet-metal manufacturing, the larger the contact the better, usually. However, when there is a tie - ie. there are several faces outside the connected component that have the same largest contact with a fringe face in the connected component, we use another heuristic. This heuristic picks a face that would minimize the diameter of the part . The diameter of the part is the largest distance between any two faces in the topology graph of the part.

*********** ********** ********** ********** ********** **********

Technical notes :

- This function probably does not work well if the part thickness is not zero.

- Note that this function creates bendline information that is consistent with facel in the algorithm (mainly, facel normal) . However, face2 normal can be inconsistent with this information, but that has to be fixed later.

See comments in PART.HXX.

********** ********** ********** ********** ********* ********* */ APPENDIX C

#include <stdio.h> #include <stdlib.h>

#include "GLOBAL.HXX" #include "ENTITY.HXX" #include "PLANE.HXX" ^include "CYLINDER.HXX" #include "LINE.HXX" #include "LOOP.HXX" #include "2D_BODY.HXX" #include "FORMING.HXX" #include "FACE.HXX" #include "BEND_OP_R.HXX" #include "BEND_PRP.HXX" #include "BENDLINE.HXX" #include "PART.HXX" #include "DBM.HXX"

/*

This function computes all contacts between the given two faces.

A contact is when two faces are touching each other (ie. within some very small distance tolerance) and the line separating them has length greaterthan 0 (ie. the common line is not a point, but a real line) .

Purpose :

To compute a list of potential bendlines between given two faces.

Assumptions :

- both faces have a non-NULL 3D version that is up to datewhich will be used for the computation.

- all loops in either faces are closed.

Requirements :

- both faces must be in the same part. Returns :

- a list of contacts

This returned list has a structure : node 0 : obj is "number of contacts" node 2i-l : obj is "head of the list on facel side, of

3 APPENDIX C

contact i" node 2i : obj is "head of the list on face2 side, of contact i" where i is in [1, "number of contacts"] .

Note that each contact is a "closed" contact, ie . there are no gaps in the contact.

Each list on either side of the contact is a list of lines (actually pointers to lines) such that the order of lines in the list is opposite to sweep direction (currently XZY sweep) .

This is what a list of contacts looks like :

# of contacts -> listl .facel .head -> listl .face2.head -> ... -> list i. facel.head -> list i.face2.head.

I I

I !

V V

V V linel in facel linel in face2 linel in facel linel in face2

I I I I

I I

V V

I I

V V

V V line i in facel line j in face2 line k in facel line 1 in face2

Note that in this function a co-linear bendline of k bends is considered as k separate contacts.

Notes :

- By default we will consider all lines in both faces as candidates for a bendline.

However, if the user already known what the bendline is (or actually, knows what the bendline APPENDIX C

edges are) he can use the name of these lines to instruct this function to consider only those lines. This function will check if any of the lines in facel or face2 (faces 1 and 2 are checked separately) have a name that is euqal to the idx of the other face. If yes, only these lines will be considered as candidates for a bendline on the side of that face.

*/ static int BM_compute_all_face_to_face_contacts ( BM_FACE & facel /^♦ IN ^♦/, BM_FACE & face2 /^♦ IN ^♦/, BM LINKED_LIST_NODE ^♦♦list_of_contacts /^♦ OUT ^♦/)

{

// initialize global variables.

// this is the contacts list we will be computing, initially list is NIL.

♦list_of_contacts = NULL ;

// this points to the beginning of the contacts list. // it does not contain the number of contacts. // ie. points to "list_of_contacts [1] " . BM_LINKED_LIST_NODE ^♦contacts = NULL ; double ^♦line_start_coordinates = NULL ; // used when the set of lines is sorted

BM_LINE ♦♦line_pointers = NULL ; // an array of lines (actually pointers to lines) used to store

// lines being swept.

// once lines are sorted according to the XZY order, we will scan this array from the end

// to the beginning -> this implements the sweep.

// at the same time, we maintain a list of open lines, this is done by leaving line

// pointers in this array "behind" as they are when we move forward in the array.

// to erase a line from the open list, we simple set the pointer in the array to NULL.

// endpoints of lines in facel and face2 that are larger with respect to XZY ordering.

// used when processing contacts. BM_POINT p_facel, p_face2 ;

// some variables used in many places, long i, j ; BM_POINT p ; APPENDIX C

/*

********** ********** ********** ********** **********

********** ********** ********** **********

Qualification check.

********** ********** ********** ********** **********

********** ********** ********** **********

*/

// both faces must be in the same part if (facel.get_part () == NULL | | facel .get_part () != face2.get_part () ) return 0 ; d o u b l e l o c a l _ t o l e r a n c e =

(facel . get_part ( ) ) - >get_distance_tolerance ( ) ;

// first try 3D-versions of faces, if both of them are empty, try flat versions.

BM_2D_B0DY ^♦bodyl = facel .get_current_3D_version () ; BM_2D_B0DY ♦body2 = face2.get_current_3D_version() ; if (NULL == bodyl && NULL == body2) { bodyl = facel .get_flat () ; body2 = face2.get flat() ;

} // if either of the 3D-bodies is NULL we are successfully done. if (NULL == bodyl | | NULL == body2) return 1 ;

// both faces must have a plane as the underlying surface BM_SURFACE ^♦surf acel = bodyl ->get_surf ace { ) ; BM_SURFACE ♦surface2 = body2->get_surf ace ( ) ; if (NULL == surface2 | j NULL == surface2) return 0 ; if ( ! sur f ace2 - > is ( BM_T Y P E_ P L ANE ) | j ! surface2->is (BM_TYPE_PLANE) ) return 0 ;

// if any of the faces has an empty bloop, we are successfully done

BM_L00P ^♦bloopl = bodyl - >get_bloop ( ) ; BM_LOOP ^♦bloop2 = body2->get_bloop ( ) ; if (NULL == bloopl ] | NULL == bloop2) return 1 ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

Here we will do a very quick check if these faces can be in contact at all.

Here we check if bloop of face2 touches (at more than one point) or intersects the plane of facel.

This will save some time when these two faces obviously cannot be in contact.

********** ********** ********** ********** ********** APPENDIX C

********** ********** ********** ********** */

// these variables will be used to do a quick check if these t o

// faces can possible touch each other double distance_plane_to_bloopl, distance_plane_to_bloop2, distance_plane_to_bloop3, distance_plane_to_bloop4 ; int distance_plane_to_bloop_count ; if (! bloop2->is_bbox_up_to_date () ) bloop2->recompute__bbox() i f (

BM_distance_between_point_and_plane (bloop2->get_bbox_pl () ,

(BM_PLANE&) (^♦surfacel) , &distance_plane_to_bloopl NULL) ) return 0 ; i f (

BM_distance_between_point_and_plane (bloop2->get_bbox_p20 ,

(BM_PLANE&) (^♦surfacel) , &distance_plane_to_bloop2 NULL) ) return 0 ; i f (

BM_distance_between__point_and_plane (bloop2->get_bbox_p30 ,

(BM_PLANE&) (^♦surfacel) , &distance_plane_to_bloop3 NULL) ) return 0 ; i f (

BM_distance_between_point_and_plane (bloop2->get_bbox_p4 () ,

(BM_PLANE&) (^♦surfacel) , &distance_plane_to_bloop4 NULL) ) return 0 ; distance_plane_to_bloop_count = 0 ; if (fabs (distance_plane_to_bloopl) >= local_tolerance) distance_plane_to_bloop_count++ ; if (fabs (distance_plane_to_bloop2) >= local_tolerance) distance_plane_to_bloop_count++ ; if (fabs (distance_plane_to_bloop3) >= local_tolerance) distance_plane_to_bloop_count++ ; if (fabs (distance_plane_to_bloop4) >= local_tolerance) distance_plane_to_bloop_count++ ; if (distance_plane_to_bloop_count > 2) {

// at most one of the bloop corners is on the plane. // that means, bbox itself is not in contact with facel. // however, it could be that different corners of bbox are on different sides of the plane.

// here we will return if all bloop2 points are on one side of planel

// ie. these two faces cannot be in contact

//

// note : at most one of the 4 distance_plane_to_bloop-i can be zero.

//

// in the next line we pick "distance_plane_to_bloopl" APPENDIX C

for checking not because it is special,

// but because any of the 4 would do. if (fabs (distance_plane_to_bloopl) >= local_tolerance) { i f

(distance_plane_to_bloopl⁺distance_plane_to_bloop2 >= 0.0 && distance_plane_to_bloopl^distance_plane_to_bloop3 >= 0.0 && distance_plane_to_bloopl^distance_j?lane_to_bloop4 >= 0 . 0 ) return 1

else { i f

(distance_plane_to_bloop2^distance_plane_to_bloop3 >= 0.0 && distance_plane_to_bloop2^distance_plane_to_bloop4 >= 0.0) return 1

}

} // if the count is 0, 1 or 2, it means that 2, 3 or 4 bbox corners are on the plane.

// that means, it is possible that there is a contact.

/*

Now we check the same thing for bboxl with respect to face2.

*/ if (! bloopl->is_bbox_up_to_date () ) bloopl->recompute_bbox () i f ( !

BM_distance_between_point_and_plane (bloopl->get_bbox_pl () ,

(BM_PLANE&) (^♦surface2) , &distance_plane_to_bloopl, NULL) ) return 0 ; i f ( !

BM_distance_between_point_and_plane (bloopl->get_bbox_p2 () ,

(BM_PLANE&) (^♦surface2) , &distance_plane_to_bloop2, NULL) ) return 0 ; i f ( !

BM_distance_between_point_and_plane (bloopl->get_bbox_p3 () ,

(BM_PLANE&) (^♦surface2) , Scdistance_plane_to_bloop3, NULL) ) return 0 ; i f ( !

BM_distance_between_point_and_plane (bloopl->get_bbox_p4 () ,

(BM_PLANE&) (^♦surface2) , &distance_plane_to_bloop4, NULL) ) return 0 ; distance_plane_to_bloop_count = 0 ; if (fabs (distance_plane_to_bloopl) >= local_tolerance) distance_plane_to_bloop_count++ ; if ⁽fabs (distance_plane_to_bloop2) >= local_tolerance) distance_plane_to_bloop_count++ ;

8 APPENDIX C

if (fabs (distance_plane_to_bloop3) >= local_tolerance) distance_plane_to_bloop_count++ ; if (f bs (distance_plane_to_bloop4) >= local_tolerance) distance_plane_to_bloop_count++ ; if (distance_plane_to_bloop_count > 2) {

// at most one of the bloop corners is on the plane. // here we will return if all bloopl points are on one side of plane2

// ie. these two faces cannot be in contact if (fabs (distance_plane_to_bloopl) >= local__tolerance) { i f

(distance_plane_to_bloopl⁺distance_plane_to_bloop2 >= 0.0 && distance_plane_to_bloopl^Λdistance_plane_to_bloop3 >= 0.0 && distance_plane_to_bloopl^distance_plane_to_bloop4 >= 0.0) return 1 ; else { i f

(distance_plane_to_bloop2^distance_plane_to_bloop3 >= 0.0 && distance_plane_to_bloop2^distance_plane_to_bloop4 >= 0.0) return 1 ;

) ^J

/*

********** ********** ********** ********** **********

********** ********** ********** **********

Count the total number of lines in both faces. We need to allocate memory.

Note : we only count edges that are lines, because only lines can make bendlines.

Note : we will check if the user has specified bend edges. User can restrict which edges will be considered as possible bend edge by setting the name of those edges.

********** ********** ********** ********** **********

********** ********** ********** **********

*/ long facel_has_user_defined_bend_edges = 0 ; // the number of user-speicifed bend edges in facel long face2_has__user_defined_bend_edges = 0 ; // the number of user-speicifed bend edges in face2 long facel_line_count = 0, face2_line_count = 0 ; long line count ; // sum of lines in faces 1 and 2 APPENDIX C

participating in the sweep. BM_LOOP ^♦holes ; BM_EDGE ^♦edges ; for (edges = bloopl->get_first_edge () ; edges ; edges = (BM_EDGE ^♦) edges->next () ) { if (! edges-_>is(BM_ENTITY_TYPE_LINE) ) continue ; if (edges- >get_name ( ) == face2. get_idx ( ) ) facel_has__user_defined_bend_edges++ ; facel line count++ ;

} for (holes = bodyl->get_first_hole () ; holes ; holes =

(BM_LOOP ^♦) holes->next () ) { for (edges = holes->get_first_edge () ; edges ; edges = (BM_EDGE ^♦) edges->next () ) { if (! edges-_>is(BM_ENTITY_TYPE_LINE) ) continue ; if (edges->get_name () == face2.get_idx() ) facel__has_user_defined_bend_edges++ ; facel line count++ ;

} ^{~ "}

} for (edges = bloop2->get_first_edge () ; edges ; edges =

(BM_EDGE ♦ ) edges - >next ( ) ) { if ( ! edges- >is (BM_ENTITY_TYPE_LINE) ) continue ; i f ( edge s - >ge t_name ( ) = = f acel . get_idx ( ) ) face2_has_user_def ined_bend_edges++ ; face2 line count++ ;

} ^{~ "} for (holes = body2->get_first_hole () ; holes ; holes =

(BM_LOOP ^♦) holes->next () ) { for (edges = holes->get_first_edge () ; edges ; edges =

(BM_EDGE ^♦) edges->next () ) { if (! edges->is(BM_ENTITY_TYPE_LINE) ) continue ; if (edges->get_name () == facel.get_idx() ) face2_has_user_defined_bend_edges++ ; face2 line count++ ;

// if there are no lines, we are successfully done if (facel_has_user_defined_bend_edges) facel_line_count facel_has_user_defined_bend_edges ; if ⁽face2_has_user_defined_bend_edges) face2_line_count face2_has_user_defined_bend_edges ; if (facel_line_count < 1 | | face2_line_count _< 1) return 1 ; line_count = facel_line_count + face2_line_count ;

/*

********** ********** ********** ********** **********

********** ********** ********** ********** Allocate memory.

10 APPENDIX C

********** ********** ********** ********** ********** ********** ********** ********** **********

*/ if (NULL == (line_pointers = new BM_LINE⁺ [line_count] ) ) return

0 ; if (NULL == (line_start_coordinates = new double [line count] ) )

{ delete [] line_pointers ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

First, fill in the line pointers array.

If required, take only user-specified edges.

********** ********** ********** ********** **********

********** ********** ********** **********

*/

11 APPENDIX C

edges->get_name () != facel.get_idx() ) continue ; line pointers [i++] = (BM LINE ^♦) edges ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

Sort the array of lines by X-coordinate, then by Z-coordinate and then by Y.

********** ********** ********** ********** **********

********** ********** ********** **********

*/

{ // this is a block of code for sorting the array

// first, construct the array of points and lines associated with them for (i = 0 ; i < line_count ; i++) {

P BM_get_greater_point_by_XZY( (line_pointers [i] ) ->get_startpt () , (li ne_pointers ti] ) ->get_endpt () ) ; line start coordinates [i] = p.X() ;

} ^" // sort by X

DBM_QuickSort_double ( line_start_coordinates , line_count , ( long ^♦ ) line_pointers ) ;

// sort by Z double first_in_set_Z = line_start_coordinates [0] ; long set_Z_size = 1, first_in_set_Z_idx = 0 ;

P BM_get_greater_point_by_XZY( (line_pointers [0] ) ->get_startpt () , (li ne_pointers [0] ) ->get_endpt () ) ; line_start_coordinates [0] = p.Z() ; for (i = 1 ; i <= line_count ; i++) { if (i < line_count) { li {

i f ⁽ s e t _ Z _ s i z e > l ) DBM_QuickSort_double ⁽line_start_coordinates + first_in_set_Z_idx, set_Z_size, (long ^♦) line_pointers ₊

12 APPENDIX C

f irst_in_set_Z_idx) ;

// sort by Y long set_Z_end = f irst__in_set_Z_idx + set_Z_size ; d o u b l e f i r s t _ i n _ s e t _ Y line_start_coordinates [f irst_in_set_Z_idx] ; long set_Y_size = 1, f irst_in_set_Y_idx f irst_in_set_Z_idx ;

B M _ P 0 I N T pY (BM_get_greater_point_by_XZY ( (line_pointers [f irst_in_set_Z_idx] ) ->get_startpt () , (line_pointers [f irst_in_set_Z_idx] ) ->get_endpt () ) ) line_start_coordinates [f irst_in_set_Z_idx] = pY.YO ; for (j = f irst_in_set_Z_idx + 1 ; j <= set Z end ; j++)

{ if (j < set_Z_end) { p Y

BM_get_greater_point_by_XZY ( (line_pointers fj] ) ->get_startpt () , (li ne_pointers [ j ] ) - >get_endpt ( ) ) ; if ( 1 i ne_s t ar t _c oor di na t e s f j ] == f irst_in_set_Y) { set_Y_size++ ; line_start_coordinates [j] = pY.YO ; continue ;

}

} // else, we have passed the end of the Z-sequence i f ( s e t _ Y _ s i z e > 1 ) DBM_QuickSort_double (line_start_coordinates + first_in_set_Y_idx, set_Y_size, (long ^♦) line_pointers + first_in_set_Y_idx) ; if (j < set_Z_end) { set_Y_size = 1 ; f irst_in_set_Y_idx = j ; f irst_in_set_Y = line_start_coordinates [j] ; line start coordinates [j ] = pY.YO ;

} ^" } if (i < line_count) ( set_Z_size = 1 ; first_in_set_Z_idx = i ; first_in_set_Z = line_start_coordinates [i] ; line start coordinates fi] = p.Z() ;

. ^{} " '}

} // end of the block of code for sorting the array

13 APPENDIX C

/*

********** ********** ********** ********** **********

********** ********** ********** **********

We don't need line_start_coordinates any more. Delete it. ********** ********** ********** ********** ********** ********** ********** ********** **********

*/ if (line_start_coordinates) { delete [] line_start_coordinates ; line start coordinates = NULL ;

} ^~

/*

********** ********** ********** ********** **********

********** ********** ********** **********

********** ********** ********** ********** **********

********** ********** ********** **********

Main loop. Sweep the two faces.

Always keep track of open lines (lines whose startpoint we have seen, but have not yet seen an endpoint - actually, not start-and-end-points, but endpoints larger and smaller with respect to XZY ordering) .

Initially the set of open lines is empty. During every iteration we will do :

- remove all lines that were open, but should be closed now,

- check if the current line overlaps with open lines,

- add this line to the set of open lines

*/

// these represent lists of contacts BM_LINKED_LIST_NODE ^♦facel_list, ♦face2_list ; BM_LINKED_LIST_NODE *last_facel_list, ♦last_face2 list j BM_LINKED_LIST_NODE ^♦left, ^♦right ;

BM_LINKED_LIST_NODE ^♦new_left = NULL, ♦new_right = NULL BM_LINE ^♦line_in_left, ^♦line_in_right ;

// these are the two lines we will be comparing BM_2D_B0DY *face_of_linel, ^♦face_of_line2 ; BM_LINE ^♦line_in_facel, ^♦line_in_face2 ;

14 APPENDIX C

BM_LINE ^♦line ; double k, dotp ;

BM_POINT p_open ; BM_VECTOR V ; long num_of_contacts = 0 ; long open_line_idx ; for (i = line_count - 1 ; i >= 0 ; i--) {

// initially face_of_linel is the face of the current line.

// however, later we will sort the current-sweep-line and current-open-line pointers. f a c e _ o f _ l i n e l ( ( line_pointers [i] ) - >get_loop ( ) ) - >get_2D_body ( ) ; p BM_get_greater_point_by_XZY( (line_pointers fi] ) ->get_startpt 0 , (li ne_pointers [i] ) ->get_endpt () ) ;

// compare this current line against all open lines in order to detect overlaps, ie. bendlines. for (open line_idx = i + 1 ; open_line_idx < line_count ; open__line_idx++) line = line_pointers [open_line_idx] ; if (NULL == line) {

// time-saving trick, if this is the last pointer, there is no need to try it again.

// reduce the index. if (open_line_idx == line_count - 1) { --line_count ; break ;

} continue ;

}

// ********** ********** ********** ********** **********

// check if this open line should really be removed.

// it should be removed if its endpoint is larger than p (ie. the

// startpoint of this line) .

// ********** ********** ********** **********

********** p _ o p e n

BM_get_smaller_j?oint_by_XZY (line- >get_startpt ( ) , line- >get_endpt ( ) ) if (BM_compare_points_by_XZY(p_open,p) ) { // time-saving trick.

15 APPENDIX C

// if the line we are removing from the open list is the last one, reduce the index

// instead of setting the pointer to NULL, next time we don't even try this element

// in the array, because it would be beyond the boundary if (open_line_idx == line_count - 1) { --line_count ; break ;

} line_pointers [open_line_idx] = NULL ; continue ; }

// ********** ********** ********** **********

**********

// first, check if both lines are in the same face. if yes, skip it.

// ********** ********** ********** **********

********** face_of_line2 = (line->get_loop() ) ->get_2D_body() if (face of linel == face of line2) continue ;

// ********** ********** ********** **********

**********

// Check if this line and the open line overlap. // ********** ********** ********** **********

**********

// First check if the startpoint is on the (open) line segment. if (! line->is_point_on_line (p, &k) ) continue ;

// Note that parameter k has to be in [0,1] now, because

// the line "line" is open still, ie. it startpoint is "before" p

// and endpoint is "after" p.

// Now check if vectors are parallel. v = (line->get_v() ) ^♦ ( (line_pointers [i] ) ->get_ () )

/ if (v.LenO > BM_PRECISION) continue ;

// Last thing, it could be that only endpoints of the lines overlap. d o t p = ( l i n e - > g e t _ v ( ) ) % ( (line_pointers [i] ) ->get_v() ) ,- if ⁽dotp < 0.0) { // lines are opposite

// startpoint-startpoint/endpoint-endpoint should not match i f ^{( (} line - >get_startpt ( ) )

16 APPENDIX C

( (line_pointers [i] ) ->get_startpt () ) | ]

( l i n e - > g e t _ e n d p t ( ) ) = =

( (line_pointers [i] ) ->get_endpt () ) ) continue ; else ( // lines have the same direction

// startpoint-endpoint/endpoint-startpoint should not match i f ( ( line - >get_endpt ( ) ) = =

( (line_pointers fi] ) ->get_startpt () ) | |

( l i n e - > g e t _s t a r t p t ( ) ) = = ( (line_pointers [i] ) ->get_endpt () ) ) continue ;

// ********** ********** ********** **********

**********

// Ok, these two lines overlap.

// First, sort lines according to which face they are in

********** ********** ********** **********

//

********** i f ( & f a c e 1 = = ( B M _ F A C E ^♦ ) face_of_linel->get_3D_body() ) { line_in_facel = line_pointers fi] ; line_in_face2 = line ; p_facel = p ; p _ f a c e 2 =

BM_get_greater_point_by_XZY(line->get_startpt 0 , line->get_endpt () )

/ else { line_in_facel = line ; line_in_f ace2 = line_pointers [i] ; p _ f a c e l =

BM_get_greater_point_by_XZY(line->get_startpt 0 , line->get_endpt () )

; p_face2 = p ;

// ********** ********** ********** **********

**********

// Scan all known contacts and check if this contact should be

// added to an existing contact

// ********** ********** ********** **********

********** for (j = 0, last_face2_list = NULL num_of_contacts ; j++) { if (NULL == last_face2_list) { last_facel list = contacts ;

17 APPENDIX C

last_face2_list = last_facel_list->next ()

else { last_f acel_list = last_f ace2_list->next 0 last_face2_list = last_facel_list->next 0 ;

} left = (BM_LINKED_LIST_NODE ^♦) last_facel_list->get_obj () ; right = ( BM_LINKED_LIST_NODE ^♦) last_face2_list->get_obj () ; line_in_left = (BM_LINE ^♦) lef t->get_obj ( ) ; line_in_right = (BM_LINE ^♦) right- >get_obj ( ) ;

// Check if the line of this existing contact overlaps the new contact .

// condition is : point on the line and vectors are parallel. i f ( ! line_in_left->is_point_on_line (p_f acel, &k) ) continue ; v = ( 1 ine_i n_l e f t - >ge t _v ( ) ) ^♦

(line_in_facel->get_v() ) ; if (v.Len() > BM_PRECISION) continue ;

// Check that lines in facel and face2 and exact extensions of left and right in this contact

// this is because we want every contact in the list of contacts to be a "closed" contact

// note it could be that one of the new lines in already in this contact.

// then we don't need to check. if (line_in_left != line_in_facel) { dotp = (line_in_left->get_v() ) % (line_in_facel->get_v() ) ; if (dotp < 0.0) { // lines are opposite // startpoints must match if ( (line_in_left->get_startpt 0 ) != (line_in_facel->get_startpt () ) &&

(line_in_left->get_endpt () ) != (line_in_facel->get_endpt () ) ) continue ; else ( // lines have the same direction // endpoints must match if { (line_in_left->get_endpt () ) != (line_in_facel->get_startpt () ) &&

(line_in_left->get_startpt () ) != (line_in_facel->get_endpt () ) ) continue ;

18 APPENDIX C

if (line_in_right != line_in_face2) { dotp = (line_in_right->get_v () ) % (line_in_face2->get_v() ) ; if (dotp < 0.0) { // lines are opposite // startpoints must match if ( (line_in_right->get_startpt () ) != (line_in_face2->get_startpt () ) &&

{1ine_in_right->get_endpt () ) != (line_in_face2->get_endpt () ) ) continue ; else { // lines have the same direction // endpoints must match if ( (line_in_right->get_endpt () ) != (line_in_face2->get_startpt () ) &&

(line_in_right->get_startpt 0 ) != (Iine_in_face2->get_endpt () ) ) continue ;

// everything is fine, add lines to this contact .

// note that is could be that the line is already in the contact . if (line_in_left != line_in_facel) { if (NULL ==, (new_left = new BM_LINKED_LIST_NODE (line_in_facel, ULL, left) ) ) goto failure ; last_facel_list->set_obj ( (BM_BASIC ^♦) new_left) ; new_left = NULL ; if (line_in_right != line_in_face2) { if (NULL == (new_right = new

BM_LINKED_LIST_NODE (line_in_face2,NULL, right) ) ) goto failure ; last_face2_list->set_obj ( (BM_BASIC ^♦) new_right) ; new right = NULL ;

} ^"

// now we have to break out of here, we added this contact to an existing contact. break ;

} // check if we checked all contacts and could not find a match if (j < num_of_contacts) continue ;

// ********** ********** ********** **********

19 APPENDIX C

**********

// Create a new contact.

// ********** ********** ********** **********

**********

// First, create nodes for both lines. i f ( N U L L = = ( l e f t = n e w

BM_LINKED_LIST_NODE (line_in_f acel) ) ) goto failure ; i f ( N U L L = = ( r i g h t = n e w

BM_ INKED_LIST_NODE ( line_in_f ace2 ) ) ) { delete left ; goto failure ;

} // Create lists for both sides of the contact. if (NULL == ( face2_list = new

BM_LINKED_LIST_NODE( (BM_BASIC ^♦) right))) { delete left ; delete right ;

TOto failure ; if ?^c (NULL == (facel_list = new

BM_LINKED_LIST_NODE( (BM_BASIC ^♦) left, NULL, face2_list) ) ) { delete left ; delete right ; delete face2_list ; goto failure ;

// add the new contact to the list of contacts if (num_of_contacts < 1) contacts = facel_list ; else last_face2_list->add_after(facel_list) ;

++num of contacts ; } ^{" "}

// we use i to index line_pointers [] array. // since i will be decremented shortly, line_pointers [i] will be "implictly" put on

// the open list of sweep lines.

/*

********** ********** ********** ********** ********** ********** ********** ********** ********** free memory if necessary

* if (line_start_coordinates) delete f] line_start_coordinates if (line_pointers) delete [] line_pointers ;

20 APPENDIX C

// finally, create a node for the number of contacts if (NULL == (left = new BM_LINKED_LIST_NODE ( (BM_BASIC ♦) num_of_contacts,NULL,contacts) ) ) goto failure ; ♦list_of_contacts = left ; return 1 ; failure :

// free memory if necessary if (line_start_coordinates) delete f] line_start_coordinates

; if (line_pointers) delete [] line_pointers ;

if (new_left) delete new_left ; if (new_right) delete new_right ; return 0 ; }

/*

This function scans the list of contacts between two faces

(most likely created by BM_compute_all_face_to_face_contacts (...)) and removes all, but the largest contact.

Basically, it "filters" the list of contacts and removes those contacts that do not belong to the longest contact. Note that, in general, the longest contact is a list of simultaneous bendlines. Thus, the basic structure of the contact list remains the same.

For example, the first node in the list will be the number of contacts in the longest contacts.

21 APPENDIX C

This function also returns the length of the longest contact.

In this function we will be using the fact that every contact is "closed", such that in the list of lines in the contact, an endpoint of a line is equal to the startpoint of the next line.

This way the only tricky part is to compute the length of the overlapping part in the first and last line. Then we just add the length of the lines in the middle of the list.

See previous function. */ s t a t i c d o u b l e

BM_keep_largest_contact_remove_the_rest ( BM_LINKED_LI ST_NODE ♦ ♦ list of contacts )

{ ^{" "} long i , j ; if (NULL == list_of_contacts) return 0.0 ; BM_LINKED_LIST_NODE ♦contacts = ^♦list_of_contacts ; if (NULL == contacts) return 0.0 ;

// some of the most important parameters we compute in this function. double length_of_largest_contact = 0.0 ; long largest_contact_i = -1 ;

// we need to keep track of the lengths of contacts. // we need to store the length of every contact. //

// NOTICE : here we try to save some heap allocation/release time by using a pre-allocated

// array if the list of contacts is not very long.

// static double static_contact_len_array[32] ; double ^♦dynamic_contact_len_array = NULL ; double ^♦contact_len_array = NULL ;

// get the number of contacts long num_of_contacts = (long) contacts->get_obj () ; contacts = contacts->next () ;

// if no contacts, delete the entire list if (num_of_contacts < 1) {

// there is no list, delete node 0 as well. delete ^♦list_of_contacts ;

22 APPENDIX C

♦list_of_contacts = NULL ; return 0.0 ;

}

// allocate memory for the contact length array if (num_of_contacts <= 32) { contact_len_array = static_contact_len_array ; else { if (NULL == (dynamic_contact_len_array = new double [num_of_contacts] ) ) goto computation_done clean_up ; contact len array = dynamic_contact_len_array ;

} ^{" "}

// Scan all known contacts and compute the length of every contact.

// later we use these values to find the longest contact (which could be a simultaneous bendline) .

BM_LINKED_LIST_NODE ^♦last_facel_list, *last_face2_list ;

BM_LINKED_LIST_NODE ^♦line_in_list ; double k ; double first_kl, first_k2 ; double last_kl, last_k2 ; last_face2_list = NULL ; for (i = 0 ; i < num_of_contacts ; i++) { if (NULL == last_face2_list) last_facel_list = contacts

; else last_facel_list = last_face2_list->next 0 ; last_face2_list = last_facel_list->next () ;

BM_LINKED_LIST_NODE ♦left = (BM_LINKED_LIST_NODE ^♦) last_facel_list->get_obj () ;

BM_LINKED_LIST_NODE ^♦right = (BM_LINKED_LIST_NODE ^♦) last_face2_list->get_obj () ;

BM_LINE ^♦first_line_in_left = (BM_LINE ^♦) left->get_obj ()

BM_LINE ^♦first_line_in_right = (BM_LINE ^♦) right->get_obj () ;

// here is what we do. we know that the first lines in left and right have to have

// an overlapping part, we compute overlapping parameters kl and k2 with respect to the

// first line on the left, between the first lines.

// the idea is that one of the points that corresponds to these parameters (kl and k2)

// is the extreme point of the contact, we just don't know yet which one.

// then we find the last points in both lists, we know

23 APPENDIX C

that they must have non-empty

// overlapping part as well, we compute kl and k2 for these two lines as well.

// then we compate kl,k2 for the first and last pairs to find the two

// extreme points of the contact, distance between them is the length of the contact.

// compute kl and k2 for the first lines. first_line_in_left-_>is_point_on_line (first_line_in_right->get_sta rtpt () , &first_kl) ; first_line_in_left-_>is_point_on_line (first_line_in_right->get_end pt () ,&first_k2) ;

// we only want kl and k2 within the first line if (first_kl < 0.0) first_kl = 0.0 ; else if (first_kl > 1.0) first_kl = 1.0 ; if (first_k2 < 0.0) first_k2 = 0.0 ; else if (first_k2 > 1.0) first_k2 = 1.0 ;

// sort kl and k2 if (first_kl > first_k2) { k = first_kl ; first_kl = first_k2 ; first_k2 = k ; }

// find the last line in the left. for (line_in_list = left->next () ; line_in_list ; line_in_list = line_in_list->next () ) { left = line in list ;

} ^{" "}

// "left" is the last node in the left now.

BMJL.INE ^♦last_line_in_left = (BM_LINE ♦) left->get_obj ()

// find the last line in the right. for (line_in_list = right->next 0 ; line_in_list line_in_list = line_in_list->next () ) { right = line in list ;

} ^{" "}

// "right" is the last node in the left now.

BM_LINE ^♦last_line_in_right = (BM_LINE ♦) right->get_obj () ;

// compute kl and k2 for the first lines. last_line_in_left->is_point_on_line (last_line_in right->get start pt () , &last_kl) ; ^~ last_line_in_left->is_point_on_line (last_line in right-_>get endpt () ,&last_k2) ; ^{~ ~}

24 APPENDIX C

// we only want kl and k2 within the last line if (last_kl < 0.0) last_kl = 0.0 ; else if (last_kl > 1.0) last_kl = 1.0 ; if (last_k2 < 0.0) last_k2 = 0.0 ; else if (last_k2 > 1.0) last_k2 = 1.0 ;

// note that we have compute last kl and k2 with respect to the last line in the left.

// we really need last kl and k2 with respect to the first line in the left.

BM_POINT lastkl (last_line_in_left->get_startpt () + last_kl^ (last_line_in_left->get_v() ) ) ;

BM_POINT lastk2 (last_line_in_left->get_startpt 0 + last_k2^ (last_line_in_left->get_v() ) ) ; first_line_in_left->is_point_on_line (lastkl, &last_kl) ; first_line_in_left->is_point_on_line (lastk2, &last_k2) ;

// sort kl and k2 if (last_kl > last_k2) { k = last_kl ; last_kl = last_k2 ; last k2 = k ; }

// now we need to take the extreme points of these two pairs if (first_kl > last_kl) first cl = last_kl ; if (first_k2 < last_k2) first_k2 = last_k2 ; contact_len_array [ i] = ( f irst_k2 first kl) ^♦ _} (first - line - in - left->get len()) ;

// find the longest contact.

// the main feature here is to check for simultaneous bendlines last_face2_list = NULL ; for (i = 0 ; i < num_of_contacts ; i++) { if (NULL == last_f ace2_list) last_f acel_list = contacts else last_facel_list = last_face2_list->next ( ) ; last face2 list = last facel list->next() ;

// make sure we don't accept contacts that have zero length. if (contact_len_array[i] < BM_PRECISION) continue ;

BM_LINKED_LIST_NODE ^♦left = (BM_LINKED_LIST_NODE ^♦) last_facel_list->get_obj () ;

BM_LINE ^♦first_line_in_left = (BM_LINE ^♦) left->get_obj () ;

// check if this contact is part of a simultaneous bendline.

// if yes, add its length to the first bendline in the

25 APPENDIX C

simultaneous bendline and

// set this length here to -1.0, so that we know it later BM_LINKED_LIST_NODE ♦temp_last_facel_list,

♦temp_last_f ace2_list = NULL ; for (j = 0 ; j < i ; j++) { if (contact_len_array [i] < BM_PRECISION) continue ;

// already part of something else if (NULL == t emp_l as t_f ace2_l i s t ) temp_last_facel_list = contacts ; e l s e t e m p _ l a s t _ f a c e l _ l i s t temp_last_f ace2_list - >next ( ) ; temp_last_face2_list = temp_last_facel_list->next ()

;

BM_L I NKED_L I ST_NODE ^♦temp_lef t

(BM_LINKED_LIST_NODE ♦) temp_last_f acel_list->get_obj () ;

BM_LINE ♦temp_first_line_in_left = (BM_LINE ^♦) temp_left->get_obj () ;

// check if the lines overlap i f ( ! first_line_in_left->is_point_on_line (temp_first_line_in_left->get _startpt () ,NULL) ) continue ; i f ( ! first_line_in_left->is_point_on_line(temp_first_line_in_left->get _endpt () ,NULL) ) continue ;

// ok, they overlap

// add the length to the first in simultaneous bend and set this to -1.0 contact_len_array[j] += contact_len_array[i] ; contact_len_array[i] = -1.0 ; break ;

» ^}

// in this loop we actually find the largest for (i = 0 ; i < num_of_contacts ; i++) {

// check to see if this new contact is the largest if ⁽length_of_largest_contact < contact_len_array[i] ) { length_of_largest_contact = contact_len_array [i] ; largest contact i = i ;

) > " ^' computation_done clean_up :

// remove all other contacts that are not largest BM_LINKED_LIST_NODE ^♦next_last_f acel_list ,

26 APPENDIX C

♦next_last_face2_list = NULL ;

BM_LINKED_LIST_NODE ^♦largest_contact_list_f1 = NULL, ♦largest_contact_list_f2 ; last_facel_list = contacts ; last_face2_list = last_facel_list->next () ; for (i = j = 0 ; i < num_of_contacts ; i++) { next_last_facel_list = last_face2_list->next 0 ; if (next_last_facel_list) next_last_face2_list next_last_facel_list->next () ; else next_last_face2_list = NULL ;

// make sure we don't delete the largest contact

// note : we skip two lists, facel list and face2 list if (contact_len_array fi] < 0.0 j | largest_contact_i == i) if (NULL == largest_contact_list_f1 ) largest_contact_list_f1 = last_facel__list ; e l s e largest_contact_list_f2->add_after (last_facel_list) ; largest_contact_list_f2 = last_face2_list ;

// count the number of contacts ++j ; last_facel_list = next_last_facel_list ; last_face2_list = next_last_face2_list ; continue ;

}

// ok, this is not the largest contact, delete both facel and face2 sides of the contact

BM_LINKED_LIST_NODE ^♦next_line_in_list ; for (line_in_list = (BM_LINKED_LIST_NODE ^♦) last_facel_list->get_obj () ; line_in_list ; line_in_list = next_line_in_list) { next_line_in_list = line_in_list->next () ; delete line in list ;

} ^{" "} for (line_in_list = (BM_LINKED_LIST_NODE ♦) last_face2_list->get_obj () ; line_in_list ; line_in_list = next_line_in_list) { next_line_in_list = line_in_list->next () ; delete line in list ;

} ^{~ "} delete last_facel_list ; delete last_face2_list ; last_facel_list = next_last_facel_list ; last_face2_list = next_last_face2_list ;

27 APPENDIX C

} // end the contact list with a NULL pointer i _f ( _{l a} r g e s t _ c o n t a c t _ l i s t _ f 2 ) _lar_gest_contact_list_f2->add_after (NULL) ;

// write the number of contacts left in node 0. also, set the correct beginning of the list.

// however, if there is no contact, erase the entire list, if (NULL == largest_contact_list_fl) {

// there is no list, delete node 0 as well. delete ♦list_of_contacts ;

♦list_of_contacts = NULL ; else {

(♦list_of_contacts) ->set_obj ( (BM_BASIC ^♦) j) ;

(*list_of_contacts) ->add_after(largest_contact_list_f1)

7

}

// delete dynamic_contact_len_array if it was allocated if ( dynami c_c on t ac t _1 en_ar r ay ) delete [] dynamic__contact_len_array ; return length_of_largest_contact ;

/*

This function is used only by static int BM_create_bendline (...) as a subroutine. a big problem now is that we have deleted some lines. the problem is that if we have a simultaneous bend, other contacts could still be refering to these lines we just deleted. we have to go and check that, and if true, replace pointers to the lines we have changed/deleted with valid references. namely, lines that are affected are the first and last lines in the left. potentially references to them need to be replaced with references to ll and 13. */ static void check_references_in_remaining_contacts (

BM_LINKED_LIST_NODE ^♦side_l, // current list we are working with

28 APPENDIX C

BM_LINKED_LIST_NODE ^♦side_2, // list on the other side of the contact

BM_LINE ♦linel,

BM_LINE ^♦lineS, // these are old references, to be checked if still valid

BM_LINE *ll,

BM_LINE ^A13 // 11 and 13 are the new references

)

{

BM_LINKED_LIST_NODE ^♦sidel = side_l, ^♦side2 = side_2 ; process_next_contact :

// find the beginnings of lists in the next contact

BM_LINKED_LIST_NODE ♦contact_list ; if (NULL == (contact_list = sidel->next ()) ) return ; // done, no more lists if (NULL == (sidel = contact_list->next () ) ) return ; // done, no more lists if (NULL == (contact_list = side2->next 0 ) ) return ; // done, no more lists if (NULL == (side2 = contact_list->next () ) ) return ; // done, no more lists

// now we need to check the beginning of this contact and the end of this contact exactly the same way.

// we describe checking the beginning, check the end is the same.

// whenever we find a match, we can go to the next list (ie. don't have to continue with the

// current list) .

//

// we take first lines is sidel and side2. if the first line in sidel is either linel or line3,

// we check which of 11 and 13 overlaps the first line in side2. we know that exactly one of

// 11 and 13 has to overlap, then we replace the reference to linel/line3 with a reference to the

// one of 11/13 that overlaps the first line in side2.

BM_LINKED_LIST_NODE ^♦start_sidel = (BM_LINKED_LIST_NODE ♦) sidel->get_obj () ;

BM_LINKED_LIST_NODE ^♦start_side2 = (BM_LINKED_LIST_NODE ♦) side2->get_obj () ;

// check the beginning of the contact

BM_LINE ^♦line_in_sidel = (BM_LINE ^♦) start_sidel->get_obj 0 ; BM_LINE ^♦line_in_side2 = (BM LINE ^♦) start_side2->get_obj 0 ; if (line_in_sidel == linel |^~| line_in_sidel == line3) {

29 APPENDIX C

if (11) { // check if 11 overlaps first line in side2 i f

(line_in_side2-_>is_point_on_line_segment (ll->get_startpt 0 , NULL)

I I I I line_in_side2-_>is_point_on_line_segment (ll->get_endpt () , NULL)) { start_sidel->set_obj (ll) ; goto process next contact ;

} if (13) { // check if 13 overlaps first line in side2 i f

(line in_side2->is_point_on_line_segment (13->get_startpt () , NULL)

I I I I line_in_side2-_>is_point_on_line_segment (13->get_endpt () , NULL) ) start sidel->set robj (13) ; goto process next contact ; }

BM_LINKED_LIST_NODE ^♦line_in_list, ^♦prev_line_in_list ;

// find the last line in side2. for (line_in_list = start_side2->next () ; line_in_list line_in_list = line_in_list->next () ) { p_}rev-line_in_list = line in list ;

// "prev_line_in_list" is the last node in the left now. Iine_in_side2 = (BM_LINE ^♦) prev_line_in_list->get_obj () ; // find the last line in sidel. for (line_in_list = start_sidel->next () ; line_in_list ; line_in_list = line_in_list->next () ) { prev_line_in_list = line_in_list ;

// "prev_line_in_list" is the last node in sidel now. line_in_sidel = (BM_LINE ^♦) prev_line_in_list->get_obj () ;

// check the end of the contact if (line_in_sidel == linel | | line_in_sidel == line3) { if (11) { // check if ll overlaps first line in side2 i f

(line_in_side2->is_point_on_line_segment (ll->get_startpt () , NULL) i I line_in_side2->is_point_on_line_segment (ll->get_endpt () , NULL) ) { prev_line_in_list->set_obj (ll) ; goto process_next_contact ;

30 APPENDIX C

if (13) { // check if 13 overlaps first line in side2 i f

(line_in_side2->is_point_on_line_segment (13->get_startpt () , NULL)

II line_in_side2->is_point_on_line_segmen (l3->get_endpt () , NULL) ) prev_line_in_list->set_obj (13) ;

} goto process_next_cont ct ;

}

/*

This function will create a simple (regular) bendline between the two faces, based on the contact list given as input. Once done, it will erase the contact list.

In this function we assume that edges of faces are sorted according to the left-hand side rule.

This function returns TRUE iff everything is fine.

NB! the key here is not to invalidate anything in the part as we create a new bendline.

In order to accomplish that, we will create the bend op is isolation (without the bendline) , and then later attach it to the bendline.

Also, we will create topology only after the bendline is built and added to the part.

Note that this function creates bendline information that is consistent with facel (mainly, facel normal) .

However, face2 normal can be inconsistent with this information, but that has to be fixed later. */ static int BM_create_bendline (BM_PART & part, BM_LINKED_LIST_NODE ^♦♦contact, BM BENDLINE ^♦♦bendline created /♦ OUT ^♦/)

{

// so far, nothing is created ^♦bendline_created = NULL ;

// it is possible, that we are building bendlines for the flat version of the part.

// if this variables is TRUE, we are building bendlines for the 3D-version, otherwise for the flat. int using_3D_version = 1 ;

31 APPENDIX C

// later we need to create a bendline centerline

// these are the start-end points of the bendline centerline

// note that they are points pi and p2 on the facel side of the bendline

// this is done because sometimes (when the bendline is not bent) we use the

// facel surface as the bendline surface, ie. the bendline orientation is defined

// with respect to facel

BM_POINT bendline_centerline_pl, bendline_centerline_p2 ;

// contact list parameters

BM_LINKED_LIST_NODE *line_in_list, ^♦next_line_in_list ; BM_LINKED_LIST_NODE ^♦left, ^♦right ; int number_of_lists = (long) (^♦contact) ->get_obj () ; BM_LINKED_LIST_NODE ^♦contact_list = (^♦contact) ->next () ; left = (BM_LINKED_LIST_NODE ^♦) contact_list->get_obj () ; r i g h t = ( B M _ L I N K E D _ L I S T _ N O D E ^♦ ) (contact_list->next () ) ->get_obj () ;

BM_LINE ^♦first_line_in_left = (BM_LINE ^♦) left->get_obj () ; BM_LINE ♦first_line_in_right = (BM_LINE ^♦) right->get_obj () ; BM_LINE ^♦last_line_in_left ; BM_LINE ^♦left_line, ^♦right^ine ;

// get loops in both bodies

BM_LOOP ^♦facel_loop = first_line_in_left->get_loop () ; BM_LOOP ^♦face2_loop = first_line_in_right->get_loop () ; if (NULL == facel_loop | | NULL == face2_loop) return 0 ;

// 2D-bodies of either face

BM_2D_BODY ^♦facel_body = facel_loop->get_2D_body() ;

BM_2D_BODY ^♦face2_body = face2_loop->get_2D_body() ; if (NULL == facel_body j j NULL == face2_body) return 0 ;

// get facel and face2

BM_FACE ^♦facel = (BM_FACE ^♦) facel__body->get_3D_body() ;

BM_FACE ^♦face2 = (BM_FACE ^♦) face2_body->get_3D_body() ; if (NULL == facel j | NULL == face2) return 0 ;

// check whether we are using the flat version of the part, or 3D-version of the part. if ⁽facel->get_current_3D_ ersion() ! = facel_body | ] face2->get_current_3D_version() != face2_body) using_3D_version = 0 ;

// get surfaces of either faces

BM_PLANE ^♦facel_plane, ^♦face2_plane ; facel_plane = ⁽BM_PLANE ^♦) facel_body->get_surface () _; face2_plane = ⁽BM_PLANE ^♦) face2_body->get_surface () '•

32 APPENDIX C

/*

First we will do everything that has to do with the bendline - create a bendline, create a bend op, set bend op parameters, create and set cylinder for the bendline, set bendline centerline, etc. */

// try to figure out whether we have a FRONT or BACK bend.

B M _ V E C T 0 R v_face2_body( (face2_plane->get_normal () ) ^♦ (first_line_in_right->ge t_v())) ; // this vector

// points toward? the inside of face2 on the plane of face2

B M _ V E C T 0 R v_face2_normal (v_face2_body^ (first_line_in_left->get_v() ) ) ; // this is the correct normal for face2 ?

// compute the angle between faces double angle = v_face2_normal ^Λ (facel_plane->get_normal () ) ;

// compute bend type (FRONT/BACK) if (fabs (angle) <= BM_ANGLE_TOLERANCE) angle = PI ; else {

BM_VECT0R v( (facel_plane->get_normal () ) ⁺v_face2_normal) if (v % (first_line_in__left->get_v() ) > 0.0) angle = PI + angle ; // BACK bend else angle = PI - angle ; // FRONT bend }

// create a new empty bendline and add it to the part BM_2D_BODY ♦new_auto_bendline = new BM_2D_B0DY ; if (NULL == new_auto_bendline) return 0 ; BM_BENDLINE ^♦new aendline = new BM_BENDLINE (&part, using_3D_version ? NULL : new_auto_bendline, using_3D_version ? new_auto_bendline : NULL) ; if (NULL == new_bendline) { delete new_auto_bendline ; return 0 ;

}

// create a regular bend op for the bendline

B M_B END_0 P_R E GUL AR ^♦new_bend_op = new

BM_BEND_OP_REGULAR (NULL, angle, 0.0 ,0.0) ; // don't attach to bendline yet if (NULL == new_bend_op) ( delete new_bendline ; return 0 ;

} // attach this bend op to the bendline. this will just assign

33 APPENDIX C

the bend_op variable in the bendline new_bendline->attach_bend_op (new_bend_op) ;

// create bendline surface, if angle is not 180 degrees, we create a cylinder, otherwise a (facel) plane. BM_SURFACE ^♦bend_surface ; if (fabs (angle - PI) <= BM_ANGLE_TOLERANCE) { // bendline not being bent

// note : facel plane is the bendline surface bend_surface = new BM_PLANE (^♦facel_plane) ; new_bendline->Do_This_Bend(0) ; // this will invalidate 3D version as well, later we validate it.

// set sense to the same as facel sense new_auto_bendline-_>set_sense (facel_body->get_sense () ) ; else {

// first, compute vx

B M _ V E C T 0 R v_facel_body ( (facel_plane->get_normal () ) ^♦ (f irst_line_in_lef t->get _v ( ) ) ) ;

// v_facel_body points towards the inside of facel on the plane of facel v_facel_body.set_length(-l .0) ; // invert and normalize v_face2_body.set_length (-1.0) ; // invert and normalize

// this is the vx vector of the bendline surface BM_VECTOR vx(v_facel_body + v_face2_body) ; b e n d _ s u r f a c e = n e w

BM_CYLINDER (f irst_line_in_lef t->get_startpt ( ) , f irst_line_in_lef t->get_endpt () , vx, 0.0) ; new_bendline->Do_This_Bend(l) ; // this will invalidate 3D version as well, later we validate it.

// set sense to 0, because we want the bendline (cylinder) to represent explicitly OUTSIDE

// ie. thickness vector is opposite to the normal (= radius vector) . new_auto_bendline->set_sense (0) ; if (NULL == bend_surface) { delete new_bendline ; return 0 ;

} new_auto_bendline->set_surface (bend_surface) ;

// lastly we need to create the bendline centerline, but for that we need to know its length, and

34 APPENDIX C

// start-end-points. However, for that we need to process the contact list which contains the

// lines that are adjacent. We will do that next, and when done, will create bendline centerline.

/*

Second, we will break some lines, if necessary, and combine all adjacent lines into a one adjacent line. We will also create adjacency information between facel-bendline and face2-bendline.

*/

// notice that processing one side of the contact is independent of the other side.

// moreover, the logic is exactly the same, therefore, we will use the following code

// to process both facel' side and face2' side. char face2_processed = 0 ; process_one_side_of_the_contact :

// in general, the contact line (on the side of either face (facel and face2) ) will consist

// of three lines - we have a list of lines, first line is the line before the contact, the

// second line is the contact line, the last line is the line after the contact.

// these variables will be used to compute these three lines. double k ; double kO, kl, k2, k3 ; // parameters k for the first line in the left double end_kl, end_k2 ; // parameters k for the last line in the left list

BM_POINT pO, pi, p2, p3 ; // pO - p3 are the endpoints of the edges that will be left (ie. endpoint of left_line, 12, 13) .

// pO and p3 and the two extreme endpoints and trivial to find, we really need to compute only pi and p2.

BM_LINE *11, +12, *13 ; // ll is the line before 12, 12 is the line that is adjacent, 13 is the one after it.

// process first line of the left side kO = 0.0 ; k3 = 1.0 ; first_line_in_left->is_point_on_line (first_line_in_right-_>get sta rtpt () , &kl) ; first__line_in_left->is_point_on_line (first_line_in_right-_>get end pt () , &k2) ; if (kl > 1.0) kl = 1.0 ;

35 APPENDIX C

else if (kl < 0.0) kl = 0.0 ; if (k2 > 1.0) k2 = 1.0 ; else if (k2 < 0.0) k2 = 0.0 ; if (k2 < kl) { k = kl ; kl = k2 ; k2 = k ; }

// note that we cannot yet compute pO, pi and p3 because we don't know the relative order

// of the first and last lines in the left list

// find and process the last line for (line_in_list = left ; line_in_list ; line_in_list = next_line_in_list) { if (NULL == (next_line_in_list = line_in_list->next () ) ) left_line = (BM_LINE ^♦) line_in_list->get_obj () ;

// first, we have to find the end of the face2 list for (line_in_list = right ; line_in_list ; line_in_list = next_line_in_list) { if (NULL == (next_line_in_list line in list->next () ) ) break ;

^{" ~} right_^}line = (BM_LINE ^♦) line_in_list->get_obj () ; left_line->is_point_on_line (right_line->get_startpt () , &end_kl) ; left_line->is_point_on_line (right_line->get_endpt () ,&end_k2) ; if (end_kl > 1.0) end_kl = 1.0 ; else if (end_kl < 0.0) end_kl = 0.0 ; if (end_k2 > 1.0) end_k2 = 1.0 ; else if (end_k2 < 0.0) end_k2 = 0.0 ; if (end_k2 < end_kl) { k = end_kl ; end_kl = end_k2 ; end_k2 = k ; } break ;

// now we have to compute points pO, pi, p2 and p3. // but first we need to convert end-k parameters to equivalent parameters with respect to the // first line in the left pO = left_line->get_startpt () ; pi = left_line->get_startpt () + end_kl ^♦ (left_line-_>get_v() )

7 p2 = left_line->get_startpt () + end_k2 ^♦ (left_line->get_v() ) p3 = left_line->get_endpt () ;

// check kO first_line_in_left->is_point_on line(p0,&k) ; if (k < kO) kO = k ;

36 APPENDIX C

// check kl first_line_in_left->is_point_on__line (pi, &k) ; if (k < kl) kl = k ;

// check k2 first_line_in_left->is_point_on_line (p2, &k) ; if (k > k2) k2 = k ,-

// check k3 first_line_in_left->is_point_on_line (p3 , &k) ; if (k > k3) k3 = k ;

// here we can compute pO, pi and p3. later we compute p2. pO = first_line_in_left->get_startpt () + kO ^♦ (first_line_in_left->get_v() ) ; pi = first_line_in_left->get_startpt () + kl ^♦ (first_line_in_left->get_v{) ) ; p2 = first_line_in__left->get_startpt () + k2 ^♦ (first_line_in_left->get_v() ) ; p3 = first_line_in_left->get_startpt () + k3 ^♦ (first__line_in_left->get_v() ) ;

// ok, we have computed p0-p3. now we will erase left list, except for the first line.

// we need the first line later.

// actually, the first line will, in the future, point to new line 12.

BM_LOOP ♦left_loop = first_line_in_left->get_loop () ; last_line_in_left = NULL ; // the last line in the left, we need it later to replace out-of-date pointers with valid ones. line_in_list = left->next() ; // this is the second line in the left list. left->add_after(NULL) ;

// note : we will erase all lines in the left list, starting from the second line. for (; line_in__list ; line_in_list = next_line_in_list) { next_line_in_list = line_in_list->next () ; // store a pointer to the last line last_line_in_left = (BM_LINE ^♦) line_in_list->get_obj ()

; left_loop->remove_edge (last__line_in_left) ; delete last_line_in_left ; delete line in list ;

} ^{" ~}

// finally, create the adjacent line 12, and two more lines (11 and 13) if necessary, if (pO != pi) {

11 = first_line_in_left ; ll->set (p0,pl) ;

12 = new BM_LINE(pl,p2) ;

37 APPENDIX C

left_loop->add_edge_after (ll, 12) ; else (

11 = NULL ;

12 = first_line_in_left ; 12->set (p0,p2) ;

^} if (p2 != p3) (

13 = new BMJLINE (p2,p3) ; left loop->add_edge_after(12, 13) ;

} else 13 = NULL ;

// last thing, create topology by making this line adjacent to the bendline

12-_>make_body_adjacent (new_bendline) ;

// a big problem now is that we have deleted some lines .

// the problem is that if we have a simultaneous bend, other contacts could still be refering

// to these lines we just deleted.

// we have to go and check that, and if true, replace pointers to the lines we have changed/deleted

// with valid references.

//

// namely, lines that are affected are the first and last lines in the left.

// potentially references to them need to be replaced with references to 11 and 13. if (11 j j 13) check_references_in_remaining_contacts ( face2_processed ? contact_list->next () : contact_list, // current list we are working with face2_processed ? contact_list : contact_list->next () , // list on the other side of the contact first_line_in_left, last_line_in_left, // these are old references, to be checked if still valid ll, 13 // ll and 13 are the new references ) ;

// we will use the same piece of code to process both sides of the contact .

// if face2 side has not been processed, do it now.

// however, first we need to set some values of variables that are commonly used. if ( ! face2_processed) { face2_processed = 1 ;

// first, 12 should be the left line.

// 12 is the real adjacent line (on the left) . now it

38 APPENDIX C

becomes the right line. left->set_obj (12)

// now we need to switch left and right sides // switch lines first_line_in_left = first_line_in_right ; first_line_in_right = 12 ;

// switch lists line_in__list = left ; left = right ; right = line_in_list ;

// store bendline centerline start-end points bendline_centerline_pl = pi ; bendline_centerline_p2 = p2 ; goto process one side_of the_contact ; }

We can delete contact lists now.

BM_LINKED_LIST_NODE ^♦list, ^♦next_list ; int lists_deleted = 0 ; // note, we can only delete first two lists for (list = contact_list ; lists_deleted < 2 && list ; list = next_list) { next_list = list->next() ;

// delete the list for (line_in_list = (BM_LINKED_LIST_NODE ^♦) list->get_obj () ; line_in_list ; line_in_list = next_line_in_list) { next_line_in_list = line_in_list->next 0 ; delete line in list ;

} ^{" "} delete list ;

++lists deleted ;

} // set the correct number of contacts left and update the pointers long num_contacts = (long) (^♦contact) ->get_obj () ; if (--num_contacts < 1) {

// if no contacts left, erase the whole contacts list delete ^♦contact ;

^♦contact = NULL ;

}

39 APPENDIX C

else {

(♦contact) -_>set_obj ( (BM_BASIC ^♦) num_contacts) ;

(♦contact) ->add after(list) ;

}

*

Finally, create centerline. Make sure direction and length are correct.

Note, pi and p2 are defined with respect to facel.

Ie. pi -_> p2 is the direction on the contact on the side of facel .

*/

B M _ L I N E centerline (bendline_centerline_pl,bendline_centerline_p2) ; if (fabs (angle - PI) <= BM_ANGLE_TOLERANCE) { // bendline not being bent

// facel plane is the bendline surface i f ( f a ce l - >ge t_idx ( ) > f ac e 2 - >ge t_id ( ) ) centerline . reverse_line ( ) ; else if (angle < PI ) { i f ( f ac e l - >ge t_i dx ( ) < f ace 2 - >get_i dx ( ) ) centerline . reverse line ( ) ; else angle is > PI

(facel - >get_idx ( ) > face2 - >get_idx ( ) ) centerline.reverse lineO ;

} new_auto_bendline->set_center_line(centerline) ;

// if we are building bendlines for the 3D-version of the part,

// make sure 3D-version of the bendline is validated because most likely we

// want to UNFOLD right away. if (using_3Diversion) new_bendline->validate_3D_version() ;

^♦bendline_created = new_bendline ; return 1 ;

}

/*

The purpose of this function is to create a minimum number of bendlines so that a part that has disconnected components consisting of faces, the part will be connected once we add these new bendlines.

40 APPENDIX C

This function assumes that the part has thickness 0.

The basic idea we use in this function is maximum spanning tree.

We model the part by an undirected graph, where faces are the nodes of the graph and there is an edge between two nodes iff these two faces are in contact (ie. are touching each other) , the cost (or length) of the edge is the length of the longest contact between these two faces

(ie. the length of the longest possible bendline between the two faces) .

In order to create bendlines we compute a maximum spanning tree for this graph and create a bendline for every edge of the graph that is in the maximum spanning tree.

There are two problems however :

It could be that the part already contains some bendlines (ie. some faces are already connected) .

- When we compute the maximum spanning tree, we can have two fringe-edges that have the same length. In that case, we don't want to break ties randomly, but instead choose an edge that would minimize the diameter of the spanning tree (the maximum distance between any two vertices of the tree) .

Algorithm :

We will process faces one-by-one, starting from the first faces of the part and proceeding until the last faces, and building the maximum spanning tree along the way as we go.

For every pair of faces, such that at least one of the is not yet in the spanning tree, we keep the longest contact between them. When we expand the spanning tree, we choose the contact between a face in the tree and a face not in the tree that is the longest contact between any face in the tree and a face outside. If there is a tie, we choose an edge that would minimize the */

/*

41 APPENDIX C

Given a face (facel) this structure contains information about the contact between this face and another face.

For any two faces, we need to know if there is a contact between these two faces and if yes, what is its length. The simplest way to handle this data is to store it as a two-dimensional array. However, this would take a lot of space, most of which would be wasted, usually, since if there are many faces, most of them do not have any contacts between them.

Instead we will store this info for every face as a list of structures. We will create this record (ie. structure) only if the length of the contact between two faces is greater than 0. */ typedef struct CONTACT_INFO {

// here we will keep track what is the length of the longest contact between any two faces

// this is really a two-dimensional array double length ;

// here we keep a pointer to the other face BM_FACE ^♦face2 ;

// here we will keep the longest contact for any two faces // this is really a two-dimensional array stored as a linked list of lists .

BM_LINKED_LIST_NODE ^♦contact ;

// a pointer to the next structure in a list CONTACT_INFO ^♦next ;

} contact_info ;

/*

This is a very basic function used by many functions in this file, It deletes a contact list.

*/ static void delete_contact_info_list (contact_info ♦list) contact_info ^♦c, ^♦c αext ;

42 APPENDIX C

BM_LINKED_LIST_NODE ^♦clist, ^♦next_clist ; BM_LINKED_LIST_NODE ^♦line_in_list, ^♦next_line_in_list ; for (c = list ; c ; c = c_next) { c_next = c->next ; if (c->contact) {

// skip the header = num of contacts for (clist = (c->contact) ->next () ; clist ; clist = next_clist) { next_clist = clist->next () ;

// delete the list for (line_in_list = (BM_LINKED_LIST_NODE ^♦) clist->get_obj () ; line_in_list ; line_in_list = next_line_in_list)

{ next_line_in_list = line_in_list->next () delete line in list ;

} ^{" "} delete clist ;

} // delete the header delete c->contact ;

} delete c ; }

static void delete_contact_info_structure (contact_info *c)

{

BM_LINKED_LIST_NODE ^♦clist, ^♦next_clist ;

BM_LINKED_LIST_NODE ^♦line_in_list, ^♦next_line_in_list ; if (NULL == c) return ; if (c->contact) {

// skip the header for (clist = (c->contact) ->next () ; clist ; clist = next_clist) { next_clist = clist->next () ;

// delete the list for (line_in_list = (BM_LINKED_LIST_NODE ^♦) clist->get obj () ; line_in list ; line in list = next line in list)

{ next_line_in_list = line_in_list->next () ;

43 APPENDIX C

delete line_in_list ;

} delete clist ;

} // delete the header delete c->contact ;

} delete c

/*

This function will compute and store the best (ie. longest) contact between two faces . */ static int compute_the_longest_contact_between_faces ( BM_FACE & facel /^♦ IN ^♦/, BM_FACE & face2 /^♦ IN ^♦/, contact__info ♦♦computed_contact_list /^♦ OUT ^♦/)

// so far we have computed nothing ♦computed_contact_list = NULL ;

BM_LINKED_LIST_NODE ♦contact_list ; // a list of all contacts between facel and face2 we will compute if (! BM_compute_all_face_to_face_contacts (facel, face2, &contact_list) ) return 0 ; d o u b 1 e x = BM_keep_largest_contact_remove_the_rest (&contact_list) ; if (x < BM_PRECISION) return 1 ; // no contact between these two faces contact_info ⁺c ; if (NULL == (c = new contact_info) ) { // delete contact list

// delete the list for ⁽line_in_list = (BM_LINKED_LIST_NODE ♦) clist->get_obj () ; line_in_list ; line_in_list = next_line_in_list) next_line_in_list = line_in_list->next () ; 44 APPENDIX C

delete line in list

} ^{" "} delete clist ;

} delete contact_list ; return 0 ; }

// initialize the structure c-_>contact = contact_^list ; c->next = NULL ; c-_>face2 = &face2 ; c->length = x ; ♦computed_contact_list = c ; return 1 ;

/*

This function adds a new face to the current connected component .

It also computes the longest contacts between this new face, and all other faces that have not been processed. */ static int add_face_to_connected_component (BM_PART ^♦part,

BM_FACE ♦face2 /^♦ this is the face we are adding to the connected component ^♦/, long ♦end_id /^♦ an index of the last face in the connected component, this is where face2 is. ^♦/, contact_info ^♦♦all_contacts /^♦ all contact info between face2 and other outside faces will be here ^♦/

/^♦ we will compute it and construct this list in this function ^♦/,

BM_FACE ^♦♦idx_to_face_mapping)

BM_FACE ^♦current_face ; contact_info ^♦last slist = NULL, ⁺cl ;

// just in case face2 has no adjacent faces outside the connected component all_contacts [^♦end_id] = NULL ;

// in this loop we want to compute all contacts between face2 and all other

// faces not yet in any connected component.

45 APPENDIX C

for (current_face = part->get_face_list () ; current_face ; current_face = (BM_FACE ^♦) current_face->next () ) { if (current_face->get_user_id () >= 0) continue ; // current_face is in some other connected component

// compute the best contact between face2 and current_face if (! compute_the_longest_contact_between_faces (⁺face2, ♦current_face, &cl) ) return 0 ; if (NULL == cl) continue ; // there is no contact between these two faces if (NULL == last_clist) all_contacts [^♦end_id] = cl ; else last_clist->next = cl ; last clist = cl ;

} ^~ face2->set_user_id(^end_id) ; idx_to_face_mapping [^♦end_id] = face2 ; ++(^end_id) ; return 1 ;

/*

These two functions allow the user to set and get the distance between two faces in the topology graph. dist_btw_faces is a upper right half of a two dimensional array. */ static long get_face2face_dist (unsigned char ^♦dist_btw_faces, long number_of_faces, long dist_array_size, long j, long i) long k ;

// distance between face_i and face_i is zero if (i == j) return 0 ;

// sort indeces if (j < i) { k = i _; i = j _; j = k ; } k = number_of_faces - i ; return dist_btw_faces [dist_array size - ((k k-1)) _>> 1) ₊ (i - i) - 1] ; } static void set_face2face_dist (unsigned char ♦dist_btw_faces,

46 APPENDIX C

long number_of_faces, long dist_array_size, long j, long i, unsigned char distance)

{ long k ;

// distance between face_i and face_i cannot be set if (i == j) return ;

// sort indeces if (j < i) { k = i ; i = j ; j = k ; } k = number_of_faces - i ; dist_btw_faces [dist_array_size - ((k*(k-D) >> 1) + (j - i) - 1] = distance ; }

/*

This function returns the largest distance between the given face and other faces in the connected component.

*/ static long get_largest_distance_for_face (unsigned char ♦dist_btw_faces, long number_of_faces, long dist_array_size, long start_id, long end id, long face id)

{ long j , k ; if (NULL == dist_btw_f aces) return 0 ; long current_f ace_to_f ace_distance = 0 ; for (j = start_id ; j < end__id ; j + + ) { k get_f ace2f ace_dist (dist_btw_faces , number_of_faces , dist_array_size , j , f ace_id) ; i f ( k > c u r r e n t _ f a c e__t o_ f a c e_d i s t an c e ) current face to face distance = k ;

^" } ^{" ~ "} return current_f ace_to_f ace_distance ;

/*

This is the main AutoBend function. Its purpose is to automatically create bendlines for the part. */

47 APPENDIX C

int BM_PART: :auto_bend(void)

{ long i, j , k ;

BM_FACE ♦current_face ;

BM_FORMING ^♦current_forming ;

BM_BENDLINE ♦current_bendline ;

// at end of the function we release memory, this code for releasing memory is shared

// between the cases when this function fails, or succeeds.

// therefore, before we get to that code, we have to know the return code. int ret_value ;

// these are the two faces we will be creating a bendline for (ie. between these two faces) . BM_FACE ^♦facel, ^♦face2 ;

// here we keep a list of information about contacts between every face and other faces.

// this is an array lists. contact_info ^♦♦all_contacts = NULL ;

// here we will keep track what face corresponds to each row in face_to_face_length[] and all_contacts [] arrays BM_FACE ^♦♦idx_to_face_mapping = NULL ;

// this (two-dimensional) array is used to store distances between faces in the TOPOLOGY GRAPH.

// these distance values are used to break ties in the maximum spanning tree algorithm when

// two edges have the same length. Note that the distance between (i,j) is the same as

// the distance between (j,i) . This means that the distance matrix is a symmetric matrix, and we

// only need to store one half (upper right triangle) to save space.

// we will use this method of breaking ties only when the number of faces is less than 256.

// this has two advantages : 1) the distance between any two faces is always less than 256,

// and therefore we only need a char (8 bits) to store the distance values, 2) since we only

// need a char for a pair, the total this array can take is 32K bytes, which is not too much memory. unsigned char ^♦dist_btw_faces = NULL ,•

// this will be the size of the distance array, given some number of faces long dist_array_size ;

48 APPENDIX C

// we will have to create a heap because some portion of the part can already be connected

// and we need to make sure we don't create any bendlines for that portion of the part.

// we will use a heap to traverse that portion of the part. long body_priority = number_of_formings + number_of_faces + number_of_bendlines ;

// Check if there are any bodies. If no, we are done. if (body_priority < 1) return 1 ;

// this is the heap we will allocate, this is used when we add a face to the connected component

// to add all other faces connected to it as well, the problem is that the part could already

// contain some bendlines.

DBM_HEAP ^♦heap = NULL ;

// this is the start-index (in face_to_face_length[] and all_contacts [] ) arrays

// of the current connected component long start_id = 0 ;

// this is the end-index (in face_to_face_length[] and all_contacts [] ) arrays

// of the current connected component long end_id = 0 ;

// if TRUE, we are starting a new connected component int starting_new_connected_component = 1 ;

// allocate memory for local variables i f ( NULL = = ( a l l _ c o n t a c t s = n e w contact_inf o* [number_of_f aces] ) ) goto failure ; if (NULL == ( idx_t o_f ace_mapp ing = new

BM_FACE* [number_of_faces] ) ) goto failure ; if (NULL == (heap = new DBM_HEAP (0 , NULL, NULL, body_priority) ) ) goto failure ; if (number_of_f aces < 256) { // allocate memory for the face -face distance array only when

// the number of faces is not too large dist_array_size = (number_of_f aces* (number_of_f aces - 1) ) >> 1 ; if (NULL == (dist_btw_faces = new unsigned charfdist array size])) goto failure ;

^"} ++body_priority ; // just in case, we don't want the priority to go to 0 later.

// initialize arrays for (i = 0 ; i < number__of_f aces ; i+ + ) { all_contacts [i] = NULL ;

49 APPENDIX C

idx to face mapping[i] = NULL ;

} ^{" " "}

// initially mark all faces as not processed for (i = 0, current_face = first_face ; current_face ; current_face = (BM_FACE ♦) current_face->next () ) { current_face-_>this_body_already_processed = 0 ;

// this will tell us whether this face has been already processed.

// if yes, this id is the index in face_to_face_length[] and all_contacts f] arrays. current face->set_user_id(-1) ;

}

/*

First, try to expand the maximum spanning tree of the current connected component of faces.

*/

// these two variables are used in tie-breaking long best_face_to_face_distance, current_face_to_face_distance

// this is the length of the best contact so far (ie. the longest contact) double best_length ;

// this is a pointer to the best contact_info structure found so far contact_info ^♦best_cntct ; expand_spanning_tree :

// first, check if all faces have been processed // end_id is where we will put the next face we add to the connected component if (end_id >= number_of_faces) goto done ; contact_info ^♦cntct, ^♦cntct_next, ♦cntct_prev ; best_length = -1.0 ; best_face_to_face_distance = 999999 ; // a really large number

// in this loop we check all faces already in the connected component and try

// to find a face that has the largest contact between itself and an adjacent face outside the

// current connected component . for (i = start_id ; i < end_id ; i+₊) (

// in this loop we are checking for a largest contact between the current face

// in the connected component and all of its adjacent faces outside the connected component.

50 APPENDIX C

cntct_prev = NULL ; for (cntct = all_contacts [i] ; cntct ; cntct = cntct_next) { cntct_next = cntct->next ;

// check if this adjacent face is already processed if ( (cntct->face2) ->get_user_id() >= 0) (

// this means that face2 has already been added to the connected component

// and we don't really need the contact info stored in this structure any more.

// we will delete it.

// this will some time in the future and free up some space. delete_contact_info_structure (cntct) ; if (NULL == cntct_prev) all_contacts [i] = cntct next else cntct_prev->next = cntct_next ; continue ;

}

// ok, this adjacent face not processed.

// check if it is better than the previously known best.

// first check if there is a tie, and if we have allocated a face-face distance array

// then use the face-to-face distances to break the tie if (fabs (cntct->length - best_length) <= distance_tolerance && dist_btw_faces) {

// if tie-breaking is used, compute the largest distance for this face (ie. facel) c u r re n t _ f a c e__t o_ f a c e_d i s t an c e get_largest_distance_for_face (dist_btw_faces, number_of_faces , dist_array_size , start_id, end_id, i) ,-

// prefer the one with the smaller largest distance if (current_face_to_face_distance < best_face_to_face_distance) { facel = idx_to_face_mapping[i] ; face2 = cntct->face2 ; best_length = cntct->length ; best_cntct = cntct ; be s t_f a c e_t o_ f a c e_d i s t anc e current face to face distance ;

51 APPENDIX C

// take simply the one that is better else if (cntct->length > best_length) { facel = idx_to_face_mappingti] ; face2 = cntct->face2 ; best_length = cntct->length ; best_cntct = cntct ;

// if tie-breaking is used, compute the largest distance for this face (ie. facel) best_face_to_face_distance = current_f ac e__t o_f ac e_d i s t anc e get_largest_distance_f or_f ace (dist_btw_f aces , number_of_f aces , di st_array_s i z e , start id, end id, i ) ;

} cntct prev = cntct ;

) ^{} "}

// at the end of this loop,

// facel, face2, best_cntct should have correct values, if best_length > 0.0,

// that means we can expand the tree.

/*

Check if we can expand the spanning tree.

Note, for this point on until we get back to expanding the connected component, we will be using only face2, ie. we won't be needing facel.

The only case is when we create a new bendline, but we can extract face id's from the contact list.

*/ if (best_length < distance_tolerance) (

// looks like we cannot expand the spanning tree

// none of the faces in the current connected component has any adjacent faces

// outside the connected component.

// most likely this connected component is finished and, since there must be still faces left,

// there is more than one connected component .

//

// next we try to find the first face in the next connected component so that we

// can continue and create all bendlines for this new conencted component. for ⁽current_face = first_face ; current_face current_face = (BM_FACE ^♦) current_face->next () ) {

52 APPENDIX C

if (current face->get user id() < 0) break ;

} if (NULL == current_face) goto done ; // we should never get here

// first, we can delete all contact information about the current (now old) connected component for (i = start_id ; i < end_id ; i++) { delete_contact_info_list (all_contacts [i] ) ; all contacts fi] = NULL ;

} ^"

// here we will be starting a new connected component starting_new_connected_component = 1 ; start__id = end_id ;

// we will skip the bendline creation normally needed when adding a face to a connected component face2 = current_face ; else {

// ok, we are in the middle of a conencted component starting_new_connected_component = 0 ;

// create a bendline between facel and face2 // notice : we might have to create a simultaneous bend l o n g n u m _ c o n t a c t s = ( l o n g )

(best_cntct->contact) ->get_obj () ;

BM_BEND_PROPERTY_SIMULTANEOUS ^♦sim_bend = NULL ; if (num_contacts > 1) { sim_bend = new BM_BEND_PROPERTY_SIMULTANEOUS (this, num_contacts, NULL) ; if (NULL == si bend) goto failure ;

} BM_BENDLINE ^♦new_bendline ; for (long contact s_done = 0 ; contacts_done < num_contacts ; contacts_done++) { i f ( ! BM_create_bendline ( * t i s ,

& (best_cntct->contact) , &new_bendline) ) goto failure ;

// note that this function will delete one contact for which a bendline was created as well if (num_contacts > 1) {

// add this bendline to the simultaneous property new_bendl ine - >add_property (sim_bend) ;

53 APPENDIX C

// store "1 + the current index of the last connected-component-face" j = end_id ;

// we could add face2 to the connected component right away, but we could save some time by

// first checking if face2 has any other stuff (outside this connected component) connected to it.

// if yes, we will be adding that stuff to this connected component right away,

// but in that case there is no need for us to compute contacts between face2 and that other

// stuff in the first place.

// Therefore, we will just remember that face2 should be added, will process adjacent stuff first,

// and then will actually add face2 (and some more stuff may-be) to the connected component.

//

// Note : we will actully put face2 in the heap, and when we remove it from the heap,

// we will remember to add it to the connected component

/*

Now we need to check if this face (ie. face2) has any other 3D-bodies connected to it.

It yes, then we need to pursue these connections in order to add all faces connected to face2 to the current connected component.

*/ heap->empty() ; if (! heap->insert (body_priority--, (long) face2) ) goto failure ; face2->this_body_already_processed = 1 ;

// check if face2 has a parent face or not. if not, set the id in local array equal to

// face2's own id. this way we know it has no parent face.

//

// IMPORTANT : normally user_id of every face is a pointer to its index in the contact-info array.

// temporarily we will use this information to keep track which of the faces already processed

// is the parent-face (in the topology graph) of every new face added to the connected component .

// this information will be used later to compute distances in the face topology graph.

// the following loop (controlled by the heap) stores all faces not yet in the connected

// component, but connected to this face2.

// later when the actualy adding of new faces to the conencted

54 APPENDIX C

component is done,

// we overwrite user_id with the correct value. if ( ! starting_new_connected__component) f ace2->set_user_id (facel->get_user_id {) ) ; else face2->set_user_id(end_id) ;

// in the following while loop we will use this_body_already_processed member variable to keep

// track whether a varible is processed; not processed, but sitting in the heap; or unseen yet.

// 0 - unseen body

// I - body in the heap

// 2 - body already finished

// also, in this while loop we will be using the user_defined_id of every 3D-body to keep

// track of the parent face of every 3D-body we traverse.

// this has two benefits : 1) we will remember the parent face of every face and can compute

// distances between faces later, and

// 2) this way when faces are later added to the spanning tree, distance between some other

// face and this face will not be computed, since it is not necessary.

BM_3D_B0DY ^♦body ; BM_TOPOLOGY_RECORD ^♦toplgy ; while (heap->get_size () > 0) { if (! heap->remove (&i, (long ^♦) &body) ) goto failure ;

// a safety check if (NULL == body) continue ;

// check if this body is already finished if (body->this_body_already_j?rocessed > 1) continue ;

// get body topology toplgy = body->get_adj_list () ; if (NULL == toplgy) continue ;

// check if the body is a face, if yes, memorize. if (body->is(BM_ENTITY_TYPE_FACE) ) { face2 = (BM_FACE ^♦) body ;

// memorize to add this face to the connected component later idx_to_face_mapping[j++] = face2 ;

body->this_body_already_processed = 2 ;

55 APPENDIX C

// process all adjacent bodies for (current_face = (BM_FACE ♦) toplgy->get_first_face () ; current_face ; current_face = (BM_FACE ^♦) toplgy->get_next faceO) {

/7 make sure that we don't process faces twice if (current_face->this_body_already_processed) continue ; if (! heap->insert (body_priority- - , (long) current_face) ) goto failure ; current_face->this_body_already_processed = 1 ;

// make sure we memorize the parent face index in the local spanning tree array c_}urrent face->set_user_id(body->get_user_id()) ; for ( current_f orming = (BM_FORMING ^♦) toplgy- >get_first_forming () ; current_f orming ; current_f orming = (BM_FORMING ^♦) toplgy- >get_next_f orming () ) { if (current_f orming- >this_body_already_processed) continue ; if (! heap- >insert (body_priority- - , (long) current_f orming) ) goto failure ; current_f orming- >this_body_already_processed = 1 ;

// make sure we memorize the parent face index in the local spanning tree array if (body- >is (BM_ENTITY_TYPE_FACE) ) current_f orming- >set_user_id ( j - 1 ) ; e l s e current_f orming- >set_user_id (body- >get_user_id ( ) ) ; for (current_bendline = (BM_BENDLINE ^♦) toplgy- >get_f irst_bendline () ; current_bendline ; current_bendline = (BM_BENDLINE ^♦) toplgy- >get_next_bendline () ) { if (current_bendline->this_body_already_processed) continue ; if (! heap->insert (body_priority- - , (long) current_bendline) ) goto failure ; current_bendline->this_body_already_processed = 1 ;

// make sure we memorize the parent face index in the local spanning tree array if ⁽ bo dy - > i s ⁽ BM_ENT I T Y_T Y P E_F ACE ) ) current_bendline->set_user_id( j -1) ; e l s e current_bendline->set_user_id⁽body->get_user_id() ) ;

} ^}

56 APPENDIX C

// finally we need to compute contacts between this face and all other faces, that we postponed

// and add them to the connected component. long facel_id ; for (i = end_id ; i < j ; i++) { face2 = idx_to_face_mapping[i] ;

// first, compute distances in the TOPOLOGY GRAPH between this face (ie. face2)

// and all faces previously in the connected component. // we will use the parent face id in the local spanning tree array for that. facel_id = face2->get_user_id() ; // note : facel_id is the parent id of face2 (ie. there is a bendline between facel_id and face2 that we count) . if (dist_btw_faces && facel_id != i) { // if equal -> face2 has no parents,

// ie. a new connected component (must be start_id = end_id) .

// actually, it must be here true that facel_id < i for (k = start_id ; k < end_id ; k++) { set_face2face_dist (dist_btw_faces , number_of_f aces, dist_array_size, k, i, get_face2face_dist (dist_btw_faces,number_of_faces,dist_array_size , k, facel id) + 1) ;

^' > ^]

// else if equal, it must be that i == 'the end_id' in the beginning of this for-loop.

/ / not e that the next funct i on (add_face_to_connected_component ( ... ) ) will increment end_id. if (! add_face_to__connected_component (this, face2, &end id, all contacts, idx to face mapping)) goto failure ; } " - - -

// ok, the current connected component has no connections to the outside.

// try to expand the spanning tree, goto expand_spanning_tree ; done : ret_value = 1 ; goto release_memory_and_return ; failure :

57 APPENDIX C

ret_value = 0 ; release_memory_and_return :

// free memory if necessary for (i = 0 ; i < number_of_faces ; i++) { delete_contact_info_list (all_contacts [i] ) ■ if (all_contacts) delete [] all_contacts ; if (idx_to_face_mapping) delete [] idx_to_face_mapping ; if (heap) delete heap ; if (dist_btw_faces) delete [] dist_btw_f aces ; return ret value ;

58 APPENDIX D /////////////////////////////////////////////////////////////

' IJ /, Example of 2d Cleanup function. u//

// //

/////////////////////////////////////////////////////////////

#include <stdio.h>

#include <stdlib.h>

#include <math.h>

#include <memory.h>

#include <string.h>

/

Function Name :

1 control level

TestFindBoundary

TestFindDrawingSheetControl TestFindViewsControl

TestFindBoundaryControl TestFindArrowContro1 TestFindOneSideOpenControi TestFindOneSideOpen2Control TestFindCenterLineControl TestFindOneSideOpenConnectControl TestFindSameColorControl TestFindSelectSingleControl TestUnselectSingleControl TestClearControi TestClearKeepFlagsControl TestDeleteControl TestUndoControl TestUndoPreControl TestDeleteViewsControl TestSetUndoFlagViewsControl TestBackColorControl

2 applicate level

F2dCleanUpFindDrawingSheet F2dCleanUpFindViews

F2dCleanUpFindSameColorEntity

F2dCleanUpFindCenterLineEntity

F2dCleanUpFindOnesideOpenConnectEntity

F2dCleanUpFindOneSideOpenEntity2

F2dCleanUpClearDrawingEntity

F2dCleanUpDeleteDrawingEntity

F2dCleanUpDeleteViews APPENDIX D

F2dCleanUpUndo F2dCleanUpUndoPre

F2dCleanUpPickUpAllOneSideOpenEntity F2dCleanUpPickUpAllArcEntity F2dCleanUpFindLongestEntity F2dCleanUpFindGroup

F2dCleanUpCheckEntityExist

F2dCleanUpFindViewsRelation F2dCleanUpFindRelativeViews

F2dCleanUpFindOutSideLoop F2dCleanUpFindInSideEntity

F2dCleanUpFindOneSideOpenEntity F2dCleanUpFindGroupMinX F2dCleanUpFindOutNextEntity F2dCleanUpFindOnesideOpenConnectEntityNext

F2dCleanUpFindArrow

F2dCleanUpFindArrowOneSideOpenEntity

F2dCleanUpFindArrowEndEntity

F2dCleanUpCheckStartEndEntity

F2dCleanUpComputeCenterLineAngle

F2dCleanUpFindArrowCenterEntity

F2dCleanUpChangeArrowVertexFlag

F2dCleanUpArcAngle

F2dCleanUpLineAngle

F2dCleanUpLineLength

F2dCleanUpChangeColor

F2dCleanUpChangeDrawingColor

F2dCleanUpChangeChildrenFlag

F2dCleanUpChangeDrawingFlag

F2dCleanUpClearDrawingKeepFlags

F2dVector F2dVectorUnit F2dDistancePointLine __ ====_™====_™====_™===_==========.

#ifdef VELLUM

#include "ck_cdlc.h"

// VCI_VLM.DLL include files

#include "system.h"

// CadKey emulation files #include "vim ck.h" APPENDIX D

#endif

#ifdef CKWIN

#include "ck_cdlc.h"

#include "ck_cde.h"

#include <ckcdll.h>

#include <ck_attr.h>

#define ck_setattr2 ck_setattr

#endif

#ifdef AUTOCAD #include <ACAD_VCI.H> #endif

// 2d3d files

#include "view3D.h"

//#include "Proto.h" // includes prototypes of all functions.

#define VLM_ID_2DT03D 32 #define Regular 0 #define ThreeViews 1 //#define SECURITY

// Some definitions

//

#undef DEBUG

//

// Following is the definition of the global variables

//

#ifdef CKWIN int TopViewNum, IsometricViewNum;

// These variables are used only with CADKEY when changing to

TopView or Isometric view.

#endif

// by Ji Feb.14, 1996 double F2dCleanUpTolerance = 0.01;

// undo flag ( set a No. for views ) extern int F2dCleanUpExcutedNo = 0; extern int WorkLevel, TrimLevel, VW3DLevel;

// Three layers are useD: WorkLevel is used to highlight views and display temporary drawings.

// TrimLevel is used to store the trimmed drawing.

// VW3DLevel is used to store the projected APPENDIX D

3D drawing.

// these are the four drawings used. extern struct drawing input_drawing, projected_drawing, trim_drawing, final_drawing;

// Three drawing structures are constructed:

// InputDrawing is constructed from the input entities

// taken from from all the layers that are turned on.

// TrimmedDrawing is constructed from the trimmed entities. Once the

// Trimmed Drawing is constructed, the Input Drawing is not used.

// ProjectedDrawing is constructed from the entities projected in a single flange.

// If 'DoneProjection' is selected, then these entities are saved on VW3DLevel.

// If 'UndoProjection' is selected, these entities are discarded. extern struct planar_flange ^♦AllFlanges; // this is a pointer to the list of flanges extern char Messages [MaxMessages] [80] ;

// This array stores the messages read from the message file. extern int message_flag;

// This variable is assigned to '1' if the messages were read successfully from the message file.

// Otherwise, it is set to '0' . extern int number_of_views;

// This vairbale stores the number of disconnected components discovered in the ♦inputs drawing. extern int number_of_entities;

// This variable stores the number of entities loaded for the current drawing. extern int number_trimmed_entities;

// This variable stores the number of entities loaded for the trimmed drawing. extern int n_edges;

// This variable stores the number of connecting edges between the trimmed entities. extern double MatchTolerance, ConnectivityTolerance;

// ConnectivityTolerance is used to determine whether two entities APPENDIX D are close enough to be considered connected.

// Matching Tolerance is used to determine whether two entities on different views are matching. extern double ArrowHeadAngle, ArrowHeadLength;

// These variuables define the structure of an arrow for the 2D clean up funciton. extern double view_boundaries [MaxNumberOfViews+1] [4] ;

// This array is used to store the bounding box of each connected component of the input drawing.

// once the views are detected, several components might be mapped to a single view. extern int type_of_view [MaxNumberOfViews+1] ;

// This array maps every view to its type, namely to one of Top,

Left, Right, Front, Back view.

// Its content is valid only after 'DoneViews' was called.

//int view_rotation_flag[MaxNumberOfViews+1] ;

// This array is used for marking rotated views.

// If the i'th element is ' 1 ' then the view is rotated.

//extern static char original_lvmask [32] ;

// At the beginning we save the (on,off) status of all the layers.

//extern static char twoDmask[32] ;

//extern static char threeDmask [32] ;

// The masks used to display the 2D and 3D models.

//extern static int original_worklevel ;

// Specifies the original work level of the user when fold is started

// static int original_setup_available = 0 ;

// true iff original (layer) setup is available, ie . can be restored

extern int dialog_open_flag = 0;

// This variable is a flag which specifies whether some dialog boxes were opened.

extern int AtrimLineTypeSwitch, AtrimColorSwitch, AtrimWidthSwitch,

AtrimCenterLineSwitch;

// these switches are used to specify in which cases AutoTrim will trim lines. APPENDIX D

//struct node ^♦F2dCleanUpInputEntity(char ^♦, struct drawing ^♦) ; //static int select_flange (struct resbuf ^♦rb) ; typedef struct Vector { double x,y,z;

}; extern double BBminX,BBmaxX,BBminY,BBmaxY,BBminZ,BBmaxZ;

/*

Function:

Name : TestFindBoundary

Description:

It is a manual function. When user selected a boundary's entity of a view, the function would find all of the connected entities for the view.

Return Value:

< 0 : Error = 0 : No proccess

> 0: Completed successfully; the Number is the counter of found entities.

Parameters : Subfunctions: int F2dCleanUpFindArrow look for arrow type entity int F2dCleanUpFindArrowEndEntity find arrow end line int F2dCleanUpFindArrowCenterEntity find arrow center line int F2dCleanUpFindArrowOneSideOpenEntity find one side open entity for arrow type int F2dCleanUpComputeCenterLineAngle compute center line angle int F2dCleanUpCheckStartEndEntity check start and end entities int F2dCleanUpFind0utSideLoop look for out side loop entities int F2dCleanUpFind0utNextEntity look for connected entity with out side loop entities struct node ^♦F2dCleanUpFindGroupMinX look for minimum X-value for a entities group struct node_list ^♦F2dCleanUpFindGroup look for a group entities connected with input entity APPENDIX D struct node ^♦F2dCleanUpInputEntity ( old function ) look for a selected entity on screen int F2dCleanUpLineAngle compute a line type entity's angle int F2dCleanUpArcAngle compute a arc type entity's angle, int F2dCleanUpChangeChildrenFlag change entity's children flags int F2dCleanUpChangeArrowVertexFlag change arrow top vertex's flags and set arrow center line coordinates int F2dCleanUpChangeColor change entity' s color int F2dCleanUpFindOneSideOpenEntity look for one side open entities and they are vertical with arrow type entities double F2dCleanUpLineLength compute line type entity's length

*/ int TestFindBoundary()

{

// external functions int F2dCleanUpFindArrow() ; int F2dCleanUpChangeColor () ; int F2dCleanUpFindOneSideOpenEntity() ; int F2dCleanUpFind0utSideLoop() ; int F2dCleanUpChangeDrawingColor () ; struct node ^♦F2dCleanUpInputEntity0 ; struct node ♦F2dCleanUpFindGroupMinX() ; struct node_list ♦F2dCleanUpFindGroup() ;

// entities counter and flags int count = 0, assistant_line_flag = 0; // link list and temp list struct node_list

♦Entities_change_color_listl; // selected entity and temp entities struct node

♦entity,

^♦MinXvalueEntity;

// Begin li¬ lt Step_l:

// check the number of entites

// get count of all entities

// if an error occurred, exit immidiately. APPENDIX D if (number_of_entities<0) return -1; for( ; ; ) { /

// Step_2 : select one entity entity = F2dCleanUpInputEntity( "Select a Boundary Entity

// check a flag for outside loop process if ( count < 1 ) continue;

MinXvalueEntity = F2dCleanUpFindGroupMinX (

Entities_change_color_listl) ; if ( MinXvalueEntity == 0 ) continue;

// Step_7:

// find outside loop entities

F2dCleanUpFind0utSideLoop( MinXvalueEntity ) ; //

// Step_X:

// change color to Magenta for a selected boundary

// turn off input levels ck_levels(CK_OFF, 1,255) ; #ifdef CKWIN

F2dCleanUpChangeColor(Entities_change color listl) ; #endif #ifdef VELLUM

F2dCleanUpChangeDrawingColor(&trim_drawing) ; #endif

// clear a flag for outside loop process assistant_line_flag = 0;

// turn ON the level used to store the trimmed entities. ck_levels (CK_ON,TrimLevel, TrimLevel) ; #ifdef CKWIN

8 APPENDIX D if ( ^♦SwitchRedraw == 1 ) ck redraw( CK PRIME VP ) ; #endif

} return count;

} /*

Function:

Name : F2dCleanUpFindGroup

Description: look for a group entities connected with input entity Return Value:

= 0: Completed successfully; Parameters :

Input :

Entity: a entity

Output:

NodeList: a node list connected with input entity

*/ struct node_list ♦F2dCleanUpFindGroup (struct node ^♦Entity, int ^Cnt

)

{

// entities counter and flags int count = 0;

// link list and temp list struct node_list

♦EntitiesListl,

// initialize

EntitiesListl = ( struct node_list ♦ ) calloc ( 1, sizeof ( struct node_list

) ) ; open_list = ( struct node_list ^♦ ) calloc ( 1, sizeof ( struct node list ) ) ;

// set a flag for the boundary entity count = 0;

// set pointer to first entity open_list->car = Entity;

// set a open flag to the node APPENDIX D open_list->car->flags = 1; // set link address open_list->cdr = NULL;

// set the pointer to output node list EntitiesListl->car = open_list->car; EntitiesListl->cdr = open_list->cdr;

// Step_3 :

// find connected entities for ( temp_list = open_list;

( temp_list

// && (temp_list->cdr != 0)

> ;

⁾ { if ( temp_list->car == NULL ) break;

// get a pointer from open list

Next_entity = temp_list->car;

// set a closed flag to the node

Next_entity->flags = 2;

// close the node ( delete the node from open list ) open_list = open_list->cdr;

// count the number of entities count++;

// look for first connected entity whose flags=0. for (temp_listl=Next_entity->connected_nodes; // temp_listl && temp_listl->car->flags==l; ( temp_listl!=0 && temp_listl->car!=0 ); temp_listl=temp_listl->cdr) { if ( temp_listl->car == NULL ) break; // if found an unmarked connected entity if (temp_listl

&& temp_listl->car->flags==0

& & E n t i t y - > l i n e _ t y p e = = temp_listl->car->line_type

&& Entity->color == temp_listl->car->color & & E n t i t y - > i s_t h i c ne s s = = temp__listl->car->is thickness

^)" {

// allocate memory for open list element open_temp_list = ( struct node_list ♦ ) calloc ( 1, sizeof ( struct node_list ) ) ;

// allocate memory for output node list element

Out_temp_list = ( struct node_list ♦ ) calloc ( 1, sizeof ( struct node_list )) ;

// add this entity to the open list open_temp_list->car = temp_listl-_>car;

10 APPENDIX D

// add this entity to the output node list Out_temp_list->car = temp_listl->car;

// set a open flag to the node open_temp_list->car->flags = 1;

// connect to the open list open_temp_list->cdr = open_list;

// move the pointer of open list to the top open_list = open_temp_list;

// connect to the output node list

Out_temp_list->cdr = EntitiesListl;

// move the pointer of output node list to the top

EntitiesListl = Out_temp_list;

}

// assign value to the loop variable temp list = open list;

}

^♦Cnt = count; return EntitiesListl;

Function:

Name : F2dCleanUpArcAngle

Description: compute a arc type entity's angle. Return Value:

RotateVertexFlag: direction flag =1: vertexl is rotate point =2 : vertex2 is rotate point Parameters:

Input :

1 OriginVertex: rotationary vertex

2 Entity: arc type entity Output :

1 EntityAngle: direction angle ( radian )

2 EntityAngleAnother: another side direction angle

*/ int F2dCleanUpArcAngle ( // input

11 APPENDIX D struct planar_vertex ^♦OriginVertex, struct node ^♦Entity,

// output double ^♦EntityAngle, double ♦EntityAngleAnother )

/ {/ return value int RotateVertexFlag=0;

// length and angle double Tolerance = 0.0;

// set tolerance

Tolerance = F2dCleanUpTolerance;

// select connected side of entity if ( Entity->vertexl == OriginVertex ) {

// get arc start-angle

// ( direction : from circle-center to start-point

)

RotateVertexFlag = 1;

♦EntityAngle = ( Entity->AngBegin ) ^♦ pi / 180.0;

// change angle to oppsite direction

// ( direction : from start-point to circle-center

)

^♦EntityAngle = ♦EntityAngle + pi;

// check angle if ( ^♦EntityAngle >(2.0*pi)) ^♦EntityAngle = ♦EntityAngle

- 2.0^pi; if ( fabs (^♦EntityAngle-2.0*pi) < Tolerance ) ♦EntityAngle = 0.0;

// compute tangent line angle for the arc end-angle

// ( direction : from start-point to end-point )

^♦EntityAngle = ^♦EntityAngle + 1.5 ^♦ pi;

// check angle if ( ^♦EntityAngle >(2.0^pi)) ^♦EntityAngle = ♦EntityAngle

- 2.0^pi; if ( fabs (*EntityAngle-2.0*pi) < Tolerance ) ♦EntityAngle = 0.0;

// get another side data

// get arc start-angle

// ( direction : from circle-center to end-point

)

^♦EntityAngleAnother = ( Entity->AngEnd ) ♦ pi / 180.0;

// change angle to oppsite direction

// ⁽ direction : from end-point to circle-center

)

^♦EntityAngleAnother = ^♦EntityAngleAnother + pi; // check angle if ( ^♦EntityAngleAnother > 2.0^pi ) ^♦EntityAngleAnother = ^♦EntityAngleAnother - 2.0 ♦ pi;

12 APPENDIX D else if ( fabs (♦EntityAngleAnother-2.O^pi) < Tolerance ) ♦EntityAngleAnother = 0.0;

// compute tangent line angle for the arc end-angle // ( direction : from end-point to start-point ) ♦EntityAngleAnother = ^♦EntityAngleAnother + 0.5 ^♦ pi;

// check angle if ( ♦EntityAngleAnother > 2.0^pi ) ♦EntityAngleAnother = ♦EntityAngleAnother - 2.0 ^♦ pi; else if ( fabs (^♦EntityAngleAnother-2. O^pi) < Tolerance ) ♦EntityAngleAnother = 0.0;

} else {

// get arc end-angle

// ( direction : from circle-center to end-point

)

RotateVertexFlag = 2;

♦EntityAngle = ( Entity->AngEnd ) ^♦ pi / 180.0;

// change angle to oppsite direction

// ( direction : from end-point to circle-center

)

^♦EntityAngle = ^♦EntityAngle + pi; if ( ♦EntityAngle > 2.0^pi ) ^♦EntityAngle = ^♦EntityAngle

- 2.0*pi; if ( fabs (^♦EntityAngle - 2.0^♦pi) < Tolerance ) ♦EntityAngle = 0.0;

// compute tangent line angle for the arc end-angle // ( direction : from end-point to start-point ) ^♦EntityAngle = ^♦EntityAngle + 0.5 ^♦ pi; if ( ^♦EntityAngle > 2.0^pi ) ^♦EntityAngle = ^♦EntityAngle

- 2.0^pi; if ( fabs (^♦EntityAngle - 2.0^pi) < Tolerance ) ^♦EntityAngle = 0.0;

// get another side data

// get arc start-angle

// ( direction : from circle-center to start-point

♦EntityAngleAnother = ( Entity->AngBegin ) ^♦ pi /

180.0,

// change angle to oppsite direction

// ( direction : from start-point to circle-center

♦EntityAngleAnother = ^♦EntityAngleAnother + pi; if ( ♦EntityAngleAnother > 2. O^ i ) ♦EntityAngleAnother = ^♦EntityAngleAnother - 2.0 ^♦ pi; if ( fabs (♦EntityAngleAnother - 2.0^pi ) < Tolerance ) ♦EntityAngleAnother = 0.0;

// compute tangent line angle for the arc start-angle

// ( direction : from start-point to end-point )

13 APPENDIX D

♦EntityAngleAnother = ^♦EntityAngleAnother + 1.5 ^♦ pi^; if ( ♦EntityAngleAnother > 2.0^pi ) ♦EntityAngleAnother = ♦EntityAngleAnother - 2.0 ^♦ pi; if ( fabs (♦EntityAngleAnother - 2.0^pi ) < Tolerance ) ♦EntityAngleAnother = 0.0;

}

// the end of this function return RotateVertexFlag; }

/*

Function:

Name : F2dCleanUpLineAngle

Description:

It is a function for control how to compute a line type entity's angle. Return Value:

Input:

1 OriginVertex: rotationary vertex

2 Entity: line type entity Output:

1 EntityAngle: direction angle ( radian )

2 EntityAngleAnother: another side direction angle

__ int F2dCleanUpLineAngle( // input struct planar_vertex ^♦OriginVertex, struct node ^♦Entity, // output double ^♦EntityAngle, double ^♦EntityAngleAnother )

{

// return value int RotateVertexFlag=0; // selected entity and temp entities struct planar_vertex ^♦VertexTemp, ^♦VertexAnotherTemp; // length and angle double diff_x = 0.0,

14 APPENDIX D diff_y = 0.0, Tolerance = 0.0; // set tolerance Tolerance = F2dCleanUpTolerance; // select connected side of entity if ( Entity->vertexl == OriginVertex ) {

// get vertexl->vertex2

RotateVertexFlag = 1;

VertexTemp = Entity->vertexl;

VertexAnotherTemp = Entity->vertex2; else { // get vertex2->vertexl RotateVertexFlag = 2;

VertexTemp = Entity->vertex2;

VertexAnotherTemp = Entity->vertexl;

}

// compute angle if ( fabs ( VertexTemp->X - VertexAnotherTemp->X ) < Tolerance

) (

// angle = 0.5 ^♦ pi if ( VertexTemp->Y < VertexAnotherTemp->Y )

^♦EntityAngle = 0.5 ^♦ pi;

// angle = 1.5 ^♦ pi else ♦EntityAngle = 1.5 ♦ pi; else { // -pi < angle < pi diff_y = VertexAnotherTemp->Y - VertexTemp->Y; if ( fabs ( diff_y ) > Tolerance ) { diff_x = VertexAnotherTemp->X - VertexTemp->X; ^♦EntityAngle = atan2 ( diff_y, diff_x ) ; // -pi < angle < 0.0 if ( ^♦EntityAngle < 0.0 ) ^♦EntityAngle ♦EntityAngle + 2.0^#pi; else { // angle = 0.0 or pi if ( VertexAnotherTemp->X > VertexTemp->X ) ^♦EntityAngle = 0.0; else ^♦EntityAngle = pi;

if ( fabs (^♦EntityAngle - 2.0*pi) < Tolerance ) ^♦EntityAngle =

0.0;

// get another side angle

^♦EntityAngleAnother = ^♦EntityAngle + pi;

// check angle if ( ^♦EntityAngleAnother > ( 2.0 ^♦ pi ) ) {

^♦EntityAngleAnother = ^♦EntityAngleAnother - 2.0 ^♦ pi, if ( fabs (^♦EntityAngleAnother - 2.0^♦pi) < Tolerance ) { ^♦EntityAngleAnother = 0.0;

15 APPENDIX D

}

// the end of this function return RotateVertexFlag;

Function:

Name : F2dCleanUpChangeColor

Description:

It is a function for changing entity's color. Return Value:

= 0: Completed successfully; Parameters :

Input :

Entities_change_color_list : a node list that will be changed color

*/ int F2dCleanUpChangeColor ( struct node_l i st

♦Entities change color list)

{

// link list and temp list struct node_list

^♦Color_temp_list; // selected entity and temp entities struct node

^♦Next_entity; // entity's specifitys CK_ENTATT attr; // specificity of entities int CountPrompt = 0; char message [64] ;

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Start Change Color Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif

#ifdef VELLUM

// clear views highlighting ck_erase_level ("VW_2D") ; // ck_erase_level ("VW TRIM"); #endif

// check all of entities on the list for⁽ Color_temp_list=Entities_change_color_list;

16 APPENDIX D

Color_temp_list;

Color_temp_list = Color_temp_list->cdr) { if ( Color_temp_list->car == NULL ) break; Next_entity = Color_temp_list->car; switch ( Next entity->flags )

{ case 1: attr.color = CK_GREEN; ck_setattr( Next_entity->id, CK_COLOR, NULL,

&attr break; case 3 : attr.color = CK_GRAY; ck_setattr( Next_entity->id, CK_COLOR, NULL,

&attr ) break; case 4 : attr.color = CK_MAGENTA; ck_setattr( Next_entity->id, CK_COLOR, NULL,

&attr ) ; break;

}

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Finished change color Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif return 0;

} /*

Function:

Name : F2dCleanUpChangeDrawingColor

Description:

It is a function for changing entity's color. Return Value:

= 0: Completed successfully; Parameters:

Input :

1 In_drawing: drawing that will be changed color __ int F2dCleanUpChangeDrawingColor(struct drawing ^♦In_drawing)

17 APPENDIX D

// extern functions void draw_entity0 ;

// selected entity and temp entities struct node ^♦entity; int CountPrompt = 0; char message [64] ;

// entity's specifitys

CK_ENTATT attr; // specificity of entities

#ifdef VELLUM

// clear views highlighting ck_erase_level ("VW_2D") ; // ck_erase_level ( "VW TRIM") ; CountPrompt++; sprintf ( message, "Start Change Color Sheet Entity %d", CountPrompt ) ; ck_prompt ( message ) ; #endif

// check all of entities on the list for( entity=In_drawing->objects; entity; entity = entity->next) { switch ( entity->flags )

#ifdef VELLUM case 0 : draw_entity( entity, entity->color ); break; case 1: draw_entity( entity, CK_GREEN ) ; break; case 2 : draw_entity( entity, entity->color ) ; break; case 3 : draw_entity( entity, CK_BLUE ) ; break; case 4 : draw_entity( entity, CK_MAGENTA ) ; break; case 5 : draw_entity( entity, entity->color ) ; break; case 6 : draw_entity( entity, CK_RED ) ; break; case 7 : draw_entity( entity, CK_YELLOW ) ; break;

18 APPENDIX D case 8 : draw_entity( entity, entity->color ) ; break; case 9: draw_entity( entity, entity->color ) ; break; case 10: draw_entity( entity, CK_CYAN ) ; break; case 11: draw_entity( entity, entity->color ) ; break; #endif #ifdef CKWIN case 1 : attr.color = CK_L_GREEN; ck_setattr( entity->id, CK_COLOR, NULL, fcattr

) ; break; case 3 : attr.color = CK_L_GRAY; ck_setattr( entity->id, CK_COLOR, NULL, kattr

) ; break; case 4 : attr.color = CK_MAGENTA; ck_setattr( entity->id, CK_COLOR, NULL, &attr

) ; break; case 6 : attr.color = CK_L_RED; ck_setattr( entity->id, CK_COLOR, NULL, &attr

) ; break; case 7 : attr.color = CK_YELLOW; ck_setattr( entity->id, CK_C0LOR, NULL, &attr

) ; break; case 10: attr.color = CK_L_CYAN; ck_setattr( entity->id, CK_COLOR, NULL, kattr

) ; break; #endif

} if ( entity->next == NULL ) break;

} #ifdef VELLUM

19 APPENDIX D

CountPrompt++; sprintf ( message, "Finished change color Sheet Entity %d", CountPrompt ) ; ck_prompt ( message ) ; #endif return 0;

} /*

Function:

Name : F2dCleanUpUndoPre

Description:

To prepare data flags for undo function. Return Value :

= 0: Completed successfully; Parameters:

Input :

1 In_drawing: drawing that will be changed status

*/ int F2dCleanUpUndoPre (struct drawing ⁺In drawing)

{

// selected entity and temp entities struct node ♦entity;

// set flag F2dCleanUpExcutedNo = 2; // check all of entities on the list for( entity=In_drawing->objects; entity; entity entity->next) ( entity->status = entity->flags; if ( entity->next == NULL ) break;

} return 0;

} /*

Function:

Name : F2dCleanUpUndo

Description:

To undo. Return Value:

= 0: Completed successfully; Parameters:

Input:

20 APPENDIX D 1 In_drawing: drawing that will be undoed status

*/ int F2dCleanUpUndo (struct drawing ^♦In_drawing)

{

// selected entity and temp entities struct node ^♦entity;

// flag int NoChangeFlag = 0;

// check flag if ( F2dCleanUpExcutedNo == 0 ) return NoChangeFlag;

// check all of entities on the list for( entity=In_drawing->objects; entity; entity entity->next) { if ( entity->status != entity->flags ) { NoChangeFlag = 1; entity->flags = entity->status;

} if ( entity->next == NULL ) break;

}

// check views flag if ( F2dCleanUpExcutedNo == 1 ) In_drawing->views

NULL;

// reset flag for views

F2dCleanUpExcutedNo = 0; return NoChangeFlag;

} /*

Function:

Name : F2dCleanUpChangeDrawingFlag

Description:

It is a function for changing entity's flag. Return Value :

= 0: Completed successfully; Parameters:

Input :

1 In_drawing: drawing that will be changed color __ int F2dCleanUpChangeDrawingFlag(struct drawing ♦In_drawing)

// selected entity and temp entities

21 APPENDIX D struct node ^♦entity; int CountPrompt = 0; char message [64] ;

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Start Flag Sheet Entity %d" CountPrompt ) ; ckjprompt ( message ) ; #endif

// check all of entities on the list for( entity=In_drawing->objects; entity; entity = entity->next) { if ( entity->flags == 4 ) entity->flags = 3; #ifdef VELLUM

//CountPrompt++;

//sprintf ( message, "Flag Sheet Entity %d", CountPrompt ) //ck_prompt ( message ) ; #endif if ( entity->next == NULL ) break;

} #ifdef VELLUM

CountPrompt++; sprintf ( message, "Finished Flag Sheet Entity %d" CountPrompt ) ; ck_prompt ( message ) ; #endif return 0;

} /*

Function:

Name : F2dCleanUpFindSameColorEntity

Description: find same color entity. Return Value:

= 0 : Completed successfully; Parameters:

Input :

1 In_entity: an entity

2 ln_drawing: drawing that will be changed color -- =======--===.-=========.====== int F2dCleanUpFindSameColorEntity( struct node ^♦In_entity, struct drawing ♦In_drawing)

22 APPENDIX D

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Start Find same color Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif

// check all of entities on the list for( entity=In_drawing->objects; entity; entity = entity->next) { if ( entity->color == In_entity->color

&& entity->flags == 0) entity->flags = 3; if ( entity->next == NULL ) break;

}

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Finished find same color Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif return 0;

} /*

Function:

Name : F2dCleanUpClearDrawingEntity

Description:

It is a function for return entity's color. Return Value:

= 0: Completed successfully; Parameters:

Input:

1 In_drawing: drawing that will be changed color

i *nt^"F2dCleanUpClearDrawingEntity(struct drawing ^♦In drawing)

{

// selected entity and temp entities struct node ^♦entity;

23 APPENDIX D

// entity's specifitys

CK ENTATT attr; // specificity of entities int CountPrompt = 0; char message [64] ;

#ifdef VELLUM

// clear views highlighting ck_erase_level ("VW_2D" ) ; // ck_erase_level ("VW_TRIM") ; // CountPrompt++;

// sprintf ( message, "Start Clear Entity %d", CountPrompt // ck_prompt ( message ) ; #endif

// check all of entities on the list for ( entity=In_drawing->objects; entity; entity = entity->next) { if ( entity == NULL ) break; switch ( entity->flags ) { case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7: case 8: case 9: case 10 entity->flags = 0; draw_entity( entity, entity->color ) ; //attr.color = entity->color;

//ck_setattr( entity->id, CK_COLOR, NULL, &attr ) ; default: break;

#ifdef VELLUM

// CountPrompt++;

// sprintf ⁽ message, "Finished clear Entity d", CountPrompt

// ck_prompt ( message ) #endif

// clear views connected

In_drawing->views = NULL;

24 APPENDIX D return 0;

}

Function:

Name : F2dCleanUpClearDrawingKeepFlags

Description:

To return entity's color but keep flags. Return Value :

= 0: Completed successfully; Parameters:

Input :

1 In_drawing: drawing that will be changed color __ int F2dCleanUpClearDrawingKeepFlags (struct drawing ^In drawing)

{

// selected entity and temp entities struct node ^♦entity;

// entity's specifitys

CK_ENTATT attr; // specificity of entities int CountPrompt = 0; char message [64] ;

#ifdef VELLUM

// clear views highlighting ck_erase_level ( "VW_2D") ; #endif

// check all of entities on the list for ( entity=In_drawing->objects; entity; entity = entity->next) { switch ( entity->flags )

{ case 0 case 1 case 2 case 3 case 4 case 5 case 6 case 7 case 8 case 9 case 10 : // center line

25 APPENDIX D draw_entity( entity, entity->color ) ;

//attr.color = entity->color;

//ck setattr( entity->id, CK_COLOR, NULL, kattr

) ;

} if ( entity->next == NULL ) break;

} return 0;

} /*

Function:

Name : F2dCleanUpDeleteViews

Description:

It is a function for delete views. Return Value:

= 0 : Completed successfully; Parameters:

Input :

1 In_drawing: drawing

*/ int F2dCleanUpDeleteViews (struct drawing ^AIn drawing)

{

// check drawing if ( In_drawing == NULL ) return 0;

// delete views ( Not release memory ) In_drawing->views == NULL; return 0;

}

Function:

Name : F2dCleanUpDeleteDrawingEntity

Description:

It is a function for delete entities. Return Value:

= 0: Completed successfully; Parameters:

Input :

1 In_drawing: drawing

_*7 ============ int F2dCleanUpDeleteDrawingEntity(struct drawing ♦In_drawing)

26 APPENDIX D

// selected entity and temp entities struct node ^♦entity; struct group ^♦Gp; // group struct node_list ^♦ETL; // Entity Temp List int CountPrompt = 0; char message [64] ;

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Start Delete Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif

// check all of entities on the list for ( entity=In_drawing->objects; entity; entity = entity->next) { if ( entity->flags == 3 | j entity->flags == 6

I j entity->flags == 0 j j entity->flags == 10 ) ( #ifdef CKWIN ck_delent ( entity->id ) ; #endif

#ifdef VELLUM

// remove_entity ( entity, In_drawing ) ;

#endif entity->flags = 99;

} if ( entity->next == NULL ) break;

} if ( In_drawing->views == NULL ) return 0; // unuse group entities for( Gp=In_drawing->views; Gp; Gp = Gp->next) { if ( Gp == NULL ) break; if ( Gp->index != 0 ) continue; for ( ETL = Gp->entities; ETL; ETL = ETL->cdr ) { #ifdef CKWIN ck_delent ( ETL->car->id ) ; #endif #ifdef VELLUM

// remove_entity ( ETL->car, In_drawing ) ;

#endif

ETL->car->flags = 99; if ( ETL->cdr == NULL ) break;

} if ( Gp->next == NULL ) break;

}

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Finished Delete Entity %d" , CountPrompt ) ;

27 APPENDIX D ck_prompt ( message ) ;

#endif return 0;

} /*

Function:

Name : F2dCleanUpCheckEntityExist

Description: check an entity for current drawing if the entity is not same to database, drawing was changed

Return Value:

= 0 : No change;

= 1: Changed; Parameters:

Input :

1 In_drawing: drawing

__ int F2dCleanUpCheckEntityExist (struct drawing ^♦In_drawing)

// selected entity and temp entities struct node ^♦entity;

// define int ReturnCode = 0, etype = 0; double Xl=0.0,yl=0.0,zl=0.0,x2=0.0,y2=0.0,z2=0.0, rad=0.0,angl=0.0,ang2=0.0; // entity's specifitys CK_ENTATT attr; // specificity of entities

// check all of entities on the list if ( In_drawing->objects == NULL ) return 1; for( entity=In_drawing->objects; entity; entity = entity->next) ( if ( entity->flags == 99 ) continue; ReturnCode = ck_getentid( entity->id, ketype ) ; if ( ReturnCode == CK_N0_ENT ) return 1; else {

// check all of entities on the list switch ( entity->type )

{ case CK_ARC : ck_getarc ( entity- _>id ,

&xl , &yl , &zl ,

&ra d , &an g l , & ang 2 , 28 APPENDIX D

&attr ) ; break; case CK_LINE: ck_getline ( entity->id, &xl, δ-yl, ScZl,

&x2 , &y2 , &z2 , fcattr ) ; break;

} if ( attr. level == 191 ) return 0; else return 1;

} if ( entity->next == NULL ) break;

} return 1;

}

Function:

Name : F2dCleanUpChangeArrowVertexFlag

Description: to change arrow top vertex's flags and set arrow center line coordinates. Return Value:

= 0: Completed successfully; Parameters:

Input :

1 Flags: flag

2 TopVertex: arrow top vertex

3 Entity: arrow center line entity

*/ int F2dCleanUpChangeArrowVertexFlag(

// input int ^♦flags, struct planar_vertex ^♦TopVertex, struct node ^♦Entity)

{

// set flag to arrow top vertex TopVertex->flags = ^♦flags; // set arrow center line coordinates TopVertex->ArrowCenterLine_Xl = Entity->X1 TopVertex->ArrowCenterLine_Yl = Entity->Y1 TopVertex->ArrowCenterLine_Zl = Entity->Z1 TopVertex->ArrowCenterLine_X2 = Entity->X2 TopVertex->ArrowCenterLine_Y2 = Entity->Y2 TopVertex->ArrowCenterLine_Z2 = Entity->Z2 // end return 0;

29 APPENDIX D

Function:

Name : F2dCleanUpFindArrowCenterEntity

Description: to find arrow center line and set flags to arrow's members. Return Value:

= 0 : No arrow;

= 1 : Arrow Parameters :

Input :

1 Nodes : connected node

2 TopVertex: arrow top vertex

3 StartEntity: arrow start line entity

4 EndEntity: arrow end line entity

5 CenterLineAngleTemp: arrow center line angle

*/ int F2dCleanUpFindArrowCenterEntity( // input struct node_list ^♦Nodes, struct planar_vertex ^♦TopVertex, struct node ^♦StartEntity, struct node ^♦EndEntity, double ^♦CenterLineAngleTemp )

{

// external functions int F2dCleanUpLineAngle () ; int F2dCleanUpChangeChildrenFlag() ; int F2dCleanUpChangeArrowVertexFlag() ; double F2dCleanUpLineLength() ; // nodes struct node ^♦CenterEntity; // link list and temp list struct node_list ^♦temp_list; // flags int ReturnFlag = 0, ArrowFlags = 3; // double double CenterParentLength = 0.0, CenterParentAngle = 0.0, CenterParentAngleAnother = 0.0; double Tolerance = 0.0;

// set value

Tolerance = F2dCleanUpTolerance;

// look for arrow center line

30 APPENDIX D for ( temp_list = Nodes; temp_list; temp_list = temp_list->cdr) { // check pointer address if ( StartEntity == temp_list->car

I j EndEntity == temp_list->car ) continue; // get new entity CenterEntity = temp_list->car;

// check flags

// if ( CenterEntity->flags != 2

II ScS CenterEntity->flags != 3 ) continue; if ( CenterEntity->type != CK_LINE ) continue; if ( CenterEntity->line_type != StartEntity->line_type && CenterEntity->color !

StartEntity- > ■c )lor

& & CenterEntity- >is__thickness ! =

StartEntity- >is_thickness ) continue; if ( CenterEntity->no_connected_l < 2

&& CenterEntity- >no_connected_2 < 2 ) continue; // check length

Cent erParent Length = F2dCleanUpLineLengt h (

CenterEntity- >parent ) ; if ( CenterParentLength < Tolerance ) continue; // arrow central-entity angle F2dCleanUpLineAngle ( // input

TopVertex, CenterEntity->parent, // output

&CenterParentAngle, ScCenterParentAngleAnother ) ; // check the angle for central entity if ( fabs ( CenterParentAngle - ^♦CenterLineAngleTemp )

> Tolerance ) continue; // set flags for arrow's entities ReturnFlag = 1;

// look for arrow entity's children F2dCleanUpChangeChildrenFlag( StartEntity->parent ) ; F2dCleanUpChangeChildrenFlag( EndEntity->parent ) ; F2dCleanUpChangeChildrenFlag( CenterEntity->parent ) ; // set flag to arrow top vertex F2dCleanUpChangeArrowVertexFlag( SArrowFlags,

TopVertex, CenterEntity ) ; return ReturnFlag; if ( temp list->cdr == NULL ) break;

} return ReturnFlag;

} /*

31 APPENDIX D

Function:

Name : F2dCleanUpFindArrowOneSideOpenEntity

Description: to find one side open entity Return Value:

= 0: No open;

= 1: open

= 2: open

Parameters:

Input:

1 Nodes: connected node

i *n7t F2dCleanUpFindArrowOneSideOpenEntity( struct node_list ^♦Nodes

)

{

// link list and temp list struct node_list ^♦temp_list; // entities counter and flags int open_entity_flag = 0; // start for ( temp_list = Nodes; temp_list; temp_list = temp_list->cdr) { // ckeck another side connected' s counter if { temp_list->car->no_connected_l == 0

&& temp_list->car->no_connected_2 > 1 ) { open_entity_flag = 1; break;

} if ( temp_list->car->no_connected_2 == 0

&& temp_list->car->no_connected_l > 1 ) { open_entity_flag = 2; break;

} if ( temp_list->cdr == NULL ) break; return open_entity_flag;

Function:

Name : F2dCleanUpCheckStartEndEntity

Description: to check start and end entities Return Value:

= 0: No arrow;

32 APPENDIX D

= 1: good Parameters :

Input:

1 StartEntity: arrow start line entity

2 EndEntity: arrow end line entity

__ int F2dCleanUpCheckStartEndEntity( // input struct node ^♦StartEntity, struct node ^♦EndEntity )

{

// external functions double F2dCleanUpLineLength() ;

// double double arrow_start_parent_length = 0.0, arrow_end_parent_length = 0.0, arrow_tolerance = 0.0; // set tolerance arrow_tolerance = F2dCleanUpTolerance;

// start

// compute entities' parent length arrow_start_parent_length = F2dCleanUpLineLength( StartEntity

) ; arrow_end_parent_length = F2dCleanUpLineLength( EndEntity ) ;

// check length if ( arrow_start_parent_length < arrow_tolerance ) return 0; if ( arrow_end_parent_length < arrow_tolerance ) return 0;

// check tloerance if ( fabs ( arrow_start_parent_length - arrow_end_parent_length

)

> arrow_tolerance ) return 0; // check angle : if entities are parallel, STOP if (sqrt (fabs( ( ( StartEntity->X1 - StartEntity->X2 )

^♦ ( EndEntity->Y1 - EndEntity->Y2 )) -(( EndEntity->X1 - EndEntity->X2 )

♦ ( StartEntity->Y1 - StartEntity->Y2 )) ) ) < arrow_tolerance ) return 0;

// normal end return 1,-

} /*

Function:

Name : F2dCleanUpComputeCenterLineAngle

33 APPENDIX D

Description: to compute center line angle Return Value:

= 0 : No arrow;

= 1 : good Parameters :

Input :

1 TopVertex: arrow top vertex

2 StartEntity: arrow start line entity

3 EndEntity: arrow end line entity Output:

1 CenterLineAngle : arrow center line angle

i *n7t F2dCleanUpComputeCenterLineAngle{

// input struct planar_vertex ^TopVertex, struct node ^♦StartEntity, struct node ^♦EndEntity,

// output double ♦CenterLineAngle )

{

// external functions int F2dCleanUpLineAngle () ;

// entities counter and flags int ReturnFlag = 0;

// double double arrow_start_parent_angle = 0.0, arrow_end_parent_angle = 0.0, arrow_center_diff_angle = 0.0,

RotateAxisAngleAnotherTemp = 0.0, arrow_center_parent_angle_temp = 0.0, arrow_tolerance = 0.0; // set tolerance arrow_tolerance = F2dCleanUpTolerance;

// start

// compute central angle of arrow //

// angle_center = 0.5 ^♦ ( angle_start + angle_end ) // atan2 returns the arctangent of y/x // atan2 returns a value in the range -pi to pi radians

// get arrow start angle F2dCleanUpLineAngle (

// input

TopVertex, StartEntity,

// output

^&arrow_start_parent_angle, &RotateAxisAngleAnotherTemp) ;

34 APPENDIX D

// arrow end angle F2dCleanUpLineAngle (

// input

TopVertex, EndEntity,

// output

&arrow_end_parent_angle , SLRotateAxisAngleAnotherTemp) ; // get difference angle arrow_center_dif f_angle = f abs ( arrow_end_parent_angle arrow_start_parent_angle ) ; // same angle if ( arrow_center_dif f_angle < arrow__tolerance ) return 0 ; if ( arrow_center_dif f_angle < pi ) {

// diff < pi arrow_center_parent_angle_temp = 0. 5 ♦ ( arrow_end_parent_angle + arrow_start_parent_angle ) ; else {

// diff > pi ( take opposite angle ) arrow_center_parent_angle_temp = 0.5 ^♦ ( arrow_end_parent_angle + arrow_start_parent_angle )

+ PL

// angle > 2^pi if ( fabs ( arrow_center_parent_angle_temp - 2 . 0 ♦ pi )

< arrow_tolerance ) { arrow_center__parent_angle_temp = 0.0 ; else { if ( arrow_center_parent_angle_temp > ( 2.0 ^♦ pi )

{ arrow_center_parent_angle_temp = arrow center parent angle temp - 2.0 ^♦ pi;

} ^"

) ^} ^♦CenterLineAngle = arrow_center_j>arent_angle_temp; // normal end return 1;

} /*

Function:

Name : F2dCleanUpFindArrowEndEntity

Description: to find arrow end line Return Value:

= 0 : No arrow;

= 1 : Arrow

35 APPENDIX D

Parameters :

Input:

1 Nodes : connected node

2 TopVertex: arrow top vertex

3 StartEntity: arrow start line entity

^♦7 int F2dCleanUpFindArrowEndEntity(

// input struct node_list ^♦Nodes, struct planar_vertex ^♦TopVertex, struct node ^♦StartEntity )

{

// external functions int F2dCleanUpFindArrowOneSideOpenEntity() ; int F2dCleanUpComputeCenterLineAngle () ; int F2dCleanUpCheckStartEndEntity() ; int F2dCleanUpFindArrowCenterEntity() ;

// link list and temp list struct node_list ^♦temp_list;

// selected entity and temp entities struct node

^♦Next_entity,

^♦arrow_start_entity_parent,

^♦arrow_end_entity,

^♦arrow_end_entity_parent;

// entities counter and flags int assistant_line_flag = 0, ReturnFlag = 0; // double double arrow_center_parent_angle_temp = 0.0;

// start for( temp_list=Nodes; temp_list,- temp_list = temp_list->cdr) { if ( temp_list->car == NULL ) break; Next_entity = temp_list->car; arrow_end_entity = temp_list->car; // if ( Next_entity->flags != 2 ) continue; if ⁽ Next_entity->type != CK_LINE ) continue; if ⁽ Next_entity->line_type != StartEntity->line_type ) continue; if ⁽ Next_entity->color != StartEntity->color ) continue; f ( Ne x t _e n t i t y - > i s_ t h i c k n e s s ! = StartEntity->is_thickness ) continue; // set parent

36 APPENDIX D arrow_start_entity_parent = StartEntity->parent,^• arrow_end_entity_parent = Next_entity->parent; // look for broken entity' s another side ReturnFlag = F2dCleanUpFindArrowOneSideOpenEntity( arrow_end_entity_parent->children ) ; if ( ReturnFlag == 0 ) continue; // check parents entities ReturnFlag = F2dCleanUpCheckStartEndEntity(

// input arrow_start_entity_parent, arrow_end_entity_parent

) ; if ( ReturnFlag == 0 ) continue;

// look for central entity of arrow

ReturnFlag = F2dCleanUpComputeCenterLineAngle (

// input

TopVertex, arrow_start_entity_parent , arrow_end_entity_parent ,

// output

&arrow_center_parent_angle__temp ) ; if ( ReturnFlag == 0 ) continue; // get central entity of arrow assistant_line_flag = F2dCleanUpFindArrowCenterEntity(

// input

Nodes, TopVertex, StartEntity, arrow_end_entity,

&arrow_center_parent_angle_temp ) ; return assistant line flag;

} return assistant_line_flag;

} /*

Function:

Name : F2dCleanUpChangeChildrenFlag

Description:

It is a function for changing entity's children flags.

Return Value:

= 0: Completed successfully; Parameters:

Input :

*/ int F2dCleanUpChangeChildrenFlag(struct node ^♦Entity)

// temp list struct node_list ^♦Temp_list;

// check all of children of the entity

37 APPENDIX D for ( Temp_list=Entity->children; Temp_list;

Temp_list = Temp_list->cdr) { // change flag to light gray Temp_list->car->flags = 3; if ( Temp_list->cdr == NULL ) break;

} return 0;

Function:

Name : F2dCleanUpLineLength

Description:

It is a function for computing line type entity's length.

Return Value :

= length; Parameters:

Input:

Entity: a parent entity

^♦7 double F2dCleanUpLineLength(struct node ^♦Entity)

{

// defination double EntityLength=0.0; // compute entities' parent length

EntityLength = sqrt ( ( Entity->X1 - Entity->X2 ) ^♦ ( Entity->X1 - Entity->X2 )

+ ( Entity->Y1 - Entity->Y2 ) ^♦ Entity->Y1 - Entity->Y2 ) ); return EntityLength; }

Function:

Name : F2dCleanUpFindArrow

Description: look for arrow type entity and set flags to them. Return Value:

= 0: Completed successfully; Parameters:

Input :

1 Entities_list : a node list that will be

38 APPENDIX D checked

*/ int F2dCleanUpFindArrow( struct node_list ^♦Entities_list )

{

// external functions int F2dCleanUpLineAngle () ; int F2dCleanUpFindArrowEndEntity() ;

// link list and temp list struct node_list

^♦temp_nodesl,

^♦arrow_list,

^♦temp_list; // selected entity and temp entities struct node ^♦Next_entity; struct planar_vertex ^♦top_vertex;

// entities counter and flags int bug_flag = 0,

ReturnFlag = 0, assistant_line_flag = 0; int CountPrompt = 0; char message [64] ;

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Start Find Arrow Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif

// Step__l : look for one side open entity // arrow start entity for ( temp_list=Entities_list; temp_list; temp_list = temp_list->cdr) { Next_entity = temp_list->car; // if ( Next_entity->flags != 2 ) continue; if ( Next_entity->type != CK_LINE ) continue; if ( Next_entity->no_connected_l == 0

SL& Next_entity->no_connected_2 < 2 ) continue; if ( Next_entity->no_connected_l < 2

&& Next_entity->no_connected_2 == 0 ) continue; // select entity's side if ( Next_entity->no_connected_l == 0 ) { // clear bug_flag bug_flag = 0;

// look for broken entity's another side for( arrow_list = Next_entity->parent->children;

39 APPENDIX D arrow_list; arrow_list = arrow_list->cdr) {

// ckeck another side vertex's pointer if ( Next_entity->parent->vertex2 == arrow__list->car->vertex2 ) { bug_flag = 1; break;

} if ( arrow list->cdr == NULL ) break;

} if ( bug_flag == 1 ) ( top_vertex = Next_entity->parent->vertex2 ; t e m p _ n o d e s l = arrow__list->car->connected_nodes_2 ;

else (

// clear bug_flag bug_flag = 0;

// look for broken entity's another side for( arrow_list = Next_entity->parent->children; arrow_list; arrow_list = arrow_list->cdr) ( // ckeck another side vertex's pointer if ( Next_entity->parent->vertexl == arrow_list->car->vertexl ) ( bug_flag = 1; break;

} if ( arrow_list->cdr == NULL ) break; if ( bug_flag == 1 ) { top_vertex = Next_entity->parent->vertexl ; t e m p _ n o d e s l arrow list->car->connected nodes 1;

if ( bug_flag != 1 ) continue;

// get end entity of arrow

ReturnFlag = F2dCleanUpFindArrowEndEntity(

// input temp_nodesl, top_vertex, Next_entity ) ; if ( ReturnFlag == 1 ) assistant_line_flag = 1; if ( temp_list->cdr == NULL ) break;

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Finished find arrows Entity %d", CountPrompt ) ;

40 APPENDIX D ck_prompt ( message ) ; #endif return assistant line_flag;

} /*

Function:

Name : F2dCleanUpFindOneSideOpenEntity

Description: look for one side open entities and they are vertical with arrow type entities. Return Value:

= 0: Completed successfully and No found;

= NUM: Open entities' number Parameters:

Input :

1 Entities_list: a node list that will be checked

* / int F2dCleanUpFindOneSideOpenEntity ( struct node_list

♦Entities list )

{

// external functions int F2dCleanUpChangeChildrenFlag() ; // link list and temp list struct node_list ^♦temp_list; // selected entity and temp entities struct node ^♦Next_entity; // entities counter and flags int count = 0; // double double assistant_line_angle = 0.0, arrow_tolerance = 0.0; int CountPrompt = 0; char message [64] ;

// set tolerance arrow_tolerance = F2dCleanUpTolerance;

ttifdef VELLUM

CountPrompt++; sprintf ( message, "Start Find One side open Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif

41 APPENDIX D

// check all of entities which connected with outside loop for ( temp_list=Entities_list; temp_list; temp_list = temp_list->cdr) { if ( temp_list->car == NULL ) break; Next_entity = temp_list->car; // check flags // if ( Next_entity->flags 1 =2 ) continue; // check vertex' flag if ( Next_entity->vertexl->flags != 3

&& Next_entity->vertex2->flags != 3 ) continue; // check lines vertically if ( Next_entity->vertexl->flags == 3 ) { assistant_line_angle = fabs (

( Next_entity->X1 - Next_entity->X2 )

^♦ ( Next_entity->vertexl->ArrowCenterLine_Xl

-Next_entity->vertexl->ArrowCenterLine_X2

)

+ ( Next_entity->Y1 - Next_entity->Y2 )

^♦ ( Next_entity->vertexl->ArrowCenterLine_Yl

-Next_entity->vertexl->ArrowCenterLine_Y2

else { assistant_line_angle = fabs (

( Next_entity->X1 - Next_entity->X2 )

^♦ ( Next_entity->vertex2->ArrowCenterLine_Xl

-Next_entity->vertex2->ArrowCenterLine_X2

+ ( Next_entity->Y1 - Next_entity->Y2 )

^♦ ( Next_entity->vertex2->ArrowCenterLine_Yl

-Next_entity->vertex2->ArrowCenterLine_Y2

) ;

}

// check the angle for central entity if ⁽ sqrt (assistant_line_angle) > arrow_tolerance ) continue;

// look for children

F2dCleanUpChangeChildrenFlag( Next_entity->parent ) ; count++;

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Finished find one side open Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif

42 APPENDIX D

// normal end return count;

Function:

Name : F2dCleanUpFindOneSideOpenEntity2

Description: look for one side open and across with views' boundaries entities after found views Return Value:

= 0: Completed successfully and No found; = NUM: Open entities' number Parameters :

Input :

1 In_drawing: drawing __ int F2dCleanUpFindOneSideOpenEntity2 ( struct drawing ^♦In_drawing )

{

// external functions int F2dCleanUpChangeChildrenFlag() ; // link list and temp list struct node_list ^♦temp_list, ^♦ConnectedNode; // selected entity and temp entities struct node ^♦Entity; // define flag int BoundaryFlag = 0; int CountPrompt = 0; char message [64] ;

#ifdef VELLUM

CountPrompt++; sprintf ( message, "2 Start Find Oneside Open Entity %d", CountPrompt ) ; ck_prompt ( message ) ; #endif

// check node list if ( In_drawing == NULL ) return -1; if ( In_drawing->objects == NULL ) return -1; // check all of entities which connected with outside loop for ( Entity=In drawing->objects;Entity;Entity = Entity->next

⁾ {

// check connected nodes count if ( Entity == NULL ) break; if (( Entity->no_connected_l == 0 )

43 APPENDIX D

&&( Entity->no_connected_2 == 0 ) ) continue; // check flags if (( Entity->flags == 3 ) Sc&( Entity->type == CK_LINE ) && ( ( Entity->no_connected_l == 0 ) j j ( Entity->no_connected_2 == 0 ) ) ) { // one side open and line type if ( Entity->no_connected_2 == 0 )

ConnectedNode = Entity->connected_nodes_l; else ConnectedNode = Entity->connected_nodes_2; // clear found boundary's flag BoundaryFlag = 0;

// check across with views' boundary if ( ConnectedNode != NULL ) { for ( temp_list = ConnectedNode; temp_list; temp_list = temp_list->cdr ) { // check wrong data if ( temp_list->car == NULL ) break; if ( temp_list->car->flags == 4 ) { // find boundary's entity BoundaryFlag = 1; break;

} if ( temp_list->cdr == NULL ) break; if ( BoundaryFlag == 1 ) // look for children F2dCleanUpChangeChildrenFlag( Entity->parent

// check arc if (( Entity->type == CK_ARC ) &&( Entity->flags != 99 ) &&( ( Entity->no_connected_l == 0 ) I j ( Entity->no_connected_2 == 0 ) ) ) { // look for children F2dCleanUpChangeChildrenFlag( Entity->parent ) ; if ( Entity->next == NULL ) break;

}

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Finished 2 one side open Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif

// normal end return 0; }

44 APPENDIX D

Function:

Name : F2dCleanUpFindGroupMinX

Description: look for minimum X-value for a entities group Return Value :

= 0: Completed successfully and no entity; Parameters :

Input :

1 Entities_list : a node list that will be checked

Output

1 Entity: entity includes minimum X-value

_____ struct node ^♦F2dCleanUpFindGroupMinX( struct node_list ♦Entities list )

{

// double F2dCleanUpLineLength() ;

// entities counter and flags int open_entity_flag = 0;

// link list and temp list struct node_list

^♦EntitiesList2,

^♦Out_temp_list2,

^♦open_temp_list2, ^♦open_list2,

^♦temp_listl,

^♦temp_list2; // selected entity and temp entities struct node

^♦Next_entity,

^♦MinXvalueEntity; // length and angle double MinXvalueGroup = 0.0; // check all of entities which connected with outside loop // initialize EntitiesList2 =

( struct node_list ^♦ ) calloc ( 1, sizeof ( struct node_list ) ) ; open_list2 =

( struct node_list ^♦ ) calloc ( 1, sizeof ( struct node_list ) ) ;

// check first entity

// search connected entity until real entity be found for ( temp_list2 = Entities_list;

// entity->connected_nodes;

45 APPENDIX D

; ;

//

// Step_6_l:

// find connected entities for ( temp_listl = open_list2; ( temp_listl

// SS (temp_list->cdr != 0) > ; ) {

// get a pointer from open list if ( temp_listl->car == NULL ) break;

Next_entity = temp_listl->car;

// set a closed flag to the node

// Next_entity->flags = 1;

// close the node ( delete the node from open list ) open_list2 = open_list2->cdr;

// look for first connected entity whose flags=0. for ⁽ temp_list2 = Next_entity->connected_nodes; ( temp_list2 !=0 &_ temp_list2->car!=0 ) ; temp_list2=temp_list2->cdr) { if ( temp_list2->car == NULL ) break; // if found an unmarked connected entity if ( temp_list2

&_ temp_list2->car->f lags==2 &_ MinXvalueEn t i ty- > 1 ine__t ype temp_list2 - >car- >line_type

&_ MinXvalueEntity- > color

46 APPENDIX D temp_l ist2 - >car- >color

_ _ MinXva lueEnt i t y - > i s_t hi ckne s s temp_list2 - >car- >is_thickness

⁾ {

// check minimum X if ( temp list2->car->Min X < MinXvalueGroup

⁾ {

// change minimum X value

MinXvalueGroup = temp_list2->car->Min_X;

MinXvalueEntity = temp list2->car;

}

// allocate memory for open list element open_temp_list2 = ( struct node_list ^♦ ) calloc ( 1, sizeof ( struct node_list )) ;

// allocate memory for output node list element

0ut_temp_list2 = ( struct node_list ^♦ ) calloc ( 1, sizeof ( struct node_list ));

// add this entity to the open list open_temp__list2->car = temp_list2->car; // add this entity to the output node list Out_temp_list2->car = temp_list2->car;

// set a open flag to the node open_temp_list2->car->flags = 1;

// connect to the open list open_temp_list2- cdr = open_list2; // move the pointer of open list to the top open_list2 = open_temp_list2; // connect to the output node list Out_temp_list2->cdr = EntitiesList2; // move the pointer of output node list to the top

EntitiesList2 = Out temp_list2;

} if ( temp list2->cdr == NULL ) break;

}

// assign value to the loop variable temp listl = open list2;

} return MinXvalueEntity;

} /*

Function:

Name : F2dCleanUpFindOutNextEntity

47 APPENDIX D

Description: look for connected entity with out side loop entities

Return Value:

= 0: Completed successfully; Parameters :

Input :

1 In_entity: rotate axis entity

2 In_Nodes: entity's node

3 In_Vertex: entity's vertex

4 ^♦In_Angle: entity's angle

*/ int F2dCleanUpFindOutNextEntity(

// input struct node ^♦In_entity, struct node_list ^♦In_Nodes, struct planar_vertex ^♦In_Vertex, double ^♦In_Angle )

{

// external functions int F2dCleanUpArcAngle () ; int F2dCleanUpLineAngle () ,- // link list and temp list struct node_list

^♦temp_list,

^♦RotateNodes,

^♦RotateNodesTemp,

^♦RotateNodesAnother,

^♦RotateNodesAnotherTemp; // selected entity and temp entities struct node

^♦Next_entity,

^♦RotateEntity; struct planar_vertex

^♦RotateVertex,

^♦RotateVertexTemp,

^♦RotateVertexAnother,

^♦RotateVertexAnotherTemp; int RotateVertexFlag = 0,

RotateAxisEntityType = 0,

RotateAxisEntityVertexFlag = 0,

FindNextFlag = 0;

// length and angle double RotateAxisAngleAnother = 0.0,

RotateAxisAngleAnotherTemp - 0.0, RotateMinAngleTemp = 0.0, RotateDiffAngleTemp = 0.0, RotateAxisAngle = 0.0,

48 APPENDIX D

RotateAxisAngleTemp = 0.0; // set tolerance Tolerance double Tolerance = 0.0; Tolerance = F2dCleanUpTolerance;

// initialize

RotateAxisAngleAnother = ^♦In_Angle;

RotateEntity = In_entity;

RotateNodesAnother = In_Nodes;

RotateVertexAnother = In_Vertex;

// check all of entities which connected with outside loop for( ; ; ) (

// set flag for first entity on loop FindNextFlag = 0; RotateEntity->flags = 4;

RotateAxisEntityType = RotateEntity->type; if ( RotateEntity->vertexl == RotateVertexAnother ) RotateAxisEntityVertexFlag = 1; else RotateAxisEntityVertexFlag = 2; // set standard axis for first search RotateAxisAngle = RotateAxisAngleAnother; RotateMinAngleTemp = 2.0 ^♦ pi; RotateNodes = RotateNodesAnother;

RotateVertex = RotateVertexAnother; // check next connected entity if ( RotateNodes == 0 ) break; // start loop process for( temp_list=RotateNodes; temp_list; temp_list = temp_list->cdr) { if ( temp_list->car == NULL ) break; Next_entity = temp_list->car; // check flags if ( Next_entity->flags == 4 ) break; if ( Next_entity->flags != 1 ) continue; switch ( Next_entity->type ) { // arc type entity case CK_ARC:

RotateVertexFlag = F2dCleanUpArcAngle ( // input

RotateVertex, Next_entity, // output

- R o t a t e A x i s A n g l e T e m p , SLRotateAxisAngleAnotherTemp) ; break ; // line type entity case CK LINE :

RotateVertexFlag = F2dCleanUpLineAngle // input RotateVertex, Next_entity ,

49 APPENDIX D

// output

- R o t a t e A x i s A n g l e T e m p , -RotateAxisAngleAnotherTemp) ; break; // wrong type entity default : break;

}

// select connected side of entity if ( RotateVertexFlag == 1 ) {

// get start->end

// set rotation vertex node

RotateNodesTemp Next_entity->connected_nodes_l;

RotateVertexTemp = Next_entity->vertexl;

// set another side vertex node

RotateNodesAnotherTemp Next_entity->connected_nodes_2;

RotateVertexAnotherTemp Next_entity->vertex2; else {

// set rotation vertex node

RotateNodesTemp = Next_entity->connected_nodes_2;

RotateVertexTemp = Next_entity->vertex2;

// set another side vertex node

RotateNodesAnotherTemp Next_entity->connected_nodes_l;

RotateVertexAnotherTemp Next entity->vertexl;

}

// compute diff angle // for change current entity RotateDiffAngleTemp = RotateAxisAngleTemp RotateAxisAngle; if ( fabs ( RotateDiffAngleTemp ) < Tolerance ) { switch( RotateAxisEntityType ) { case CK_ARC: switch( RotateAxisEntityVertexFlag

⁾ { case 1 : RotateDiffAngleTemp = 2.0 ^♦ pi; break; case 2 : RotateDiffAngleTemp =

0.0; break;

} break; case CK_LINE:

50 APPENDIX D switch( RotateVertexFlag ) ( case 1: RotateDiffAngleTemp =

0 . 0 ; break; case 2 : RotateDiffAngleTemp =

2 . 0 PL break;

} break;

}

} if ( RotateDiffAngleTemp < 0.0 ) {

RotateDiffAngleTemp = RotateDiffAngleTemp + 2.0

* Pi

} if ( fabs ( RotateDiffAngleTemp - RotateMinAngleTemp

) _< Tolerance ) ( switch( Next_entity->type ) { case CK_ARC: switch( RotateEntity->type ) { case CK_ARC: switch( RotateVertexFlag

) ( case 1: // n o change break; case 2 :

RotateEntity

= Next_entity;

RotateNodes

= RotateNodesTemp;

RotateVertex

= RotateVertexTemp;

RotateNodesAnother = RotateNodesAnotherTemp;

RotateVertexAnother = RotateVertexAnotherTemp;

RotateAxisAngleAnother = RotateAxisAngleAnotherTemp; break;

} break; case CK_LINE: switch( RotateVertexFlag

⁾ { case 1: // n o change break; case 2:

RotateEntity

51 APPENDIX D Next_entity;

RotateNodes

= RotateNodesTemp; RotateVertex RotateVertexTemp;

RotateNodesAnother = RotateNodesAnotherTemp;

RotateVertexAnother = RotateVertexAnotherTemp;

RotateAxisAngleAnother = RotateAxisAngleAnotherTemp; break;

} break;

} case CK_LINE: switch( RotateEntity->type ) { case CK_ARC: if ( RotateEntity->vertexl

== RotateVertex ) {

RotateEntity

= Next_entity;

RotateNodes

= RotateNodesTemp;

RotateVertex = RotateVertexTemp;

RotateNodesAnother = RotateNodesAnotherTemp;

RotateVertexAnother = RotateVertexAnotherTemp;

RotateAxisAngleAnother = RotateAxisAngleAnotherTemp;

} break;

}

} if ( RotateDiffAngleTemp < RotateMinAngleTemp

FindNextFlag = 1;

RotateMinAngleTemp = RotateDiffAngleTemp;

// set rotation entity

RotateEntity = Next_entity;

RotateNodes RotateNodesTemp;

RotateVertex = RotateVertexTemp;

RotateNodesAnother RotateNodesAnotherTemp;

RotateVertexAnother RotateVertexAnotherTemp;

52 APPENDIX D

RotateAxisAngleAnother RotateAxisAngleAnotherTemp;

// check flags

// if loop meet closed entity if ( Next__entity->flags == 4 ) break; if ( FindNextFlag == 0 ) break;

} return 0;

Function:

Name : F2dCleanUpFindOutSideLoop

Description: look for out side loop entities Return Value:

= 0: Completed successfully and No found;

= NUM: Outside entities' number Parameters:

Input:

1 Entity: entity includes the loop's minimum X-value

*/ int F2dCleanUpFindOutSideLoop ( struct node ^♦MinXvalueEntity )

{

// external functions int F2dCleanUpArcAngle () ; int F2dCleanUpLineAngle () ; int F2dCleanUpFindOutFirstEntity() ; int F2dCleanUpFindOutNextEntity() ;

// entities counter and flags int count = 0, count2 = 0, open_entity_flag = 0,

RotateVertexFlag = 0, assistant_line_flag = 0; // link list and temp list struct node_list

^♦outside_loop_temp_list3 ,

^♦RotateNodes,

^♦RotateNodesTemp,

^♦RotateNodesAnother,

^♦RotateNodesAnotherTemp; // selected entity and temp entities

53 APPENDIX D struct node

♦Next_entity, ^♦RotateEntity; struct planar_vertex ^♦RotateVertex, ♦RotateVertexTemp, ♦RotateVertexAnother, ♦RotateVertexAnotherTemp; // length and angle double assistant_line_angle = 0.0, MinXvalueGroup = 0.0, RotateAxisAngleAnother = 0.0, RotateAxisAngleAnotherTemp = 0.0, RotateMinAngleTemp = 0.0, RotateDiffAngle = 0.0, RotateDiffAngleTemp = 0.0, RotateAxisAngle = 0.0, RotateAxisAngleTemp = 0.0; // set tolerance Tolerance double loop_tolerance = 0.00; loop_tolerance = F2dCleanUpTolerance;

// loop direction : counter-clock

// check a flag for outside loop process // if ( count2 > 0 ) { // step_7_l: find first entity for the group // set standard axis for first search RotateAxisAngle = 1.5 ^♦ pi; RotateMinAngleTemp = 2.0 ^♦ pi; // step_7_l_l : only one entity // check entities' angle

// arc case : special case ( no vertex on min_x ) if ( MinXvalueEntity->type == CK_ARC

&_ ( fabs ( MinXvalueEntity->Min_X MinXvalueEntity->vertexl->X ) > loop_tolerance )

_& ( fabs ( MinXvalueEntity->Min_X MinXvalueEntity->vertex2->X ) > loop_toleranee ) ) { // set rotation vertex node

RotateNodesAnother =MinXvalueEntity->connected_nodes_2; RotateVertexAnother = MinXvalueEntity->vertex2; RotateEntity = MinXvalueEntity; // get arc end-angle

// ( direction : from circle-center to end-point )

RotateAxisAngleAnotherTemp = ( MinXvalueEntity->AngEnd ) ^♦ pi / 180.0;

// change angle to oppsite direction

// ( direction : from end-point to circle-center

RotateAxisAngleAnotherTemp = RotateAxisAngleAnotherTemp

54 APPENDIX D

+ pi;

// check angle if ( RotateAxisAngleAnotherTemp > ( 2.0 ^♦ pi ) ) { R o t a t e Ax i s A n g l e An o t h e r T e mp = RotateAxisAngleAnotherTemp - 2.0 ^♦ pi; if ( fabs ( RotateAxisAngleAnotherTemp - ( 2.0 ♦ pi ) ) < loop_tolerance ) {

RotateAxisAngleAnotherTemp = 0.0;

}

// compute tangent line angle for the arc end-angle

// ( direction : from end-point to start-point )

RotateAxisAngleAnotherTemp = RotateAxisAngleAnotherTemp + 0.5 ^♦ pi;

// check angle if ( RotateAxisAngleAnotherTemp > ( 2.0 ^♦ pi ) ) {

R o t a t e Ax i s An g l e An o t h e r T e mp RotateAxisAngleAnotherTemp - 2.0 ^♦ pi;

} if ( fabs ( RotateAxisAngleAnotherTemp - ( 2.0 ^♦ pi ) ) < loop_tolerance ) {

RotateAxisAngleAnotherTemp = 0.0;

}

// set rotation standard angle ( clock direction )

RotateAxisAngleAnother = RotateAxisAngleAnotherTemp;

} // step_7_l_2 : multiple entities around Min_x else {

// select connected side of entity if ( fabs ( MinXvalueEntity- >Min_X

MinXvalueEntity->vertexl->X ) < loop_tolerance )

// get start->end

// set rotation vertex node

RotateVertex = MinXvalueEntity->vertexl; else RotateVertex = MinXvalueEntity->vertex2 ; switch ( MinXvalueEntity->type ) {

// arc type entity case CK_ARC:

RotateVertexFlag = F2dCleanUpArcAngle (

// input

RotateVertex, MinXvalueEntity,

// output

- R o t a t e A x i s A n g l e T e m p ,

-RotateAxisAngleAnotherTemp) ; break;

// line type entity case CK_LINE:

RotateVertexFlag = F2dCleanUpLineAngle (

// input

RotateVertex, MinXvalueEntity,

55 APPENDIX D

// output - R o t a t e A x i s A n g l e T e m p ,

-RotateAxisAngleAnotherTemp) ; break ; def ault : break ;

// select connected side of entity if ( RotateVertexFlag == 1 ) (

// get start->end

// set rotation vertex node

RotateNodesTemp MinXvalueEntity-_>connected_nodes_l;

RotateVertexTemp = MinXvalueEntity->vertexl;

// set another side vertex node

RotateNodesAnotherTemp MinXvalueEntity->connected_nodes_2;

RotateVertexAnotherTemp MinXvalueEntity->vertex2; else {

// set rotation vertex node

RotateNodesTemp MinXvalueEntity->connected_nodes_2;

RotateVertexTemp = MinXvalueEntity->vertex2;

// set another side vertex node

RotateNodesAnotherTemp MinXvalueEntity->connected_nodes_l;

RotateVertexAnotherTemp = MinXvalueEntity->vertexl;

}

// compute diff angle // for change current entity

RotateDiffAngleTemp = RotateAxisAngleTemp RotateAxisAngle; if ( RotateDiffAngleTemp < 0.0 ) {

RotateDiffAngleTemp = RotateDiffAngleTemp + 2.0 ♦

Pi,

} if ( RotateDiffAngleTemp < RotateMinAngleTemp ) RotateMinAngleTemp = RotateDiffAngleTemp; // set rotation entity

RotateEntity = MinXvalueEntity;

RotateNodes = RotateNodesTemp;

RotateVertex = RotateVertexTemp;

RotateNodesAnother = RotateNodesAnotherTemp;

RotateVertexAnother = RotateVertexAnotherTemp;

RotateAxisAngleAnother

RotateAxisAngleAnotherTemp;

56 APPENDIX D

// step_7_l_3 : find next connected entity with first entity around Min_x-vertex

// check all of entities which connected with first entity if ( RotateNodes != 0 ) { for ( outside_loop_temp_list3=RotateNodes; outside__loop_temp_list3 ; o u t s i d e _ l o o p _ t e m p _ l i s t 3 outside_loop_temp_list3->cdr) { if ( outside_loop_temp_list3->car == NULL ) break ;

Next_entity = outside_loop_temp__list3 ->car ; // check flags if ( Next_entity->f lags !=1 ) continue; switch ( Next_entity->type ) { // arc type entity case CK_ARC:

R o t a t e V e r t e x F l a g F2dCleanUpArcAngle (

// input

RotateVertex, Next_entity, // output

-RotateAxisAngleTemp , -RotateAxisAngleAnotherTemp) ; break; // line type entity case CK_LINE:

R o t a t e V e r t e x F l a g F2dCleanUpLineAngle (

// input

RotateVertex, Next_entity, // output

-RotateAxisAngleTemp , -RotateAxisAngleAnotherTemp) ; break; // wrong type entity default : break;

}

// select connected side of entity if ( RotateVertexFlag == 1 ) { // get start->end // set rotation vertex node RotateNodesTemp = Next_entity->connected_nodes_l;

RotateVertexTemp = Next_entity->vertexl;

// set another side vertex node RotateNodesAnotherTemp Next_entity->connected_nodes_2;

57 APPENDIX D RotateVertexAnotherTemp

Next entity- >vertex2 ;

} else {

// set rotation vertex node

RotateNodesTemp Next_entity->connected_nodes_2;

RotateVertexTemp Next_entity->vertex2;

// set another side vertex node

RotateNodesAnotherTemp Next_entity->connected_nodes_l;

RotateVertexAnotherTemp Next entity->vertexl;

}

// compute diff angle

// for change current entity

RotateDiffAngleTemp = RotateAxisAngleTemp

RotateAxisAngle; if ( RotateDiffAngleTemp < 0.0 ) {

RotateDiffAngleTemp = RotateDiffAngleTemp + 2.0 ^♦ pi;

} if ( f abs ( RotateDif fAngleTemp

RotateMinAngleTemp ) < loop_tolerance ) { switch { Next_entity- >type ) { case CK_ARC : switch ( RotateEntity- >type ) case CK_ARC : s w i t c h RotateVertexFlag ) { case 1 : // change break ; case 2 :

RotateEntity = Next_entity;

RotateNodes = RotateNodesTemp;

RotateVertex = RotateVertexTemp;

RotateNodesAnother = RotateNodesAnotherTemp;

RotateVertexAnother = RotateVertexAnotherTemp;

RotateAxisAngleAnother = RotateAxisAngleAnotherTemp; break;

}

58 APPENDIX D break; case CK_LINE: s w i t c h (

RotateVertexFlag ) { case 1: // D change break; case 2 :

RotateEntity - Next_entity;

RotateNodes = RotateNodesTemp;

RotateVertex = RotateVertexTemp;

RotateNodesAnother = RotateNodesAnotherTemp;

RotateVertexAnother = RotateVertexAnotherTemp;

RotateAxisAngleAnother - RotateAxisAngleAnotherTemp; break;

} break;

} case CK_LINE: switch( RotateEntity->type ) { case CK_ARC: i f (

RotateEntity->vertexl === RotateVertex ) {

RotateEntity = Next_entity;

RotateNodes = RotateNodesTemp;

RotateVertex = RotateVertexTemp;

RotateNodesAnother = RotateNodesAnotherTemp;

RotateVertexAnother = RotateVertexAnotherTemp;

RotateAxisAngleAnother = RotateAxisAngleAnotherTemp;

} break;

}

} if ( RotateDiffAngleTemp < RotateMinAngleTemp

59 APPENDIX D

RotateMinAngleTemp = RotateDiffAngleTemp;

// set rotation entity

RotateEntity = Next_entity;

RotateNodes

RotateNodesTemp;

RotateVertex

RotateVertexTemp;

RotateNodesAnother

RotateNodesAnotherTemp;

RotateVertexAnother

RotateVertexAnotherTemp;

RotateAxisAngleAnother RotateAxisAngleAnotherTemp;

// step_7_2 : find next connected entity with first entity along the loop

F2dCleanUpFindOutNextEntity( // input

RotateEntity, RotateNodesAnother, RotateVertexAnother, -RotateAxisAngleAnother ) ; return 0;

}

Function:

Name : TestFindArrowControl

Description:

It is a manual function.

To look for all arrow type entities for the view.

Return Value :

< 0 : Error = 0 : No proccess

> 0: Completed successfully; the Number is the counter of found entities. Parameters: Subfunctions : struct node_list ^♦F2dCleanUpPickUpAllOneSideOpenEntity look for one side open entity int F2dCleanUpFindArrow look for arrow type entity

60 APPENDIX D int F2dCleanUpFindArrowEndEntity find arrow end line int F2dCleanUpFindArrowCenterEntity find arrow center line int F2dCleanUpFindArrowOneSideOpenEntity find one side open entity for arrow type int F2dCleanUpComputeCenterLineAngle compute center line angle int F2dCleanUpCheckStartEndEntity check start and end entities int F2dCleanUpLineAngle compute a line type entity's angle int F2dCleanUpArcAngle compute a arc type entity's angle, int F2dCleanUpChangeChildrenFlag change entity's children flags int F2dCleanUpChangeArrowVertexFlag change arrow top vertex's flags and set arrow center line coordinates int F2dCleanUpChangeDrawingColor change entity's color double F2dCleanUpLineLength compute line type entity's length

int TestFinciArrowControl ( int ♦SwitchRedraw )

{

// external functions int F2dCleanUpFindArrow() ; int F2dCleanUpChangeDrawingColor() ; struct node_list ^♦F2dCleanUpPickUpAll0neSide0penEntity() ;

// entities counter and flags int count = 0, assistant_line_flag = 0; // link list and temp list struct node__list

^♦Entities_change_color_listl;

// Begin

//

// Step_l:

// check the number of entites // get count of all entities // if an error occurred, exit immidiately. if (number_of_entities<0) return -1;

//

// Step_3 : pickup all one side open entities

E n t i t i e s _ c h a n g e _ c o l o r _ l i s t l

61 APPENDIX D

F2dCleanUpPickUpAllOneSideOpenEntity( &trim_drawing ) ; if (Entities_change_color_listl == 0) return 0; /

// Step_4 : find arrow line

F2dCleanUpFindArrow( Entities_change_color_listl ) ;

// Step_X:

// change color to Magenta for a selected boundary

// turn off input levels #ifdef CKWIN if ( ♦SwitchRedraw == 1 ) ck_levels (CK_OFF, 1,255) ;

#endif if ( ♦SwitchRedraw == 1 ) F2dCleanUpChangeDrawingColor(_trim_drawing) ;

// clear a flag for outside loop process assistant_line_flag = 0;

// turn ON the level used to store the trimmed entities. #ifdef CKWIN if ( ♦SwitchRedraw == 1 ) ck_l eve l s ( CK_ON , Tr imLeve l ,

TrimLevel ) ; if ( ♦SwitchRedraw == 1 ) ck_redraw( CK_PRIME_VP ) ; #endif return count;

} /*

Function:

Name : TestFindOneSideOpen2Control

Description:

It is a manual function.

To find all one side open and across with views' boundaries entities Return Value :

< 0 : Error = 0 : No proccess Parameters: Subfunctions: int F2dCleanUpFindOneSideOpenEntity2 look for all one side open and across with views' boundaries entities int F2dCleanUpChangeDrawingColor change entity' s color -- =================_______

int TestFindOneSideOpen2Control ( int ♦SwitchRedraw )

62 APPENDIX D

{

// external functions int F2dCleanUpFindOneSideOpenEntity2 () ; int F2dCleanUpChangeDrawingColor () ;

// Step_l : check the number of entites // if an error occurred, exit immidiately. if (number_of_entities<0) return -1;

// Step_2: pickup all one side open entities

F2dCleanUpFindOneSideOpenEntity2 ( _trim_drawing ) ;

// Step_X: change color to Magenta for a selected boundary

// turn off input levels #ifdef CKWIN if ( ♦SwitchRedraw == 1 ) ck__levels (CK_0FF, 1, 255) ; #endif if ( ♦SwitchRedraw == 1 ) F2dCleanUpChangeDrawingColor(_trim_drawing) ;

// turn ON the level used to store the trimmed entities. #ifdef CKWIN if ( ♦SwitchRedraw == 1 ) ck_levels (CK_ON, TrimLeve1 , TrimLevel) ; if ( ♦SwitchRedraw == 1 ) ck_redraw( CK_PRIME_VP ) ; #endif return 0;

}

Function:

Name : TestFindOneSideOpenControl

Description:

It is a manual function.

To find all one side open and vertical with arrow type entities for the view.

Return Value:

< 0 : Error = 0: No proccess

> 0: Completed successfully; the Number is the counter of found entities. Parameters: Subfunctions : struct node_list ^♦F2dCleanUpPickUpA110neSideOpenEntity look for one side open entity int F2dCleanUpChangeChildrenFlag change entity' s children flags

63 APPENDIX D int F2dCleanUpChangeDrawingColor change entity' s color int F2dCleanUpFindOneSideOpenEntity look for one side open entities and they are vertical with arrow type entities

* / int TestFindOneSideOpenControl ( int ^♦SwitchRedraw )

{

// external functions int F2dCleanUpFindOneSideOpenEntity() ; int F2dCleanUpChangeDrawingColor() ; struct node_list ♦F2dCleanUpPickUpAllOneSideOpenEntity() ;

^♦Entities_list;

// Begin

//

// Step_l :

// check the number of entites // if an error occurred, exit immidiately. if (number_of_entities<0) return -1;

//

// Step_2 : pickup all one side open entities

Entities_list = F2dCleanUpPickUpAllOneSideOpenEntity (

_trim_drawing ) ; if (Entities_list == 0) return 0; //

// Step_3 : find lines vertical with arrow type entities

F2dCleanUpFindOneSideOpenEntity( Entities list ) ; //

// Step_X:

// change color to Magenta for a selected boundary

// turn off input levels #ifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck levels (CK OFF, 1,255) ; #endif if ( ^♦SwitchRedraw == l ) F2dCleanUpChangeDrawingColor(_trim_j_rawing) ;

// clear a flag for outside loop process assistant_line_flag = 0;

// turn ON the level used to store the trimmed entities. #ifdef CKWIN if ^{( ♦}SwitchRedraw == 1 ) ck_levels (CK_ON, TrimLevel ,

64 APPENDIX D

TrimLevel) ; if ( ^♦SwitchRedraw == 1 ) ck_redraw( CK_PRIME_VP ) ; #endif return count;

} /*

Function:

Name : F2dCleanUpPickUpAllOneSideOpenEntity

Description: look for all one side open entities in selected drawing Return Value:

= 0: Completed successfully; Parameters :

Input :

In_drawing: drawing Output:

NodeList: a node list for all one side open entities __ struct node_list ♦F2dCleanUpPickUpAllOneSideOpenEntity(struct drawing ^♦In_drawing )

{

// entities counter and flags int count = 0;

// link list and temp list struct node_list

^♦EntitiesListl,

♦Out_temp_list; // selected entity and temp entities struct node ^♦entity;

// initialize

EntitiesListl = ( struct node list ^♦ ) calloc ( 1, sizeof ( struct node list

) )

// set a flag for the boundary entity count = 0;

// check pointer about first entity if ( In_drawing->objects == 0 ) return 0; // set a open flag to the node // open_list->car->flags = 1; // set the pointer to output node list // set pointer to first entity for (entity=In_drawing->objects; entity __ (count<number_trimmed_entities) ;

65 APPENDIX D entity=entity->next) { if ( entity == NULL ) break; if ( entity->no_connected_l > 0 &_ entity->no_connected_2

_> 0 ) continue; i f ( ent i t y - > no_c onne c t e d_l = = 0 Sc ¬ ent ity- >no_connected_2 == 0 ) continue ; if ( entity->type != CK_LINE ) continue; EntitiesListl->car = entity; break;

}

// set link address

EntitiesListl->cdr = NULL; /

// find connected entities if ( EntitiesListl == 0 ) return 0; if ( entity == 0 ) return 0; if ( entity->next == 0 ) return EntitiesListl; for (entity-entity->next; entity &_ (count<number_trimmed_entities) ; entity=entity->next) { count++; if ( entity == NULL ) break; if ( entity->type != CK_LINE ) continue; if ( entity->no_connected__l > 0 _SL entity->no_connected_2

> 0 ) continue; i f ( ent i ty - >no_conne c t e d_l = = 0 _ _ entity- >no_connected_2 == 0 ) continue ;

// allocate memory for output node list element

Out_temp_list = ( struct node_list ^♦ ) calloc ( 1, sizeof ( struct node_list ) ) ;

// add this entity to the output node list

Out_temp_list->car = entity;

// connect to the output node list

Out_temp_list->cdr = EntitiesListl;

// move the pointer of output node list to the top

EntitiesListl = Out_temp_list; if ( entity->next == NULL ) break;

} if (co_nt>=number_trimmed_entities) return 0; else return EntitiesListl;

Function:

Name : F2dCleanUpPickUpAllArcEntity

Description: look for all arc entities in selected drawing

66 APPENDIX D

Return Value:

= 0: Completed successfully; Parameters :

Input :

In_drawing: drawing Output:

NodeList : a node list for all arc entities

// link list and temp list struct node_list

♦EntitiesListl,

♦Out_temp_list; // selected entity and temp entities struct node ^♦entity;

// initialize

EntitiesListl = ( struct node_list ^♦ ) calloc ( 1, sizeof ( struct node_list

> ) ;

// set a flag for the boundary entity count = 0;

// check pointer about first entity if ( In_drawing->objects == 0 ) return 0; // set pointer to first entity for (entity=In_drawing->objects; entity __ (count<number_trimmed_entities) ; entity=entity->next) { if ( entity == NULL ) break; if ( entity->type != CK_ARC ) continue; EntitiesListl->car = entity; break;

}

// set link address

EntitiesListl->cdr = NULL; //

// find connected entities if ( EntitiesListl == 0 ) return 0; if ( entity == 0 ) return 0; if ( entity->next == 0 ) return EntitiesListl; for (entity-entity->next; entity _S_ (count<number_trimmed_entities) ; entity=entity->next) { count++; if ( entity == NULL ) break;

67 APPENDIX D if ( entity->type != CK_ARC ) continue;

// allocate memory for output node list element Out_temp_list = ( struct node_list ^♦ ) calloc ( 1, sizeof ( struct node_list ) ) ;

// add this entity to the output node list

Out_temp_list->car = entity;

// connect to the output node list

Out_temp_list->cdr = EntitiesListl;

// move the pointer of output node list to the top

EntitiesListl = Out_temp_list; if ( entity->next == NULL ) break;

} if (count>=number_trimmed_entities) return 0; else return EntitiesListl;

Function:

Name : F2dCleanUpFindCenterLineEntity

Description: look for arc's center line entities. Return Value:

= 0: Completed successfully and No found; = NUM: Open entities' number Subfunctions : int F2dCleanUpChangeChildrenFlag set flags for unuse entities double F2dDistancePointLine compute distance from arc's center point to entity Parameters:

Input:

1 In_Drawing: a drawings

2 In_Entities_list: a node list for all arcs __ int F2dCleanUpFindCenterLineEntity( struct drawing ♦In_drawing, struct node list ^In Entities list )

{ ^"

// external functions int F2dCleanUpChangeChildrenFlag() ; double F2dDistancePointLine () ; double F2dCleanUpLineLength() ; // link list and temp list struct node_list ^♦temp_list, ^♦temp_list2, ♦temp_list3, ♦temp_list ; // selected entity and temp entities

68 APPENDIX D struct node

^♦ArcEntity, ♦ViewEntity; // define 3D coordinates struct Vector Vsp, Vep, Vp; // entities counter and flags int count = 0; // double double Distance = 0.0,

DistanceViewEntity2Center = 0.0, DX = 0.0, DY = 0.0,

ViewEntityLength = 0.0, Tolerance = 0.0; int CountPrompt = 0 ; char message [64] ;

// set tolerance

Tolerance = F2dCleanUpTolerance;

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Start Find Center Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif

// check all of entities which connected with outside loop for( temp_list=In_Entities_list; temp_list; temp_list = temp_list->cdr) { if ( temp_list->car == NULL ) break; ArcEntity = temp_list->car; // check all of entities in the view for( ViewEntity=In_drawing->objects; ViewEntity;

ViewEntity = ViewEntity->next) { // check flags : only no marks entities if ( ViewEntity->flags !=0

__ ViewEntity->flags !=10 ) continue; // check type : only line if ( ViewEntity->type != CK_LINE ) continue; // check open side : only open line if ( ViewEntity->no_connected_l > 0

_Sc ViewEntity->no_connected 2 > 0 ) continue,

// check open side : two side open line ( delete ) if ( ViewEntity->no_connected_l == 0

__ ViewEntity->no_connected_2 == 0

69 APPENDIX D

Sc- ViewEntity->line__type != CK_CENTER ) { ViewEntity->flags = 3; continue;

// check parent open side if ViewEntity->no_connected_l == 0

_6c ViewEntity->line_type != CK_CENTER ) {

// set center line flags ( look for children for ( temp_list2=ViewEntity->parent->children; temp_list2 __ temp_list2->cdr; temp_list2 = temp_list2->cdr) { // change flag to light gray if ( temp_list2->car == NULL ) break; if ( temp_list2->car->vertex2

ViewEntity->parent->vertex2 ) break; if ( temp_list2-_>car-_>no_connected_2 != 0) continue;

} if ( ViewEntity->no_connected_2 == 0

_Sc ViewEntity->line_type != CK_CENTER ) {

// set center line flags ( look for children for( temp_list3=ViewEntity->parent->children; temp_list3 _Sc temp_list3->cdr; temp_list3 = temp_list3->cdr) { // change flag to light gray if ( temp_list3->car == NULL ) break; if ( temp_list3->car->vertexl

== ViewEntity->parent->vertexl ) break;

} if ( temp_list3->car->no_connected_l != 0) continue;

} // check distance from arc's center to line entity

Vsp.x=ViewEntity->X1;Vsp.y=ViewEntity->Y1;Vsp. z=ViewEntity->Z1;

70 APPENDIX D

Sc& ( ViewEntity->parent- >Max_Y > =

ArcEntity- >Min_Y

__ ViewEntity- >parent - >Max_Y < =

ArcEntity- >Max_Y )) {

F 2 dC l e anUpChange Ch i l dre nF l ag ( ViewEntity- >parent ) ; continue;

} if (( ViewEntity- >parent - >Min_X >= ArcEntity- >Min_X

ScSc ViewEntity- >parent - >Min_X < =

ArcEntity- >Max_X )

_δc ( ViewEntity- >parent - >Min_Y > =

ArcEntity- >Min_Y

_δc ViewEntity- >parent- >Min_Y < =

ArcEntity- >Max_Y )) {

F 2 dC l e anUp Change Ch i l dren F l ag ( ViewEntity- >parent ) ; continue;

} if (( ArcEntity->Max_X >= ViewEntity->parent->Min_X

Sc _ A r c E n t i t y - > M a x _ X < =

ViewEntity->parent->Max_X )

SC SL ( A r c E n t i t y - > M a x _ Y > =

ViewEntity- >parent - >Min_Y

Sc _ Ar c En t i t y - > Ma x_Y < =

ViewEntity- >parent->Max_Y ) ) {

F 2 dC l e anUpChange Ch i l dre nF l ag ( ViewEntity- >parent ) ; continue;

} if (( ArcEntity- >Min_X >= ViewEntity- >parent->Min_X

_ 5c A r c E n t i t y - > M i n _ X < =

ViewEntity- >par ent ->Max_X )

_ SL ( A r c E n t i t y - > M i n_Y > =

ViewEntity- >parent - >Min_Y

Sc _ ArcEnt i ty - >M in_Y < =

ViewEntity- >parent - >Max_Y ) ) {

F 2 dCle anUpChangeChi ldrenF l ag (

ViewEntity- >parent ) ; continue;

}

// check distance from arc to entity

V i e w E n t i t y L e n g t h F2dCleanUpLineLength (ViewEntity- >parent) ;

D X =

0.5* (ViewEntity- >parent->Min_X+ViewEntity->parent->Max_X)

- ArcEntity->CenterX;

D Y

0.5* (ViewEntity->parent->Min_Y+ViewEntity->parent->Max_Y)

- ArcEntity->CenterY;

71 APPENDIX D

DistanceViewEntity2Center = sqrt (DX+DX+DY^DY)

- ArcEntity->Radius - 0.5♦ViewEntityLength; if ( DistanceViewEntity2Center <= ViewEntityLength

F 2 dC l e anUp Change Chi l drenF l ag ( ViewEntity- >parent ) ; continue;

// set center line flags ( look for children ) for ( temp_list4=ViewEntity->parent->children; temp_list4; temp_list4 = temp_list4->cdr) ( // change flag to light gray if ( temp_list4->car == NULL ) break; temp_list4->car->flags = 10; count++;

} if ( ViewEntity->next == NULL ) break;

} }

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Finished find center Entity %d", CountPrompt ) ; ck_prompt ( message ) ; #endif

// normal end return count;

} /*

Function:

Name : F2dCleanUpFindInSideEntity

Description: look for in side entities for outside loop. Return Value :

= 0: Completed successfully and No found;

= NUM: Open entities' number Subfunctions: int F2dCleanUpChangeChildrenFlag set flags for unuse entities Parameters:

Input :

1 In_Drawing: a drawings

2 In_Entities_list : a node list for outside loop

72 APPENDIX D

*/ int F2dCleanUpFindInSideEntity( struct drawing ^♦In_drawing, struct node_list ^♦In_Entities_list )

{

// external functions int F2dCleanUpChangeChildrenFlag() ;

// link list and temp list struct node_list

♦temp_list, ^♦temp_list2,

♦NewNodeList, ^♦NewNodeListTemp; struct group

^♦NewGroup,

♦temp_group_list; struct bounding_box ^♦NewBoundBox; // selected entity and temp entities struct node

^♦OutsideEntity,

^♦ViewEntity; // entities counter and flags int count = 0,

NoFlag = 0; // double double Tolerance = 0.0;

// set tolerance

Tolerance = F2dCleanUpTolerance;

// initialize

NewGroup = (struct group ^♦) calloc (1, sizeof (struct group )) ; NewBoundBox = (struct bounding_box ^♦) calloc (1, sizeof (struct bounding_box ) ) ;

NewNodeList = (struct node_list ^♦)calloc (1, sizeof (struct node_list ) ) ;

// set group data NewGroup->next = NULL; NewGroup->box = NewBoundBox;

// check all of entities which connected with outside loop for ( temp_list=In_Entities_list; temp_list; // __ temp_list->cdr; temp_list = temp__list->cdr) { if ( temp__list->car == NULL ) break; OutsideEntity = temp_list->car; // check entity : only boundary group ( flags = 1 and

4 ) if ( OutsideEntity->flags != 1

_Sc OutsideEntity->flags != 4 ) continue; if ( !OutsideEntity ) {

NoFlag = 1; break;

}

73 APPENDIX D

NewGroup-_>box->Min_X = OutsideEntity->Min_X

NewGroup->box->Max_X = OutsideEntity->Max_X

NewGroup->box->Min_Y = OutsideEntity->Min_Y

NewGroup->box->Max_Y = OutsideEntity->Max_Y break;

} if ( NoFlag == 1 ) return 0;

NewNodeList->car = OutsideEntity;

NewNodeList->cdr = NULL;

// check first group if ( In_drawing->views == NULL ) {

NewGroup->index = 1;

In drawing->views = NewGroup;

} else {

// check group number for( temp_group_list= In drawing->views; temp_group_list; 7/ && temp_group_list->next; temp_group_list = temp_group_list->next) { if ( temp_group_list->next == NULL ) {

NewGroup->index = (temp_group_list->index) +1; temp_gro_p_list->next = NewGroup; break;

}

// check connected entity if ( !temp_list->cdr __ temp_list->car->type != CK_ARC ) return 0;

// check all of entities which connected with outside loop if ( temp_list->cdr ) { for ( temp_list2 = temp_list->cdr; temp_list2; // __ temp_list2->cdr; temp_list2 = temp_list2->cdr) { if ( temp_list2->car == NULL ) break; OutsideEntity = temp_list2->car; // check entity : only boundary group ( flags = 1 and 4 ) if ( OutsideEntity->flags != 1

_Sc OutsideEntity->flags != 4 ) continue if ( NewGroup->box->Min_X > OutsideEntity->Min_X )

NewGroup->box->Min_X = OutsideEntity->Min_X if ⁽ NewGroup->box->Max_X < OutsideEntity->Max_X

NewGroup->box->Max_X = OutsideEntity-_>Max_X if ⁽ NewGroup->box->Min_Y > OutsideEntity->Min_Y )

NewGroup->box->Min_Y = OutsideEntity->Min^~Y if ⁽ NewGroup->box->Max_Y < OutsideEntity->Max^~Y )

NewGroup->box->Max_Y = OutsideEntity-_>Max_Y // allocate memory for node list NewNodeListTemp = (struct node_list

74 APPENDIX D

♦) calloc (1, sizeof (struct node_list )) ;

// add this entity to node list NewNodeListTemp->car = OutsideEntity; // connect to the node list NewNodeListTemp->cdr - NewNodeList; // move the pointer of the node list to the top NewNodeList = NewNodeListTemp; count++; }

}

// check all of entities in the view for ( ViewEntity=In_drawing->objects; ViewEntity;

ViewEntity = ViewEntity->next) { // check flags : only no marks entities if ( ViewEntity == NULL ) break; if ( ViewEntity->flags !=0 ) continue; // check bounding box: only inside entities if ( ( NewGroup->box->Min_X <= ViewEntity->Min_X ) __( NewGroup->box->Max_X >= ViewEntity->Max_X ) Sc_( NewGroup->box->Min_Y <= ViewEntity->Min_Y ) __( NewGroup->box->Max_Y >= ViewEntity->Max_Y )) // set flag ViewEntity->flags = 7; // allocate memory for node list

NewNodeListTemp = (struct node_list ♦) calloc (1, sizeof (struct node_list )) ;

// add this entity to node list NewNodeListTemp->car = ViewEntity; // connect to the node list NewNodeListTemp->cdr = NewNodeList; // move the pointer of the node list to the top NewNodeList = NewNodeListTemp; count++,- }

}

// set node list to group

NewGroup->entities = NewNodeList;

// normal end return count;

Function:

Name : F2dCleanUpFindOnesideOpenConnectEntityNext

Description: look for next entity connected with one side open entity

75 APPENDIX D

Return Value:

= Out_entity: Completed successfully; Parameters :

Input :

1 In_entity: entity

s *truct node ♦F2dCleanUpFindOnesideOpenConnectEntityNext (struct node

♦In entity)

{

// selected entity and temp entities struct node ♦Out_entity;

// link list and temp list struct node_list

^♦TempListl, ^♦TempList2, ^♦TempList3, ^♦NextNodes;

// define int Flagl = 0,

Flag2 = 0,

Count = 0; int CountPrompt = 0; char message [64] ;

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Start Find One Side Open Connected Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif

// initialize Out_entity = NULL;

// check input entity if ( !In_entity ) return Out_entity;

// check open side for( TempListl=In_entity->connected_nodes_l;

TempListl; // ScSc TempListl->cdr;

TempListl = TempListl->cdr) { if ( TempListl->car == NULL ) break; if ⁽ TempListl->car->flags == 6 ) Flagl = 1; for ( TempList2=In_entity->connected_nodes_2 ; TempList2; // __ TempList2->cdr; TempList2 = TempList2->cdr) { if ( TempList2->car == NULL ) break; if ⁽ TempList2->car->flags == 6 ) Flag2 = 1;

// first entity case if ( Flagl == 0 __ Flag2 == 0 ) {

76 APPENDIX D if ( In_entity->no_connected_l == 0 ) Flagl = 1; if ( In entity->no_connected_2 == 0 ) Flag2 = 1;

}

// f inish case : return if ( ( Flagl == 0 __ Flag2 == 0 ) j I ( Flagl == 1 &&. Flag2 == 1 ) ) return Out_entity;

// check all of entities connected with input entity if ( Flagl == 1 ) NextNodes = In_ent ity- >connected_nodes_2 ; else N e x t N o d e s

In_entity- >connected_nodes_l ; for ( TempList3= NextNodes;

TempList3; // SL TempList3->cdr; TempList3 = TempList3->cdr) { if ( TempList3->car == NULL ) break; if ( TempList3->car->flags == 0 ) { Out_entity = TempList3->car; Count++; }

}

// check result if ( Count != 1 ) Out_entity = NULL;

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Finished one side open connected Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif return Out entity;

} /*

Function:

Name : F2dCleanUpFindOnesideOpenConnectEntity

Description: look for one side open entity and single connected with it Return Value:

= 0: Completed successfully; Subf nctions: struct node ^♦F2dCleanUpFind0neside0penConnectEntityNext look for next entity connected with one side open entity Parameters :

Input :

1 In_drawing: drawing

77 APPENDIX D int F2dCleanUpFindOnesideOpenConnectEntity(struct drawing ♦In drawing)

{

// external functions struct node ♦F2dCleanUpFindOnesideOpenConnectEntityNext () ;

// selected entity and temp entities struct node ♦entity, ^♦FindEntity, ^♦NextEntity;

// check all of entities on the list for( entity=In_drawing->objects; entity; entity = entity->next) { if ( entity == NULL ) break; if ( entity->flags != 0 ) continue; if ( entity->type != CK_LINE ) continue; if ( entity->no_connected_l > 0

__ entity->no_connected_2 > 0 ) continue; if ( entity->no_connected_l == 0

__ entity- >no_connected_2 == 0 ) continue; if ( ( en t i t y - >no_c onne c t ed_l == 0 &.& entity- >no_connected_2 >= 1)

I I (entity- >no_connected_2 == 0 __ entity- >no_connected_l >= 1) ) {

// set flag entity->flags = 6; FindEntity = entity; // find connected entity for (;;) {

NextEntity =

F2dCleanUpFindOnesideOpenConnectEntityNext ( FindEntity ) ; if ( NextEntity == NULL ) break; // set flag FindEntity = NextEntity; FindEntity->flags = 6;

return 0;

Function:

Name : F2dCleanUpFindLongestEntity

Description: look for a longest entity for a drawing Return Value :

78 APPENDIX D

- Entity: Completed successfully; 0 : No entity; Parameters:

Input:

1 In_drawing: drawing

s *7tr ^"uct node ^♦F2dCleanUpFindLongestEntity(struct drawing ♦In drawing)

{ ^"

// selected entity and temp entities struct node ^♦entity, ^♦LongestEntity;

// define double LongestLength = 0.0,

ArcLength = 0.0;

// check all of entities on the list for( entity=In_drawing->objects; entity; entity = entity->next) ( // only not used if ( entity == NULL ) break; if ( entity->flags != 0 ) continue; // check length if ( entity->parent->length <= LongestLength ) continue;

// check arc length if ( entity->type == CK_ARC ) {

ArcLength = 2.0 ^♦ entity->parent->Radius; if ( entity->parent->length < ArcLength )

ArcLength = entity->parent->length; if ( ArcLength > LongestLength ) { // change max value and entity LongestLength = ArcLength; LongestEntity - entity;

else {

// change max value and entity LongestLength = entity->parent->length; LongestEntity = entity;

if ( LongestLength == 0.0 ) return NULL; else return LongestEntity;

/*

79 APPENDIX D

Function:

Name : F2dCleanUpFindViews

Description: look for view's groups for a drawing Return Value:

Entity: Completed successfully; 0 : No entity; Subfunctions : struct node ♦F2dCleanUpFindLongestEntity look for a longest entity for a drawing struct node_list ^♦F2dCleanUpFindGroup look for a group entities connected with input entity struct node ♦F2dCleanUpFindGroupMinX look for minimum X-value for a entities group int F2dCleanUpFindOutSideLoop look for out side loop entities int F2dCleanUpFindInSideEntity look for in side entities for outside loop Parameters :

Input:

1 In_drawing: drawing

*/^~~ int F2dCleanUpFindViews ( struct drawing ⁺In drawing )

{

// external functions struct node ^♦F2dCleanUpFindLongestEntity() ; struct node_list ^♦F2dCleanUpFindGroup() ; struct node ^♦F2dCleanUpFindGroupMinX() ; int F2dCleanUpFindOutSideLoop () ; int F2dCleanUpFindInSideEntity() ;

// selected entity and temp entities struct node ^♦LongestEntity, ^♦MinXvalueEntity; struct node_list ^♦ViewGroup;

// define int ReturnFlag = 1,

NoGroup = 0,

Count = 0; double LongestLength = 0.0; int CountPrompt = 0; char message [64] ;

// check all views while ( ReturnFlag ) { #ifdef VELLUM

CountPrompt++; sprintf ⁽ message, "Start Find Views Entity %d" , CountPrompt ) ; ck_prompt ( message ) ;

80 APPENDIX D

#endif

// Step_l: find a longest entity for a drawing LongestEntity = F2dCleanUpFindLongestEntity( In_drawing ) ; if ( LongestEntity == NULL ) ReturnFlag = 0; else {

// set flag

NoGroup = 1;

// clear counter

Count = 0;

// Step_2 : find connected group with the longest entity

ViewGroup = F2dCleanUpFindGroup ( LongestEntity, -Count )

// Step_3 : find outside loop entities

// check a flag for outside loop process if ( Count > 1 ) M i n Xv a l u e E n t i t y

F2dCleanUpFindGroupMinX( ViewGroup ) ; else MinXvalueEntity = LongestEntity; if ( MinXvalueEntity == 0 ) continue;

// Step_4 : find outside loop entities if ( Count > 1 ) F2dCleanUpFind0utSideLoop (

MinXvalueEntity ) ; else MinXvalueEntity->flags = 4; // Step_5 : find inside entities F2dCleanUpFindInSideEntity( In drawing, ViewGroup ) ;

} #ifdef VELLUM

CountPrompt++; sprintf ( message, "Finished find view Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif

} if ( NoGroup == 0 ) return -1; else return 0;

}

/*

Function:

Name : F2dCleanUpFindRelativeViews

Description: look for relative views. Return Value:

= 0: Completed successfully;

= 1 : Found relative Views Parameters:

Input :

1 In_Drawing: a drawing

81

APPENDIX D

2 In_Group: a group

3 In ToleranceViews : tolerance for views

*/ int F2dCleanUpFindRelativeViews ( struct drawing ^♦In_drawing, struct group ^♦In_Group, double ♦In_ToleranceViews )

{

// link list and temp list struct group ^♦TempGroup; struct group_list ^♦Te pGroupList; // entities counter and flags int ReturnFlag = 0; // double double Tolerance = 0.0;

// set tolerance

Tolerance = ^♦In_ToleranceViews;

// check views if ( In_drawing->views == NULL ) return 0; if( In_Group == NULL ) return 0; for ( TempGroup= In_drawing->views;

TempGroup; // _& TempGroup->next; TempGroup = TempGroup->next) { // check same group if ( TempGroup == NULL ) break; if ( TempGroup == In_Group ) continue; // Step_l: check top and bottom direction i f

( (TempGroup->box->Max_X>= ( (In_Group->box->Max_X) -Tolerance ) )

SE_(TempGroup->box->Max_X<= ( (In_Group->box->Max_X) +Tolerance ) )

__ (TempGroup->box->Min_X>= ( (In_Group->box->Min_X) -Tolerance ) )

_& (TempGroup->box->Min_X<= ( (In_Group->box->Min_X) +Tolerance ) ) ) { // set return flag ReturnFlag = 1;

// Step_l_l : check top direction if ⁽ TempGroup->box->Min Y > In Group->box->Max Y ⁾ { ^"

// check top direction relative views if ( In_Group->to_top == NULL ) {

TempGroupList = (struct group_list ^♦) calloc (1, sizeof (struct group_list )) ;

TempGroupList->car = TempGroup; TempGroupList->cdr = NULL; In_Group->to_top - TempGroupList;

82 APPENDIX D else {

// check distance if ( In_Group->to_top->car->box->Min_Y > TempGroup->box->Min_Y ) {

// change top view In_Group->to_top->car = TempGroup;

else

// Step_l_2 : check bottom direction if ( TempGroup->box->Max Y < In Group->box->Min Y

⁾ { ^"

// check top direction relative views if ( In_Group->to_bottom == NULL ) {

TempGroupList = (struct group_list ♦) calloc (1, sizeof (struct group_list ));

TempGroupList->car = TempGroup; TempGroupList->cdr = NULL; In_Group->to_bottom = TempGroupList; else {

// check distance if ( In_Group->to_bottom->car->box->Max_Y < TempGroup->box->Max_Y ) (

// change top view

In Group->to bottom->car = TempGroup;

} else

// Step_2 : check left and right direction i f

( (TempGroup->box->Max_Y>= ( (In_Group->box->Max_Y) -Tolerance ) )

__(TempGroup->box->Max_Y<= ( (In_Group->box->Max_Y) +Tolerance ) )

( (In_Group->box->Min_Y) -Tolerance ) )

_δc(TempGroup->box->Min_Y<= ( (In_Group->box->Min_Y) +Tolerance ) ) ) {

// set return flag ReturnFlag = 1;

// Step_2_l: check right direction if ( TempGroup->box->Min_X > In_Group->box->Max_X

⁾ {

// check right direction relative views if ( In_Group->to_right == NULL ) {

TempGroupList = (struct group_list

83 APPENDIX D

♦) calloc (1, sizeof (struct group_list )) ;

TempGroupList->car = TempGroup;

TempGroupList->cdr = NULL;

In Group->to right = TempGroupList;

} else {

// check distance if ( In_Group->to_right->car->box->Min_X

_> TempGroup->box->Min_X ) {

// change left view

In Group->to right->car = TempGroup;

}

} else

// Step_2_2 : check left direction if ( TempGroup->box->Max_X < In_Group->box->Min_X

) {

// check left direction relative views if ( In_Group->to_left == NULL ) {

TempGroupList = (struct group_list

♦) calloc (1, sizeof (struct group_list )) ;

TempGroupList->car = TempGroup;

TempGroupList->cdr = NULL;

In Group->to left = TempGroupList;

} else {

// check distance if ( In_Group->to_left->car->box->Max_X <

TempGroup->box->Max_X ) {

// change right view

In_Group->to_left->car = TempGroup;

» ' '

// normal end return ReturnFlag;

/*

Function:

Name : F2dCleanUpFindViewsRelation Description: look for views relation for a drawing. Return Value:

= 0: Completed successfully and No found;

84 APPENDIX D

= NUM: Views' number Subfunctions: int F2dCleanUpFindRelativeViews look for relative views Parameters:

Input :

1 In_Drawing: a drawings __ int F2dCleanUpFindViewsRelation( struct drawing ^♦In_drawing )

{

// external functions int F2dCleanUpFindRelativeViews 0 ;

// link list and temp list struct group

^♦MaxGroup,

^♦TempGroup,

^♦NextGroup; struct group_list

♦OpenGroupList,

♦TempOpenGroupList,

♦TempGroupList; struct node_list ^♦ETL; // Entity Temp List

// entities counter and flags int NoFlag = 0;

// double double BiggestSize = 0.0,

BiggestSizeTemp = 0.0; double Tolerance = 0.0,

ToleranceViews = 0.0;

// set tolerance

Tolerance = F2dCleanUpTolerance;

// check views if ( In_drawing->views -= NULL ) return 0;

// Step_l : get the biggest group's size for ( TempGroup= In_drawing->views;

TempGroup; // ScSc TempGroup->next; TempGroup = TempGroup->next) { // clear flags TempGroup->index = 0; // compute size for box BiggestSizeTemp = fabs ( TempGroup->box->Max_X - TempGroup->box->Min_X )

+ fabs ( TempGroup->box->Max_Y - TempGroup->box->Min_Y

) ;

// check size if ( BiggestSizeTemp > BiggestSize ) { BiggestSize = BiggestSizeTemp;

85 APPENDIX D

MaxGroup = TempGroup;

}

// check last one if ( TempGroup->next == NULL ) break;

}

// check views tolerance

ToleranceViews = 0.01 ^♦ BiggestSize; if ( ToleranceViews < Tolerance ) ToleranceViews = Tolerance;

// Step_2 : find relationship for each view // initialize a open list

OpenGroupList = (struct group_list ^♦) calloc (1, sizeof (struct group_list ) ) ;

OpenGroupList->car = MaxGroup; OpenGroupList->cdr = NULL; // set a open flag to the group MaxGroup->index = 1; for ( TempGroupList = OpenGroupList; TempGroupList; ) ( // get a pointer from open list if ( TempGroupList->car == NULL ) break; NextGroup = TempGroupList->car; // set a closed flag to the group NextGroup->index = 2;

// close the group ( delete the group from open list ) OpenGroupList = OpenGroupList->cdr; // get relation

NoFlag = F2dCleanUpFindRelativeViews ( In_drawing, NextGroup, -ToleranceViews ) ; if ( NoFlag == 0 ) break;

// check each direction for the view if ( NextGroup->to_top != NULL ) { if ( NextGroup->to_top->car->index == 0 ) {

TempOpenGroupList = (struct group_list ^♦) calloc(1, sizeof (struct group_list ));

T e m p O p e n G r o u p L i s t - > c a r NextGroup- >to_t op- >car ;

TempOpenGroupList->cdr - OpenGroupList; TempOpenGroupList->car->index = 1; OpenGroupList = TempOpenGroupList;

if ( NextGroup->to_bottorn != NULL ) { if ⁽ NextGroup->to_bottom->car->index == 0 ) {

TempOpenGroupList = (struct group_list ^♦) calloc (1,sizeof (struct group_list ));

T e m p O p e n G r o u p L i s t - > c a r NextGroup- >to_bottom->car ;

TempOpenGroupList->cdr = OpenGroupList; TempOpenGroupList->car->index = 1_; OpenGroupList = TempOpenGroupList;

86 APPENDIX D

if ( NextGroup->to_left != NULL ) { if ( NextGroup->to_left->car->index == 0 ) {

TempOpenGroupList = (struct group_list ♦) calloc (1, sizeof (struct group_list ) ) ;

T e m p O p e n G r o u p L i s t - > c a r NextGroup- >to_lef t - >car ;

TempOpenGroupList->cdr = OpenGroupList; TempOpenGroupList->car->index = 1; OpenGroupList = TempOpenGroupList;

if ( NextGroup->to_right != NULL ) { if ( NextGroup->to_right->car->index == 0 ) {

TempOpenGroupList = (struct group_list ♦) calloc (1, sizeof (struct group_list )) ;

T e m p O p e n G r o u p L i s t - > c a r NextGroup- >to_right - >car ;

// assign value to the loop variable TempGroupList = OpenGroupList;

}

// Step_3 : clear flags for no relative groups for ( NextGroup=In_drawing->views; NextGroup; NextGroup NextGroup->next) { if ( NextGroup == NULL ) break; if ( NextGroup->index != 0 ) continue; for ( ETL=NextGroup->entities; ETL; ETL=ETL->cdr ) { ETL->car->flags = 0; if ( ETL->cdr == NULL ) break;

// normal end return 0;

} /*

Function:

Name : F2dCleanUpFindDrawingSheet

Description: look for drawing sheet for a drawing

87 APPENDIX D

Return Value:

= -1 : Error data;

0: Completed successfully; 1: No sheet; Subfunctions: struct node_list ^♦F2dCleanUpFindGroup look for a group entities connected with input entity

Parameters:

Input :

1 In_drawing: drawing

___ _ int F2dCleanUpFindDrawingSheet ( struct drawing ^♦In_drawing, struct bounding_box ^♦InOut_BB, int ^ΛIn TimesFlag )

{

// external functions struct node_list ^♦F2dCleanUpFindGroup() ;

// selected entity and temp entities struct node ^♦entity, ^♦LongestEntityX, ^♦LongestEntityY; struct node_list ^♦SheetGroup, ^♦temp_list2;

// bounding box struct bounding_box BB, EBB;

// define int SheetFlag = 0, count = 0,

CountPrompt = 0; double LongestLengthX = 0.0,

LongestLengthY = 0.0,

AlphaX = 0.0, AlphaY = 0.0; char message [64] ; double Tolerance = 0.0; // set tolerance Tolerance = F2dCleanUpTolerance;

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Start Find Sheet Entity %d", CountPrompt ) ; ck__prompt ( message ) ; #endif

// STEP_1 : look for the longest line type parent entity

// check all of entities on the list for( entity=In_drawing->objects; entity;

88 APPENDIX D entity = entity->next) { // only line type if ( entity == NULL ) break; if ( entity->type != CK_LINE ) continue; // only not used if ( entity->flags != 0 ) continue; // check length if ( ( entity->parent->length <= LongestLengthX ) -S ( entity->parent->length <= LongestLengthY )) continue,

// check horizontally and vertically

// if angle (alpha) was smaller than ( sin ( alpha ) alpha )

AlphaX = fabs ( ( entity->parent->Max_X entity->parent->Min__X )

/ entity->parent->length ) ; AlphaY = fabs ( ( entity->parent->Max_Y entity->parent->Min_Y )

/ entity->parent->length ) ; if ( ( AlphaX > Tolerance ) ScSc ( AlphaY > Tolerance ) ) continue;

// change max value and entity if ( AlphaY < Tolerance ) { if ( entity->parent->length > LongestLengthX ) {

LongestLengthX = entity->parent->length; LongestEntityX = entity;

else { if ( entity->parent->length > LongestLengthY ) {

LongestLengthY = entity->parent->length; LongestEntityY = entity;

#ifdef VELLUM

//CountPrompt++;

//sprintf ( message, "Sheet Entity %d" , CountPrompt ) ;

//ck_prompt ( message ) ; #endif

}

// check the longest entitiy's length if ( ( LongestLengthX < Tolerance )

I I ( LongestLengthY < Tolerance )) return -1;

// set bounding box

BB.Min_X = LongestEntityX->parent->Min_X BB.Max_X = LongestEntityX->parent->Max_X BB.Min_Y = LongestEntityY->parent->Min_Y

89 APPENDIX D

BB.Max_Y = LongestEntityY->parent->Max_Y;

// check inside drawing sheet size if ( ♦In_TimesFlag == 1) ^♦InOut_BB = BB; // set values else { if ( (InOut_BB->Min_X =- 0.0 ) &_ (InOut_BB->Max_X == 0.0

)

__(InOut_BB->Min_Y == 0.0 ) __ (InOut_BB->Max_Y == 0.0

) ) return 0; if (( BB.Max_X - BB.Min_X ) < Tolerance ) return ξ> if ( f bs ( ( InOut_BB->Max_X - InOut_BB->Min_X ) / (

BB.Max X - BB.Min X )) > 1.2 ) return 0;

/ r/ set a flag to entity LongestEntityX->flags = 3;

// STEP_2 : look for connected entities group and make boundary

SheetGroup = F2dCleanUpFindGroup( LongestEntityX, -count ) ; if ( SheetGroup == 0 ) return 0;

// STEP_3 : clear flags for all entities for ( entity=In_drawing->objects; entity; entity = entity->next) { if ( entity =- NULL ) break; if ( entity->flags == 3 ) continue; if ( entity->flags == 4 ) continue; entity->flags = 0;

}

// STEP_4 : set flags the group for ( temp_list2 = SheetGroup; ( temp_list2 !=0 Sc& temp_list2->car!=0 ) ; temp_list2=temp_list2->cdr) { if ( temp_list2->car == NULL ) break; temp list2->car->flags = 3;

}

// STEP_5 : set expanding box ( 20% for +X,-X,+Y,-Y )

EBB.Min_X = ( BB.Min_X ) - 0.2 ^♦ fabs ( ( BB.Max_X ) - ( BB.Min_X ) ) ;

EBB.Max_X = ( BB.Max_X ) + 0.2 ^♦ fabs ( ( BB.Max_X ) - ( BB.Min_X ) ) ;

EBB.Min_Y = ( BB.Min_Y ) - 0.2 ^♦ fabs ( { BB.Max_Y ) - ( BB.Min_Y ) ) ;

EBB.Max_Y = ( BB.Max_Y ) + 0.2 ^♦ fabs ( ( BB.Max_Y ) - ( BB.Min_Y ) ) ;

// STEP_6: check outside entity( out of the bounding box

90 APPENDIX D

) in the DRAWING

// if any entity was found, no sheet was found for this drawing for ( entity=In_drawing->objects; entity; entity = entity->next) { if ( entity == NULL ) break; if ( entity->flags != 0 ) continue; // look for out side entity if ( entity->Max_X > EBB.Max_X ) { SheetFlag = 1; break;

} if ( entity->Max_Y > EBB.Max_Y ) {

SheetFlag = 1; break;

} if ( entity->Min_X < EBB.Min_X ) (

SheetFlag = 1; break;

} if ( entity->Min_Y < EBB.Min_Y ) {

SheetFlag = 1; break;

}

// STEP_7 : look for sheet data ( out side of the boundary

) if ( SheetFlag == 0 ) {

// look for sheet data for ( entity=In_drawing->objects; entity; entity = entity->next) { if ( entity == NULL ) break; if ( entity->flags == 3 ) entity->flags = 4; if ( entity->flags != 0 ) continue; if (( entity->Min_X <= BB.Min_X ) i j ( entity->Max_X

>= BB.Max_X )

! ( entity->Min_Y <= BB.Min_Y ) | | ( entity->Max_Y >= BB.Max Y ) )

// set flag entity->flags =

// Good return 0; }

// STEP_8 : clear flags ( No sheet was found ! ) for( entity=In_drawing->objects;

91 APPENDIX D entity; entity = entity->next) { if ( entity == NULL ) break; if ( entity->flags == 4 ) continue; entity->flags = 0;

} if ( SheetFlag != 0 ) return 1;

#ifdef VELLUM

CountPrompt++; sprintf ( message, "Finished find Sheet Entity %d" , CountPrompt ) ; ck_prompt ( message ) ; #endif return 0;

}

/*

Function:

Name : TestFindDrawingSheetControl

Description:

It is a manual function. To find drawing sheet for a drawing. Return Value:

< 0 : Error = 0 : No proccess

> 0: Completed successfully; the Number is the counter of found entities. Parameters: Subfunctions: int F2dCleanUpFindDrawingSheet look for drawing sheet for a drawing, int F2dCleanUpChangeDrawingFlag change entities' flags ( 4 >> 3 ) __ int TestFindDrawingSheetControl ( int ^♦SwitchRedraw )

{

// external functions int F2dCleanUpFindDrawingSheet () ; int F2dCleanUpChangeDrawingColor() ; int F2dCleanUpChangeDrawingFlag() ;

// define struct bounding_box BB; int ReturnFlag = 0; int TimesFlag = 0;

//

92 APPENDIX D

// Step_2 : find center line entities

TimesFlag = 1;

ReturnFlag = F2dCleanUpFindDrawingSheet ( _trim_drawing, _BB, ScTimesFlag ) ; if ( ReturnFlag == 1 ) return 0; if ( ReturnFlag == -1 ) return 0;

TimesFlag = 2;

F2dCleanUpFindDrawingSheet ( _trim_drawing, _BB, ScTimesFlag ) ; //

// Step_3 : change flags

F2dCleanUpChangeDrawingFlag( _trim_drawing ) ;

// Step_X: change color to light gray for a selected center lines

// turn off input levels #ifdef CKWIN if ( ♦SwitchRedraw == 1 ) ck_levels (CKJDFF, 1, 255) ; #endif if ( ^♦SwitchRedraw == 1 ) F2dCleanUpChangeDrawingColor(_trim_drawing) ;

// turn ON the level used to store the trimmed entities. #ifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck_levels (CK_ON, TrimLevel , TrimLevel) ;

//if ( ^♦SwitchRedraw == 1 ) ck_redraw ( CK_PRIME_VP ) ; #endif return 0;

} /*

Function:

Name : TestFindViewsControl

Description:

It is a manual function. To find views for a drawing. Return Value:

< 0 : Error = 0 : No proccess

> 0: Completed successfully; the Number is the counter of found entities. Parameters : Subfunctions : int F2dCleanUpFindViews

93 APPENDIX D look for view's groups for a drawing int F2dCleanUpFindViewsRelation look for views relation for a drawing int F2dCleanUpChangeDrawingColor

*/^" int TestFindViewsControl ( int ^♦SwitchRedraw )

{

// external functions int F2dCleanUpFindViews () ; int F2dCleanUpFindViewsRelation() ; int F2dCleanUpChangeDrawingColor () ;

// Step_2 : find views' group if ( F2dCleanUpFindViews ( _trim_drawing ) == -1 ) return

-l;

// Step_3 : find views' relationship

// move to delete function ( Mar. 29,1996 ) //F2dCleanUpFindViewsRelation( _trim_drawing ) ;

// Step_X: change color to light gray for a selected center lines

// turn off input levels #ifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck_levels (CK_OFF, 1, 255) ; #endif if ( ^♦SwitchRedraw == 1 ) F2dCleanUpChangeDrawingColor(_trim_drawing) ;

// turn ON the level used to store the trimmed entities. #ifdef CKWIN if ( ^♦SwitchRedraw _= 1 ) ck_levels (CK_ON, TrimLevel , TrimLevel) ; if ( ^♦SwitchRedraw == 1 ) ck_redraw( CK_PRIME_VP ) ,- #endif return 0;

}

Function:

Name : TestFindBoundaryControl

Description:

It is a manual function. When user selected a

94 APPENDIX D boundary' s entity of a view, the function would find all of the connected entities for the view.

Return Value:

< 0 : Error = 0 : No proccess

> 0: Completed successfully; the Number is the counter of found entities. Parameters : Subfunctions: int F2dCleanUpFind0utSideLoop look for out side loop entities int F2dCleanUpFindInSideEntity look for in side entities for outside loop int F2dCleanUpFind0utNextEntity look for connected entity with out side loop entities struct node ^♦F2dCleanUpFindGroupMinX look for minimum X-value for a entities group struct node_list ^♦F2dCleanUpFindGroup look for a group entities connected with input entity struct node ^♦F2dCleanUpInputEntity ( old function ) look for a selected entity on screen int F2dCleanUpLineAngle compute a line type entity' s angle int F2dCleanUpArcAngle compute a arc type entity's angle, int F2dCleanUpChangeChildrenFlag change entity's children flags int F2dCleanUpChangeArrowVertexFlag change arrow top vertex's flags and set arrow center line coordinates int F2dCleanUpFindViewsRelation look for views relation for a drawing int F2dCleanUpChangeColor change entity's color double F2dCleanUpLineLength compute line type entity's length __

int TestFindBoundaryControl ( int ^♦SwitchRedraw )

{

// external functions int F2dCleanUpChangeDrawingColor() ; int F2dCleanUpFind0utSideLoop () ;

95 APPENDIX D int F2dCleanUpFindInSideEntity() ; int F2dCleanUpFindViewsRelation () ; struct node ♦F2dCleanUpInputEntity() ; struct node ♦F2dCleanUpFindGroupMinX() ; struct node_list ^♦F2dCleanUpFindGroup() ;

// entities counter and flags int count = 0;

// link list and temp list struct node_list ^♦Entities_list;

// selected entity and temp entities struct node

^♦entity,

^♦MinXvalueEntity;

// Begin

//

// Step_l:

// check the number of entites // get count of all entities // if an error occurred, exit immidiately. if (number_of_entities<0) return -1; for( ; ; ) { //

// Step_2 : select one entity entity = F2dCleanUpInputEntity( "Select a Boundary Entity for a view",

_trim_drawing) ; if ( entity == 0 ) return 0; if ( entity->flags != 0 ) continue;

//

// Step_3 : find a connected entities group

Entities list = F2dCleanUpFindGroup( entity, -count ); //

// Step_4 : find outside loop entities without assistant entities

// check a flag for outside loop process if ( count < 1 ) continue;

MinXvalueEntity = F2dCleanUpFindGroupMinX (

Entities_list) ; if ( MinXvalueEntity == 0 ) continue; //

// Step_5 : find outside loop entities

F2dCleanUpFindOutSideLoop( MinXvalueEntity ) ; //

// Step_6 : find inside entities

F2dCleanUpFindInSideEntity( _trim_drawing, Entities list

⁾;

//

96 APPENDIX D

// Step_7 : find views' relationship

// move delete function (Mar.29, 1996 By Ji )

//F2dCleanUpFindViewsRelation( _trim_drawing ) ; /

// Step_X: change color to Magenta for a selected boundary

// turn off input levels #ifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck_levels (CK_OFF, 1, 255) ; #endif

F2dCleanUpChangeDrawingColor(Sctrim_drawing) ;

// turn ON the level used to store the trimmed entities. #ifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck_leveIs (CK_ON,TrimLevel, TrimLevel) ; if ( ^♦SwitchRedraw == 1 ) ck_redraw( CK_PRIME_VP ) ; #endif

} return count;

}

/*

Function:

Name : TestFindCenterLmeControl

Description:

It is a manual function.

To find center line for all arcs in the view. Return Value:

< 0: Error = 0 : No proccess

> 0: Completed successfully; the Number is the counter of found entities. Parameters : Subfunctions : struct node_list ^♦F2dCleanUpPickUpAllArcEntity look for all arc entities in selected drawing int F2dCleanUpChangeDrawingColor change entity's color __

int TestFindCenterLmeControl ( int ^♦SwitchRedraw )

{

// external functions int F2dCleanUpFindCenterLineEntity() ; int F2dCleanUpChangeDrawingColor() ; struct node_list ♦F2dCleanUpPickUpAllArcEntity() ;

97 APPENDIX D

// link list and temp list struct node_list

♦Entities_list;

// Begin

// Step_l: check the number of entites // if an error occurred, exit immidiately. if (number_of_entities<0) return -1;

// Step_2: pickup all arc entities

Entities_list = F2dCleanUpPickUpAllArcEntity( _trim_drawing ) ; if (Entities_list == 0) return 0; /

// Step_3 : find center line entities

F2dCleanUpFindCenterLineEntity( _trim_drawing, Entities_list

) ;

//

// Step_X: change color to light gray for a selected center lines

// turn off input levels #ifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck_levels (CK_OFF, 1, 255) ; #endif if ( ^♦SwitchRedraw -= 1 ) F2dCleanUpChangeDrawingColor(_trim_drawing) ;

// turn ON the level used to store the trimmed entities. #ifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck_leve l s ( CK_ON , TrimLeve l ,

TrimLevel) ; if ( ^♦SwitchRedraw == 1 ) ck_redraw( CK_PRIME_VP ) ; #endif return 0; }

/ *

Function:

Name : TestFindOneSideOpenConnectControl

Description:

It is a manual function.

To find one side open connected entities. Return Value:

< 0 : Error

= 0: No proccess

> 0: Completed successfully; the Number is the counter of found entities.

98 APPENDIX D

Parameters: Subfunctions : int F2dCleanUpFindOnesideOpenConnectEntity look for one side open connected entities. ____ int TestFindOneSideOpenConnectControl { int ^♦SwitchRedraw )

{

// external functions int F2dCleanUpFindOnesideOpenConnectEntity() ; int F2dCleanUpChangeDrawingColor() ;

/

// Step_l : check the number of entites

// if an error occurred, exit immidiately. if (number_of_entities<0) return -1; /

// Step_2 : find center line entities

F2dCleanUpFindOnesideOpenConnectEntity( Setrim_drawing ) ; //

// Step_X: change color to light gray for a selected center lines

// turn off input levels #ifdef CKWIN if ( ♦SwitchRedraw == 1 ) ck_levels (CK_OFF, 1, 255) ; #endif if ( ^♦SwitchRedraw == 1 ) F2dCleanUpChangeDrawingColor (_trim_drawing) ;

// turn ON the level used to store the trimmed entities. #ifdef CKWIN if ( ♦SwitchRedraw == 1 ) ck_levels ( CK_ON, TrimLevel , TrimLevel) ; if ( ^♦SwitchRedraw == 1 ) ck_redraw( CK_PRIME_VP ) ; #endif return 0;

}

/ *

Function:

Name : TestFindSameColorControl

Description:

It is a manual function.

To find same color entities.

Return Value:

= 0: Completed successfully; the Number is the counter of

99 APPENDIX D found entities. Parameters: Subfunctions: struct node ^♦F2dCleanUpInputEntity ( old function ) look for a selected entity on screen int F2dCleanUpFindSameColorEntity look for same color entities by a selected entity int F2dCleanUpChangeDrawingColor change entity's color

* / int TestFindSameColorControl ()

{

// external functions int F2dCleanUpChangeDrawingColor() ; int F2dCleanUpFindSameColorEntity() ; struct node ^♦F2dCleanUpInputEntity() ;

// selected entity and temp entities struct node ^♦entity;

// Begin

//

// Step_l:

// check the number of entites // if an error occurred, exit immidiately. if (number_of_entities<0) return -1; for( ; ; ) { //

// Step_2 : select one entity entity = F2dCleanUpInputEntity( "Select a Unuse Color Entity",

//

// Step_3 : find a connected entities group

F2dCleanUpFindSameColorEntity( entity, Sctrim drawing ) ; //

// Step_X: change color to Magenta for a selected boundary

// turn off input levels #ifdef CKWIN ck_levels(CK_OFF, 1,255) ; #endif

F2dCleanUpChangeDrawingColor(_trim_drawing) ; // turn ON the level used to store the trimmed entities. #ifdef CKWIN ck_levels (CK_ON,TrimLevel, TrimLevel) ;

100 APPENDIX D if ( ^♦SwitchRedraw == 1 ) ck_redra ( CK_PRIME_VP ) ; #endif

} return 0;

}

/ *

Function:

Name : TestFindSelectSingleControl

Description:

It is a manual function. To find a use entity.

Return Value :

= 0: Completed successfully; Parameters: Subfunctions : struct node ^♦F2dCleanUpInputEntity ( old function ) look for a selected entity on screen __ int TestFindSelectSingleControl ()

{

// external functions struct node ^♦F2dCleanUpInputEntity() ;

// selected entity and temp entities struct node ^♦entity;

// entity' s specifitys

CK_ENTATT attr; // specificity of entities

// Begin

//

// Step_l: check the number of entites // if an error occurred, exit immidiately. if (number_of_entities<0) return -1; for( ; ; ) { //

// Step_2: select one entity entity = F2dCleanUpInputEntity( "Select an Unuse Entity", _trim_drawing) ; if ( entity == 0 ) return 0; //

// Step_3 : change flag if ( entity->flags == 7 ) { entity->flags = 0; attr.color = CK BLUE;

} else if ( entity->flags == 0 ) {

101 APPENDIX D entity->flags = 7; attr.color = CK YELLOW;

} else return 0; /

// Step_X: change color to light gray for unuse entity

// turn off input levels // ck_levels(CK_OFF, 1,255) ; ck_setattr( entity->id, CK_COLOR, NULL, _attr ) ;

// turn ON the level used to store the trimmed entities, // ck_leveIs (CK_ON,TrimLevel, TrimLevel) ; ck redraw( CK PRIME VP ) ;

} ^" return 0;

} /*

Function:

Name : TestUnselectSingleControl

Description:

It is a manual function. To clear an unuse entity. Return Value:

= 0: Completed successfully; Parameters: Subf nctions: struct node ^♦F2dCleanUpInputEntity ( old function ) look for a selected entity on screen

int TestUnselectSingleControl ()

// external functions struct node ^♦F2dCleanUpInputEntity() ;

// selected entity and temp entities struct node ^♦entity;

// entity's specifitys

CKJ3NTATT attr; // specificity of entities

// Begin

//

// Step_l : check the number of entites // if an error occurred, exit immidiately. if (number of entities<0) return -1; for(

//-

102 APPENDIX D

// Step_2 : select one entity entity = F2dCleanUpInputEn ity( "Select an Entity to release" ,

_trim_drawing) ; if ( entity -= 0 ) return 0;

// Step_3 : change flag if ( entity->flags != 3 __ entity->flags != 4 cSc entity->flags != 10 ) continue; entity->flags = 0;

// Step X: change color to light gray for unuse entity 7/ turn off input levels ck_levels(CK_OFF, 1,255) ; attr.color = entity->color; ck_setattr( entity->id, CK_COLOR, NULL, _attr );

// turn ON the level used to store the trimmed entities. ck_levels (CK_ON,TrimLevel, TrimLevel) ; ck redra ( CK PRIME VP ) ;

} return 0;

Function:

Name : TestClearKeepFlagsControl

Description:

It is a manual function.

To return all selected entities to back

Return Value:

= 0: Completed successfully;

Parameters: Subfunctions: int F2dCleanUpClearDrawingKeepFlags change entity's color to back but keep flags _____

int TestClearKeepFlagsControl ( int ^♦SwitchRedraw )

{

// external functions int F2dCleanUpClearDrawingKeepFlags () ;

// Begin

//

103 APPENDIX D

// check the number of entites

// if an error occurred, exit immidiately. if (number_of_entities<0) return -1;

// turn off input levels

#ifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck_levels (CK_0FF, 1,255) ; #endif

F2dCleanUpClearDrawingKeepFlags (_trim_drawing) ; if ( ^♦SwitchRedraw == 1 ) ck_redra ( CK_PRIME_VP );

// turn ON the level used to store the trimmed entities. #ifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck_levels (CK_0N, TrimLevel , TrimLevel) ; if ( ^♦SwitchRedraw == 1 ) ck_redraw( CK_PRIME_VP ) ; #endif return 0; }

*

Function:

Name : TestClearControl

Description:

It is a manual function.

To return all selected entities to back

Return Value:

= 0: Completed successfully;

Parameters: Subfunctions: int F2dCleanUpClearDrawingEntity change entity's color to back __

int TestClearControl ( int ^♦SwitchRedraw )

{

// external functions int F2dCleanUpClearDrawingEntity() ;

// Begin

//

// check the number of entites

// if an error occurred, exit immidiately.

104 _

APPENDIX D if (number_of_entities<0) return -1; /

// turn off input levels #ifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck_levels(CK_OFF,1,255) ; #endif

F2dCleanUpClearDrawingEntity(Sctrim_drawing) ;

// turn ON the level used to store the trimmed entities. #ifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck_levels (CK_ON, TrimLevel , TrimLevel) ; if ( ^♦SwitchRedraw == 1 ) ck_redraw( CK_PRIME_VP ) ; #endif return 0; }

/ *

Function:

Name : TestDeleteControl

Description:

It is a manual function. To delete all selected entities.

Return Value:

= 0: Completed successfully;

Parameters: Subfunctions: int F2dCleanUpDeleteDrawingEntity delete selected entity

*/ int TestDeleteControl( int ^♦SwitchRedraw )

{

// external functions int F2dCleanUpDeleteDrawingEntity() ;

// Begin

//

// check the number of entites

// if an error occurred, exit immidiately. if (number_of_entities<0) return -1; // Step_7: find views' relationship

// move from find boundary function (Mar.29, 1996 By Ji

105 APPENDIX D

F2dCleanUpFindViewsRelation( Sctrim_drawing );

#ifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck_levels (CK_OFF, 1,255) ; #endif

// turn ON the level used to store the trimmed entities.

F2dCleanUpDeleteDrawingEntity(_trim_drawing) ; #ifdef VELLUM if ( ^♦SwitchRedraw == 1 ) F2dCleanUpChangeDrawingColor(_trim_drawing) ; #endif #ifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck_levels (CK_ON, TrimLevel , TrimLevel) ; if ( ^♦SwitchRedraw == 1 ) ck_redra ( CK_PRIME_VP ); #endif return 0; }

*

Function:

Name : TestBackColorControl

Description:

To undo. Return Value:

= 0 : Completed successfully; Parameters : Subfunctions : int F2dCleanUpUndo undo

*/ int TestBackColorControl ( int ^♦SwitchRedraw )

{

// external functions int F2dCleanUpChangeDrawingColor() ; if ( ^♦SwitchRedraw == 1 ) F2dCleanUpChangeDrawingColor(Sctrim_drawing) ; return 0; }

/*

Function:

106 APPENDIX D

Name : TestUndoControl Description:

To undo. Return Value :

= 0: Completed successfully; Parameters : Subfunctions: int F2dCleanUpUndo undo

*/ int TestUndoControl ( int ^♦SwitchRedraw )

{

// external functions int F2dCleanUpUndo() ;

// check the number of entites : if an error occurred, exit immidiately. if (number_of_entities<0) return -1; // Step_l: get flags if ( F2dCleanUpUndo ( _trim_drawing ) == 0 ) return 0; ttifdef CKWIN if ( ^♦SwitchRedraw == 1 ) ck_levels (CK_OFF, 1, 255) ; #endif

// turn ON the level used to store the trimmed entities. #ifdef VELLUM if ( ^♦SwitchRedraw == 1 ) F2dCleanUpChangeDrawingColor(_trim_drawing) ; #endif #ifdef CKWIN if ( ♦SwitchRedraw == 1 ) ck_levels (CK_ON, TrimLevel , TrimLevel) ; if ( ♦SwitchRedraw == 1 ) ck_redra ( CK_PRIME_VP ); #endif return 0; }

/*

Function:

Name : TestUndoPreControl

Description:

To prepare undo. Return Value:

= 0: Completed successfully; Parameters: Subfunctions : int F2dCleanUpUndoPre

107 APPENDIX D to prepare undo

*/ int TestUndoPreControl ( int ^♦SwitchRedraw )

{

// external functions int F2dCleanUpUndoPre () ;

// check the number of entites: if an error occurred, exit immidiately. if (number_of_entities<0) return -1; // Step_l: get flags

F2dCleanUpUndoPre ( _trim_drawing ) ; return 0;

} /*

Function:

Name : TestSetUndoFlagViewsControl

Description:

To set a flag for undo. flag=0: views are not exist flag=l : views are exist. Return Value:

= 0: Completed successfully; Parameters:

'/ int TestSetUndoFlagViewsControl ( int ^♦SwitchRedraw )

// set flag F2dCleanUpExcutedNo = 1; return 0;

}

/'

Function:

Name : TestDeleteViewsControl

Description:

To delete views. Return Value:

= 0: Completed successfully; Parameters:

108 APPENDIX D

Subfunctions: int F2dCleanUpDeleteViews to delete views __—

int TestDeleteViewsControl ( int ^♦SwitchRedraw )

{

// external functions int F2dCleanUpDeleteViews () ;

// check the number of entites: if an error occurred, exit immidiately. if (number_of_entities<0) return -1; // delete views

F2dCleanUpDeleteViews ( _trim_drawing ) ; return 0; } struct node ♦F2dCleanUpInputEntity(char ^♦prompt, struct drawing ♦select drawing)

{ int etype, count, stat; unsigned long entid; struct node ♦entity;

// CK_ENTATT attr; if (( ck_getent (prompt, _etype, _entid) != CK_N0ERR0R) I ( stat == CK_BACKUP ) j ( stat == CK_ESCAPE )) return 0;

#ifdef DEBUG printC'ID: %ld; #",entid);

#endif count=0; for (entity=select_drawing->objects; entity St_ (count<number_trimmed_entities) 6c_ (entity->id !=entid) &_ (entity->highlighted_id !=entid) ; entity=entity->next) count++; if (count>=number_trimmed_entities) return 0; else return entity;

} /*

Function:

109 APPENDIX D

Name : F2dVector

Description:

To compute line's unit vector.

Return Value :

- Out_v: Completed successfully;

Parameters:

Input :

1 In_sp: line entity start point coordinates

2 In_ep: line entity end point coordinates Output :

1 Out_v: line entity vector

*7 struct Vector F2dVector( struct Vector In_sp, struct Vector In_ep

)

{ struct Vector Out_v;

Out_v.x - In_ep.x - In_sp.x; Out_v.y = In_ep.y - In_sp.y; Out_v.z = In_ep.z - In_sp.z; return Out v; }

*

Function:

Name : F2dVectorUnit

Description:

To compute line's unit vector.

Return Value:

- Out_n: Completed successfully; < 0 : Error. Parameters:

Input :

1 In_v: line entity vector Output:

1 Out_n: line entity unit vector __

struct Vector F2dVectorUnit ( struct Vector In_v)

110 APPENDIX D

struct Vector Out_n; double Length = 0.0;

// tolerance double tolerance = 0.0; tolerance = F2dCleanUpTolerance;

Length = sqrt( In_v.x^In_v.x + In_v.y^In_v.y + In_v.z^ΛIn_v.z

// check length if ( Length < tolerance ) return Out n;

Out_n.x = In_v.x / Length Out_n.y - In_v.y / Length Out n.z = In_v.z / Length return Out n;

*

Function:

Name : F2dDistancePointLine

Description:

To compute distance from a point to a line.

Return Value:

= 0: Completed successfully;

Parameters:

Input:

1 In_sp: line entity start point coordinates

2 In_ep: line entity end point coordinates

3 In_p: point coordinates. Output:

1 Out_d: distance __

double F2dDistancePointLine(

// input struct Vector In_sp, struct Vector In_ep, struct Vector In_p

)

{

// external functions struct Vector F2dVector(); struct Vector F2dVectorUnit () ;

// define

111 APPENDIX D struct Vector Ve, Vne, Vp; double Out_d=0.0;

// get vector

Ve = F2dVector( In_sp, In_ep ) ;

Vp = F2dVector( In_sp, In_p ) ;

// get unit vector

Vne = F2dVectorUnit ( Ve ) ;

// compute distance

Out_d = sqrt ( Vp.x+Vp.x + Vp.y^Vp.y

-⁽Vp.x^Vne.x + Vp.y⁺Vne.y) ^♦ (Vp.x^Vne.x ₊ Vp.y^Vne.y) ) ,- return Out d;

112 APPENDIX E iiiinmiiiiiii in 117 ii II i nu 'iiiiiin 'illinium

II II

1/ Example of bend model viewer implementation; contains // // implementation of CBendCADViewPart member functions, // // and implementation of menu-driven functions. //

// //

////////////////////////////////////////////////////////////////// #include "stdafx.h"

// include all BMAPI files #include "ALLBMAPI .HXX"

// RW_DRAW library include file. #include "RW_DRAW.^■HXX"

#include "BendCAD.h" #include "BendCADDoc.h" #include "BendCADViewPart .h"

// RenderWare include files; found in \rwwin\include #include "rwlib.h"

// OpenGL include files. #include "gl\gl.h" #include "gl\glu.h"

// GL_LIST library include file. #include "GL_LIST.HXX"

#include <stdlib.h>

// Orthoview dialog include file #include "OrthoViewDlg.h"

minium

#define WINDOWTITLE "BAJA CAD SYSTEM" static void multiply_pt_by__matrix(BM_POINT ^♦point, float rot_matrix[4] [4] ) ; static RwClump ^♦ RWCALLBACK do__transformations (RwClump ♦clump, void

^♦matrix) ,- static RwClump ^♦ RWCALLBACK clear_scene (RwClump ♦clump) ; static RwPolygon3d ^♦ RWCALLBACK SetPolygonColor (RwPolygon3d

^♦polygon, void ^♦color) ;

// The following variables are defined to set up the color palette

1 APPENDIX E of the screen. unsigned char threetoδ [8] =

0, 0111>>1, 0222>>1, 0333>>1, 0444>>1, 0555>>1, 0666>>1, 0377

unsigned char twotoθ [4] = 0, 0x55, Oxaa, Oxff

unsigned char onetoθ [2] = 0, 255

static int defaultOverride [13] =

0, 3, 24, 27, 64, 67, 88, 173, 181, 236, 247, 164, 91

static PALETTEENTRY defaultPalEntry[20] =

192, 220, 192, 0

166, 202, 240, 0

255, 251, 240, 0

160, 160, 164, 0

void PrintString(char ^♦s) APPENDIX E

}

// This function initializes the RenderWare library, creates a scene, camera and light. int CBendCADViewPart: :Init3D(HWND window)

{

BOOL berr = FALSE;

// Open the RenderWare library, if ( IRwOpen("MSWindows" , NULL) )

{

: :MessageBox(window,

"Could not open the RenderWare library",

WINDOWTITLE, MB_OK | MB_ICONSTOP |

MB_APPLMODAL) ; return 0; '

}

// Create a camera.

Rwlnt32 x_size, y_size; x_size = GetSystemMetrics(SM_CXFULLSCREEN) ; // The camera should be able to display y_size = GetSystemMetrics (SM_CYFULLSCREEN) ; // on the maximum size allowed by the screen. m_camera = RwCreateCamera (x_size, y_size, NULL) ; if ( ! m camera)

{

: :MessageBox(window, "Could not create camera",

WINDOWTITLE, MB_OK | MB_ICONSTOP j

MB_APPLMODAL) ; return 0; else {

RwSetCameraProjection(m_camera, rwPARALLEL) ;

RwWCMoveCamera(m_camera, O.Of, O.Of, 10000.Of) ;

// Move camera far away from xy plane. RwSetCameraViewport (m__camera, 0, 0, x_size, y_size) ;

// Default sizes of viewing volume RwSetCameraViewwindow (m_camera , (float) (x_size) , (float) (y size)) ; // and viewport.

} ^"

// Create a scene. m_scene = RwCreateScene () ; if (m scene == NULL)

{

RwDestroyCamera (m_camera) ; m_camera = NULL; RwClose () ; APPENDIX E

: :MessageBox(window, "Could not create the 3D Scene",

WINDOWTITLE, MB_OK j MB_IC0NST0P | MB_APPLMODAL) ; return 0;

}

berr = TRUE; if (light2)

{ m_lights.Add(light2) ; RwAddLightToScene (m_scene, light2) ,- else berr = TRUE; if (light3)

{ m_lights.Add(light3) ,- RwAddLightToScene (m_scene, light3) ; else berr = TRUE; if ⁽berr⁾ AfxMessageBoxCCould not create light source") ;

//RwSetLightState(m_light, rwON) ;

//RwSetLightColor (m_light, CREAL (1.0) , CREAL (1.0) , CREAL(1.0)) ; return 1; }

// This function is used by the document class to initialize certain APPENDIX E

// parameters whenever a new part is loaded. What we do here is compute the centroid,

// prepare RenderWare clumps and show the part as requested - wireframe/shaded if show_solid is 0/1,

// flat/3d version of part if show_3d is 0/1. void CBendCADViewPart: :new_part_init (int show_solid, int show_3d)

{ int i;

BM_PART ^part = NULL;

CBendCADDoc^ pDoc = GetDocument () ;

ASSERT_VALID(pDoc) ; part = pDoc->get_part () ,- if(! part) { mi_part_is__up_to_date = 0; return;

} mi__jpart_is_up_to_date = 1; compute_part_centroid(part, 0, _m_flat_part_centroid) ; / /

Compute 3d and flat part centroids . - compute_part_centroid(part, 1, _m_3d_part_centroid) ; m_part_centroid = m_3d_part_centroid; mi_show_3d = show_3d; mi_show_solid = show_solid;

// Clear out any RenderWare clumps around, if (m_flat_base_clump) {

RwRemoveClumpFromScene (m_flat_base_clump) ;

RwDestroyClump(m_flat_base_clump) ; m flat base clump - NULL; i ^}f (^"m_3d_^~base_^"clump) {

RwRemoveClumpFromScene (m_3d_base_clump) ;

RwDestroyClump(m_3d_base_clump) ; m 3d base clump = NULL;

} ^{" " "} m_base_clump = NULL; if (mi_show_3d) { // Prepare

3d solid or 3d wireframe. m_part_centroid - m_3d_part_centroid; if ( ! mi_show_solid) { // Prepare wireframe . i = prepare_display__list (part, mi_show_3d, 1, 0, 1, 1, md_face_color, md_bendline_color) ; if(! i) { mi_display_list_is_up_to_date = 0; return; } APPENDIX E

w re rame. i = prepare_display_list (part, mi_show_3d, 1, 0, 1, 1, md_face_color, md_bendline_color) ; if(! i) { mi_display_list_is_up_to_date = 0; return;

} mi_display_list_is_up_to_date = 1; mi_flat_is_up_to_date = 1; mi_3d_is_up_to_date = 0; else { //

Prepare solid. i - facet_part_rw(part, _m_flat_base_clump, mi_show_3d, md_part_color) ; i ⁽ (! i) M (! m_flat_base_clump) ) return; m_base_clump = m_flat_base_clump; RwAddClumpToScene (m_scene, m_base_clump) ; mi_display_list_is_up_to_date = 0; mi_3d_is_up_to_date = 0; mi_flat_is_up_to_date = 0;

mi_bbox_is_up_to_date = 0; ml_num_faces = part->get_number_of_faces () ; ml_num_bendlines = part->get_number_of_bendlines 0 ; ml_n_m_formings = part->get_number_of_formings() ; clear_selection_set () ; reset_selection bufferO ; APPENDIX E OnViewIsometric ( ) ,-

}

// This function destroys all parent clumps in the scene, and hence all clumps in the scene.

// It always returns NULL. static RwClump ♦ RWCALLBACK clear_scene(RwClump ♦clump) int i = RwGetClumpNumChildren(clump) ,- if(i) RwDestroyClump(clump) ; return NULL; }

// This allows other classes to signal that the part has changed, and hence

// current clumps/display lists are no longer valid. We recompute centroids, prepare new clumps

// or display lists, and show the part in its current orientation and zoom. void CBendCADViewPart::invalidate display(void)

{ int i;

BM_PART ^part = NULL;

CBendCADDoc^ pDoc = GetDocument() ;

ASSERT_VALID(pDoc) ; part = pDoc->get_part() ; if(! part) { mi_part_is_up_to_date - 0; return;

} mi__part_is_up_to_date = 1; compute_part_centroid(part, 0, _m_flat_part_centroid) ; / / Compute 3d and flat part centroids. compute_part_centroid(part, 1, _m_3d_part_centroid) ; mi_3d_is_up_to_date = 0; mi_flat_is_up_to_date - 0;

// Clear out any RenderWare clumps around, if (m_flat_base_clump) {

RwRemoveClumpFromScene(m_flat_base__clump) ;

RwDestroyClump(m_flat_base_clump) ; m flat base clump = NULL; i ^}f(^"m_3d_^"base_^"cl_mp) {

RwRemoveClumpFromScene (m_3d_base_clump) ; APPENDIX E

RwDestroyClump(m_3d_base_clump) ; m 3d__base_clump = NULL; m }_base__clump = NULL; if (mi_show_3d) { // Prepare 3d solid or 3d wireframe. m_part_centroid = m_3d_part_centroid; if ( ! mi_show_solid) ( // Prepare wireframe . i = prepare_display_list (part, mi_show_3d, 1, 0, 1, 1, md face_color, md_bendline_color) ; if(! i⁾ { mi_display_list_is_up_to_date = 0; return;

} mi_display_list_is_up_to_date = 1; mi_3d_is_up_to_date = 1; mi_flat_is_up_to_date = 0;

} , else ( //

Prepare solid. i = facet_part_rw(part, _m_3d_base_clump, mi show 3d, md_part_color) ; if ( (! i) | | (! m_3d_base_clump) ) return; m_base_clump = m_3d_base_clump;

RwAddClumpToScene (m_scene, m_base_clump) ;

} else { // Prepare flat solid or flat wireframe. m_part_centroid = m_flat_part_centroid; if ( ! mi_show_solid) { // Prepare wireframe. i = prepare_display_list (part, mi_show_3d, 1, 0, 1, 1, md_face_color, md_bendline color) ; if(! i) { mi_display_list_is__p_to_date = 0; return;

Prepare solid. i = facet_part_rw(part, _m_flat_base_clump, mi show 3d, md_part_color) ; ^~ if^{( (}! i⁾ | J ⁽! m_flat_base_clump) ) return; m_base_clump = m_flat_base_clump; APPENDIX E RwAddClumpToScene (m_scene, m_base_clump) ;

} } mi_bbox_is_up_to_date = 0; ml_num_faces = part->get_number_of_faces() ; ml_num_bendlines = part->get_number_of_bendlines () ; ml_num_formings = part->get_number_of_formings () ; clear_selection_set () ; reset_selection_buf er () ; ml_last_function_chosen = RETURN_FROM_DIALOG_MODE;

DrawPart (m_hdc) ; ml last function chosen = NO FUNCTION MODE;

void CBendCADViewPart: :OnShowFlat () { if ( 1 mi_show_3d) return; // Flat is already being shown. int i;

BM_PART ^♦part = NULL;

CBendCADDoc^ pDoc = GetDocument () ;

ASSERT_VALID(pDoc) ; part = pDoc->get_part () ,- if(! part) { mi_part_is_up_to_date = 0; return;

}

// Store the current 3d view parameters. md_old_x_trans = md_x_trans; md_old_y_trans = md_y trans; for(i=0; i < 3; ++i) J md_old__part_bbox_size [i] = md_part_bbox_size [i] ; md old view vol [i] = md current view vol [i] ;

} ^{" "} - ^{" " "} if (md_old__rot_matrix) RwDestroyMatrix(md_old_rot_matrix) ; md_old_rot_matrix = RwDuplicateMatri (md_rot_matrix) ; mi_show_3d = 0; m_part_centroid = m_flat_part__centroid; if ( ! mi_show_solid) ( // Compute the flat's display list. if ( ! mi_flat_is_up_to_date) { i = prepare_display_list (part, mi_show_3d, 1, 0, 1, 1, md face color, md bendline color) ; APPENDIX E if(! i) { mi_display_list_is_up_to_date = 0; return;

}

} mi_display_list_is_up_to_date = 1; mi_flat_is_up_to_date = 1; mi_3d_is_up_to_date = 0; else { // Compute flat's clump if not already done. if (m_3d_base_clump) RwRemoveClumpFromScene (m_3d_base_clump) ; m_base_clump = NULL; if ( ! m_flat_base_clump) { i = facet_part_rw(part, _m_flat_base_clump, mi_show_3d, md_part_color) ; if( (! i) | | (! m_flat_base_clump) ) { mi_display_list_is_up_to_date = 0; return;

' » m_base_clump = m_flat_base_clump; // Set the base clump to be the flat's base clump. _} RwAddClumpToScene (m-scene, m-base-clump) ;

mi_bbox_is_up_to_date = 0; OnViewTop () ;

// Here we show the 3d version, in its old orientation and zoom

(before flat was

// asked to be shown) . void CBendCADViewPart : :0nShow3d()

{ if (mi_show_3d) return; // 3D view is already being shown. int i;

BM_PART ^♦part = NULL;

CBendCADDoc^ pDoc = GetDocument () ;

ASSERT_VALID(pDoc) ; part = pDoc->get_part () ; if(i part) { mi_part_is_up_to_date = 0; return;

}

10 APPENDIX E

// Restore the old 3d view parameters. md_x_trans = md_old_x_trans; md_y_trans = md_old_y trans; for(i=0; i < 3; ++i) md_part_bbox_size [i] = md_old_part_bbox_size [i] ; md_current_view_vol [i] = md_old_view_vol [i] ; if (md_rot_matrix) RwDestroyMatrix(md_rot_matrix) ; md_rot_matrix = RwDuplicateMatrix(md_old_rot_matrix) ; mi_show_3d = 1; _part_centroid = m_3d_part_centroid; if ( ! mi_show_solid) { if ( ! mi_3d_is_up_to_date) { i = prepare_display_list (part, mi_show_3d, 1, 0, 1, 1, md_face_color, md_bendline_color) ; if(! i) { mi_display_list_is_up_to_date = 0; return;

} mi_display_list_is_up_to_date = 1; mi_3d_is_up_to_date = 1; mi_flat_is_up__to_date = 0;

else { // Compute the 3d_version's clumps if have not already done so. i f ( m _ f l a t _ b a s e _ c l u m p ) RwRemoveClumpFromScene (m_flat_base_clump) ; m_base_clump = NULL; if ( ! m_3d_base_clump) { i = facet_part_rw(part, _m_3d_base_clump, mi_show_3d, md_part_color) ; if( (! i) | | (! m_3d_base_clump) ) ( mi_display_list_is_up_to_date = 0; return;

) ' m_base_clump = m_3d_base_clump; // Set the base clump to be the 3d version's base clump.

RwAddClumpToScene (m_scene, m_base_clump) ;

mi_bbox__is__up_to_date = 0; md_current_view_vol [1] = md_current_view_vol [0] /md_aspect;

RwSetCameraViewwindo (m_camera, (float) (md_current_view_vol [0] ) , (float) (md current view vol [1] ) ) ;

11 APPENDIX E glMatrixMode (GL_PROJECTION) ; glLoadldentity() ; glOrtho(- md_current_view_vol [0] , md_current_view_vol [0] , md_current_view_vol [ 1 ] , md_current_view_vol [1] , // - 50000.0, 50000.0) ;

- md_current_view_vol [2] , md_current_view_vol [2] ) ; glMatrixMode (GL_MODELVIEW) ;

DrawPart (m hdc) ;

void CBendCADViewPart: :OnHide3d()

{

OnShowFlat () ; // For now, just switch to showing flat. }

void CBendCADViewPart : :OnHideFlat ()

{

OnShow3d(); // For now, just switch to showing

3D.

}

void CBendCADViewPart: :OnViewWireframe ()

{ if ( ! mi_show_solid) return; // Wireframe already being shown, int i;

CreateRGBPalette(m_hdc) ; // when wireframe -- recreate the palette -- AK

HGLRC hrc - wglGetCurrentContext {) ; // The rendering context of OpenGL. wglMakeCurrent (NULL, NULL) ; wglMakeCurrent (m_hdc, hrc) ; // Associate the rendering context with this view.

BM_PART ^part = NULL; CBendCADDoc^ pDoc = GetDocument () ; ASSERT_VALID(pDoc) ; part = pDoc->get_part () ; if(! part) { mi_part_is_up_to_date = 0; return;

}

12 APPENDIX E mi_show_solid = 0;

ml_last_function_chosen = IREFRAME_MODE;

DrawPart (m_hdc) ; ml_last_function_chosen = NO_FUNCTION_MODE;

}

void CBendCADViewPart: :OnViewShaded() { if (mi_show_solid) return; // Solid already being shown, int i;

BM_PART ^♦part = NULL; CBendCADDoc^ pDoc = GetDocument () ; ASSERT_VALID(pDoc) ; part = pDoc->get_part () ; if(! part) { mi_part_is_up_to_date = 0; return;

}

13 APPENDIX E mi_show_solid = 1; if (mi_show_3d) { if(! m_3d_base_clump) { i = facet_part_rw(part, „m_3d_base_clump, mi_show_3d, md_part_color) ; if( (! i) | | (! m_3d_base_clump) ) { mi_display_list_is_up_to_date = 0; return;

> >

RwRemoveClumpFromScene (m_base_clump) ; m_base_clump = m_3d_base_clump; // Set the base clump to be the 3d version's base clump.

RwAddClumpToScene (m_scene, m_base_clump) ; else { if(! m_flat_base_clump) { i = facet_part_rw(part, _m_flat_base_clump, mi_show_3d, md_part_color) ; if( (! i) | | (! m_flat_base_clump) ) { mi_display_list_is_up_to_date = 0; return;

> >

RwRemoveClumpFromScene (m_base_clump) ; m_base_clump = m_flat_base_clump; // Set the base clump to be the 3d version's base clump.

RwAddClumpToScene (m scene, m base clump);

} ^{" "} ml_last_function_chosen = SOLID_MODE;

DrawPart (m_hdc) ; ml_last_function_chosen - NO_FUNCTION_MODE;

void CBendCADViewPart: :OnZoomWindow()

{ ml_last_function_chosen = ZOOM_WINDOW_MODE; mi bbox is up to date = 0; } ^{~ " ~ " ~} void CBendCADViewPart : :OnUpdateZoomWindo (CCmdUI^ pCmdUI) pCmdUI->SetCheck⁽ (Z00M_WIND0W_M0DE == ml last function chosen) ? TRUE : FALSE) ; - - _

}

14 APPENDIX E void CBendCADViewPart: :OnZoomInOut ()

{ ml_last_function_chosen = ZOOM_IN_OUT_MODE; mi bbox_is_up_to date = 0;

} ^" void CBendCADViewPart : :OnUpdateZoomInOut (CCmdUI^ pCmdUI)

{ pCmdUI->SetChec ( (ZOOM_IN_OUT_MODE == ml_last_function_chosen)

? TRUE : FALSE) ; } void CBendCADViewPart: :OnZoomAll () {

BM_PART ^♦part = NULL; CBendCADDoc^ pDoc; BM_VECTOR offset; if ( ! mi_bbox_is_up_to_date) ( pDoc = GetDocument () ; ASSERT_VALID(pDoc) ; part = pDoc->get_part () ; if ( ! part) return; if ( I mi_part_is_up_to_date) { // By now we expect the centroid to have already compute part_centroid (part , 0 ,

_m_f lat_part_centroid) ; (J been computed; this is only a safety check. compute_part_centroid(part, 1, _m_3d_part_centroid) ; if (mi_show_3d) m_part_centroid = m_3d_part_centroid; else m_j?art_centroid = m_f lat_part_centroid; compute_part_bbox (part , mi_show_3d, md_rot_matrix, m_part_centroid, md_part_bbox_size, -offset) ; mi_bbox_is_up_to_date = 1; md_x_trans = - (offset.X() ) ; // Adjust for the fact that bbox center may md_y_trans = - (offset.Y() ) ; // not coincide with centroid of part . } set_new_view_volume (md_part_bbox_size [0] ,md_part_bbox_size [1] ) ; ml_last_f nction_chosen = ZOOM_ALL_MODE;

DrawPart (m_hdc) ; ml last function chosen = NO FUNCTION MODE;

15 APPENDIX E void CBendCADViewPart : :OnUpdateZoomAll (CCmdUI^ pCmdUI)

{ pCmdUI->SetCheck( (ZOOM_ALL_MODE == ml_last_function_chosen) ?

TRUE : FALSE) ; }

void CBendCADViewPart: :OnViewIsometric ()

{ ml_last_function_chosen - ISOMETRIC VIEW MODE; mi_bbox_is_up_to_date = 0; md_z_angle = 270.0; md_y_angle = 305.2643897; // 270 + tan l(l/sqrt(2) )

RwRotateMatrix (RwScratchMatrix ( ) , 0. Of, 0.Of, 1. Of CREAL(md_z_angle) ,• rwREPLACE) ;

RwRotateMatrix (RwScratchMatrix{) , O.Of, l.Of O.Of, CREAL(md_y_angle) , rwPRECONCAT); md_z_angle = 45.0;

RwRotateMatrix (RwScratchMatrix ( ) , 0. Of O.Of l.Of CREAL (md_z_angle) , rwPRECONCAT) ; if (md_rot_matrix) RwDestroyMatrix (md_rot_matrix) ; md_rot_matrix = RwDuplicateMatrix (RwScratchMatrix ( ) )

// glPushMatrix() ;

// glLoadldentityO ;

// glRotated(md_z_angle, 0.0, 0.0, 1.0) ; // Rotation about z axis.

// glRotated(md_y_angle, 0.0, 1.0, 0.0) ; // Rotation about y axis .

// md_z_angle = 45.0;

// glRotated(md_z_angle, 0.0, 0.0, 1.0) ;

// glGetDoublev(GL_MODELVIEW_MATRIX, md_rot_matrix) ;

// glPopMatrix() ;

OnZoomAll () ;

void CBendCADViewPart: :OnViewRight () ml_last_function_chosen = RIGHT_VIEW_MODE; mi_bbox_is_up_to_date = 0; md_z_angle = 270.0; md_y_angle = 270.0; md_x_angle = 0.0;

RwRotateMatrix (RwScratchMatrix ( ) , O.Of O.Of, 1. Of, CREAL(md_z_angle) , rwREPLACE) ;

RwRotateMatrix (RwScratchMatrix () , O.Of 1. Of, O.Of, CREAL(md_y_angle) , rwPRECONCAT) ;

16 APPENDIX E if (md_rot_matrix) RwDestroyMatrix(md_rot_matrix) ; md_rot_matrix = RwDuplicateMatrix(RwScratchMatrix()) ;

// glPushMatrixO ;

// glLoadldentity() ;

// glRotated(md_z_angle, 0.0, 0.0, 1.0) ; // Rotation about z axis.

// glRotated(md_y_angle, 0.0, 1.0, 0.0) ; // Rotation about y axis.

// glGetDoublev(GL_MODELVIEW_MATRIX, md_rot_matrix) ;

// glPopMatrixO ;

OnZoomAll () ; }

void CBendCADViewPart: :OnViewFront ()

{ ml_last_function_chosen = FRONT_VIEW_MODE; mi_bbox_is_up_to_date = 0; md_y_angle = 0.0; md__z_angle = 0.0; md_x_angle = 270.0;

RwRotateMatrix (RwScratchMatrix () , l.Of, O.Of, O.Of, CREAL(md_x_angle) , rwREPLACE) ; if (md_rot_matrix) RwDestroyMatrix(md_rot_matrix) ; md_rot_matrix = RwDuplicateMatrix(RwScratchMatrix()) ;

// glPushMatrixO;

// glLoadldentity() ;

// glRotated(md_x_angle, 1.0, 0.0, 0.0) ; // Rotation about x axis.

// glGetDoublev(GL_MODELVIEW_MATRIX, md_rot_matrix) ;

// glPopMatrixO ,-

OnZoomAll 0 ; }

void CBendCADViewPart : :OnViewTop()

{ ml__last_function_chosen = TOP_VIEW_MODE; mi_bbox_is_up_to_date = 0; RwIdentityMatrix(md_rot_matrix) ;

OnZoomAll () ; } void CBendCADViewPart : :OnViewsOrtho()

{

17 APPENDIX E if (NULL == GetDocument () ->get_part () ) { // no part return ; }

COrthoViewDlg orhodlg(NULL, GetDocument () ->get_part , this) int ret = orhodlg.DoModal ;

// if we are in a Client Mode, the user cannot have entered anything.

// we have to redisplay the main client menu bar. if (BENDCAD_CLIENT_MODE) {

: :PostMessage(this->GetSafeHwnd() ,WM_COMMAND,MAKEWPARAM(ID_DISPLA Y_CLIENTJMENU, 1) ,NULL) ; return ;

}

// when OK pressed, get return values if (IDOK == ret) {

void CBendCADViewPart : :OnViewBack ()

{ ml_last_function_chosen = BACK_VIEW_MODE; mi_bbox_is_up_to_date = 0; md_x_angle = 270.0; md_y_angle = 0.0; md_z_angle - 180.0;

RwRotateMatrix (RwScratchMatrix() , l.Of, O.Of, O.Of CREAL(md_x_angle) , rwREPLACE) ;

RwRotateMatrix (RwScratchMatrix( ) , O.Of, O.Of, O.lf CREAL(md_z_angle) , rwPRECONCAT); if(md_rot_matrix) RwDestroyMatrix(md_rot_matrix) ; md_rot_matrix = RwDuplicateMatrix(RwScratchMatrix()) ;

OnZoomAll () ;

} void CBendCADViewPart: :OnViewBottom() ml_last_function_chosen = BOTTOM_VIEW_MODE; mi_bbox_is_up_to_date = 0; md_x_angle = 0.0; md_y_angle = 180.0; md_z_angle = 0.0;

18 APPENDIX E

RwRotateMatrix (RwScratchMatrix ( ) , O . O f , l . O f , O . Of ,

CREAL (md_y_angle) , rwREPLACE) ; if (md_rot_matrix) RwDestroyMatrix (md_rot_matrix) ; md_rot_matrix = RwDuplicateMatrix (RwScratchMatrix ( ) ) ;

OnZoomAll O ; } void CBendCADViewPart : :OnViewLeft 0

{ ml_last_function_chosen = LEFT_VIEW_MODE; mi_bbox_is_up_to_date = 0,- md_z__angle = 90.0; md_y_angle = 90.0; md_x_angle = 0.0;

RwRotateMatrix(RwScratchMatrix () , O.Of, O.Of, l.Of, CREAL(md_z_angle) , rwREPLACE) ;

RwRotateMatrix (RwScratchMatrix () , O.Of, l.Of, O.Of, CREAL( d_y_angle) , rwPRECONCAT); if (md_rot_matrix) RwDestroyMatrix(md_rot_matrix) ; md_rot_matrix = RwDuplicateMatrix(RwScratchMatrix0 ) ;

OnZoomAll 0 ; }

void CBendCADViewPart : :OnRotate ()

{ ml_last_function_chosen = ROTATION_MODE; mi bbox is up to date = 0; } ^{" "} void CBendCADViewPart : :OnUpdateRotate (CC dUI^ pCmdUI)

{ pCmdUI->SetCheck( (ROTATION_MODE == ml_last_function_chosen) ? TRUE : FALSE) ;

}

void CBendCADViewPart: :OnPa () { ml_last_function_chosen - PAN_MODE; mi_bbox_is_up_to_date = 0;

} void CBendCADViewPart: :OnUpdatePan(CCmdUI^ pCmdUI) pCmdUI->SetCheck ( (PAN_MODE == ml_last_function_chosen) ? TRUE

19 APPENDIX E

: FALSE) ; }

// This function computes the rotation angles about x and y axes. // It is called on mouse movement, if the rotation mode is on. void CBendCADViewPart: :compute_rotation_angles(int x_flag)

{

// else

// glRotated(md_z_angle, 0.0, 0.0, 1.0); // Rotation about z axis.

// glMultMatrixd(md_rot_matrix) ;

// glGetDoublev(GL_MODELVIEW_MATRIX, md_rot_matrix) ;

// glPopMatri ;

}

// This function computes the x and y translation values. // It is called on mouse movement, if the pan mode is on. void CBendCADViewPart: :compute_pan_translation(void) double x; if( (x = (double)m_old_rect.right) > 0.0 ) md_x_trans += (m_current_point.x - m_previous_point.x) ♦ md_current_view_vol [0] /x; if ( (x = (double)m_old_rect.bottom) > 0.0 ) md_y_trans += - (m_current_point.y - m_previous_point.y) ♦ //

20 APPENDIX E

-ve sign for md_y_trans md_current_view_vol[1] /x; // bee. y axis points up screen.

}

// This funciton computes translation values when user zooms in/out. void CBendCADViewPart: :compute_zoom_in_out_translation(void)

{ double x; if ( (x = (double)m_old_rect.right) > 0.0 ) md_x_trans += - (m_current_point.x - m_previous_point.x) ^♦ md_current_view_vol[0] /x;

}

// This funciton computes translation values when user zooms into a

// selected rectangular region. void CBendCADViewPart: :compute_zoom_region_translation(void)

{ double xl - ((double) (m_lt_btn_down_point . + m_lt_btn_up_point.x) ) /2.0; double delta_x = xl - ( (double)m__old_rect.right) /2.0; double yl = ((double) (m_lt_btn_down_point .y + m_lt_btn_up_point.y) )/2.0; double delta_y = - ( yl - ( (double)m_old_rect.bottom) /2.0 ); md_x_trans -= //m_part_centroid.X0 - md_x_trans + delta_x⁺2.0⁺md_current_view_vol[0] / ( (double)m_old_rect.right) ; md_y_trans -= //m_part_centroid.Y() - md_y_trans + delta y^A2.0^md current view vol [1] /( (double)m old rect.bottom) ; }

// This function computes the new view volume when the user zooms in/out.

// The part has already been translated appropriately. v_{oid CBendCADViewPart::compute-zoom-in-out--view-volume(void) double x = - (m__current_point.y - m_previous_point.y) ; m d _ c u r r e n t _ v i e w _ v o l [ 0 ] - = md_current_view_vol [0] ^♦x/ (double) m_old_rect . bottom; md_current_view_vol[1] = md_current_ iew_vol [0] /md_aspect; RwSetCameraViewwindow(m_camera, (float) (md_current_view_vol [0] ) , (float) (md_current_view_vol[1] ) ) ;

21 APPENDIX E glMatrixMode (GL_PROJECTION) ; glLoadldentity() ; glOrtho(- md_current_view_vol [0] , md_current_view_vol [0] , md_current_vi ew_vol [ l ] , md_current_view_vol [1] , // - 50000.0, 50000.0) ;

}

// This function computes and sets the new view volume when the user zooms into

// a selected rectangular region. (In the case of zoom all, the selected

// region is simply the bounding box of the part) .

// The part has already been translated appropriately. void CBendCADViewPart: :set_new_view_volume (double x_extent, double y extent)

^{" double longer = max(x_extent, y_extent) ; double shorter = min(x_extent, y_extent) ; if (longer == x_extent) { // The x extent is longer. if (longer/md_aspect < shorter) // y extent does not fit screen. longer = shorter^md_aspect;

lo fi

longer = shorter/md_aspect; md_current_view_vol [0] = 0.55^Λlonger^md_aspect; md_current_view__vol [1] = 0.55^♦longer; md_current_view_vol [2] = maxdonger, md_part_bbox_size [2] ) ;

RwSetCameraViewwindow(m_camera, (float) (md_current_view_vol [0] ) , (float) (md_current_view_vol [1] ) ) ; glMatrixMode (GL_PROJECTION) ; glLoadldentity() ; glOrtho⁽- md_current_view_vol [0] , md_current_view_vol [0] , md_current_view_vol [ 1 ] , md_current_view_vol [1] ,

22 APPENDIX E

// - 50000.0, 50000.0) ;

- md_current_view_vol [2] , md_current_view_vol [2] ) ; glMatrixMode (GL_MODELVIEW) ; glLoadldentityO ;

// This function draws a part on the view window. It is called by

OnDraw. void CBendCADViewPart : :DrawPart (HDC draw_dc)

{ long size, i, parent_tag, key; static long ♦selected_items = NULL;

RwClump ♦pic ed^lump, ^♦parent_clump; static double red_color[3] = {1.0, 0.0, 0.θ}; static double green_color[3] = (0.0, 1.0, 0.0} ; double array[16] ;

BV_SELECT_DATA +data = NULL; static RwMatrix4d ^♦old_matrix[50] ; static float rw_array[4] [4] ; if (mi_simulation_is_on) { // Set camera parameters to view window values.

RwSetCameraViewport (m_camera, 0, 0, m_old_rect.right, m_old_rect.bottom) ; glViewport (0, 0, m_old_rect.right, m_old_rect.bottom) ;

Rw S e t C a m e r aV i e ww i n do w ( m_ c a m e r a , (float) (md_current_view_vol [0] ) , (float) (md_current_view_vol [1] ) ) ; glMatrixMode (GL_PR0JECTI0N) ; glLoadldentity 0 ; glOrtho(- md_current_ iew_vol [0] , md_current_view_vol [0] , md_cu r r e n t _v i ew_vo 1 [1] , md_current_view_vol [1] , // - 50000.0, 50000.0) ;

- md_current_view_vol [2] , md_current_view_vol [2] ) ; glMatrixMode (GL MODELVIEW) ;

} if ( ! mi_show_solid) { // Wireframe drawing; use OpenGL drawing commands. glPushMatrix() ; glTranslated (m_camera_pos it ion . X ( ) , m_camera_position. Y () , m_camera_position. Z () ) ; gluLookAt (m_camera_pos it ion. X () , m_camera_position. Y 0 , m_camera_position. Z () , m_camera_target . X ( ) ,

23 APPENDIX E m_camera_target .Y0 , m_camera_target.Z () , m_camera_look_up.X() , m_camera_look_up.Y0 , m_camera_look_up.Z () ) ; glTranslated(md_x_trans, md_y_trans, 0.0) ; convert_rw_matrix_to_gl_array(md_rot_matrix, array) ; glMultMatrixd(array) ; glTranslated(- m_part_centroid.X 0 , - m_part_centroid.Y() , - m_part_centroid.Z () ) ; glClearColor(0.0f, O.Of, O.Of, l.Of) ; glClear(GL_COLOR_BUFFER_BIT j GL_DEPTH_BUFFER_BIT) ; if (ml_last_function_chosen == SELECT_OBJECTS_MODE) glRenderMode(GL_SELECT) ; else glRenderMode(GL_RENDER) ;

glSelectBuffer(512, mw_selected_objects) ; gllnitNames 0 ; glPushName (0) ; if ( ! mi_simulation_is_on) {

// First draw the selected items in red, if any. size = selection set. list size () ; if (size) { selected_items = new long [size^2 + 1] ; // +1 for safety. m_selection_set.display(size, selected_items) ; glColor3d(1.0, 0.0, 0.0) ; for(i=0; i < size; ++i) { data = (BV SELECT DATA ^♦)selected items [i^Λ2

+1] ; if ( ! data) continue; if (data->edge) { // Edge is available, => not a trivial g l C a l l L i s t ( ( u n s i g n e d int) selected_items [i<<l] ) ; // bendline or body of something . else {

// Body of face/forming/bendline picked; highlight whole thing. key = selected_items [i<<l] /100000 ; glCallList ( (unsignedint) key^lOOOOO) ;

I ' '

24 APPENDIX E glCallList (1) ; glFinishO ; glPopMatrixO ; SwapBuffers(wglGetCurrentDC0 ) ;

} else ( // Solid drawing; use RenderWare.

// Transform the part if necessary, if (ml_last_function_chosen != NO_FUNCTION_MODE) { RwTranslateMatrix (RwScratchMatrix () , CREAL (md_x_trans) , CREAL (md_y_trans ) , CREAL (0.0) , rwREPLACE) ;

RwTransf ormMatrix (RwScratchMatrix ( ) , md_rot_matrix, rwPRECONCAT) ;

RwTrans lateMatrix ( RwScratchMatrix ( ) , CREAL ( - m_part_centroid . X () ) , CREAL ( - m_part_centroid . Y 0 ) ,

CREAL (- m_ art_centroid.Z() ) , rwPRECONCAT) ;

// RwForAllClumpsInScenePointer (m_scene, do_transf ormations, (void ♦ ) RwScratchMatrix () ) ;

RwTransf ormClump(m_base_clump, RwScratchMatrix ( ) , rwREPLACE) ;

}

/*

RwClump ^♦clump; for(i=0; i < ml_num_faces + ml_num_bendlines + ml_num_formings;

++i) { clump = RwFindTaggedClump (m_base_clump, i + 10001) ; if ( ! clump) continue; old_matrix[i] = RwCreateMatrix 0 ; / /

Store the current modeling matrix of the clump. RwGetClumpMatrix (clump, old_matrix [i] ) ;

RwGetMatrixElements (old_matrix [ij , rw_array) ; RwTranslateMatrix (RwScratchMatrix () , CREAL (md_x_t ran s) , CREA (md_y_trans) , CREAL (0.0) , rwREPLACE);

RwTransf ormMatrix (RwScratchMatrix ( ) , md_rot_matrix, rwPRECONCAT) ;

RwTrans la t eMat rix ( RwScr at chMat rix ( ) , CREAL ( - m_part_centroid . X () ) , CREAL ( - m_part_centroid . Y () ) ,

CREAL (- m_part_centroid.Z() ) , rwPRECONCAT) ; //RwForAllClumpsInScenePointer (m_scene , do_transf ormations, (void ^♦) RwScratchMatrix ()) ;

RwTransformClump(clump, RwScratchMatrix() , rwPOSTCONCAT) ;

} */

// Now set the color of the clumps in the selection set to red.

25 APPENDIX E size - m_selection_set. list_size () ; if (size) { selected_items = new long [(size << 1) + 1] ; // +1 for safety. m_selection_set .display(size, selected_items) ; for(i=0; i < size; ++i) { data = (BV_SELECT_DATA ^♦) selected_items [i^2

+1] ; if ( ! data) continue; if (data->edge) { // Edge is available,

=_> not a trivial bendline or body of something. parent_tag = (selected_items [i<<l] /100000) + 10000; p a r e n t _ c l u m p RwFindTaggedClump(m_base_clump, parent_tag) ; else ( key = selected_items [i<<l] %100000; // Check if it is a trivial bendline. if ( (key/10000) _= 6) { p a r e n t _ t a g (selected_items [i<<l] /100000) + 10000; p a r e n t _ c l u m p =

RwFindTaggedClump (m_base_clurnp, parent_tag) ; else parent_clump = m_base_clump; key = selected_items [i<<l] %100000; picked_cl_mp = RwFindTaggedClump (parent_clump, key) ,-

RwForAllPolygonsInClumpPointer(picked_clump, SetPolygonColor, (void ^#)red color);

) '

// In case simulation is going on, we don't want to show the m_sim_base_clump, only m_base_clump. if (mi_simulation_is_on) { i f ( m _ s i m _ b a s e _ c l u m p ) RwRemoveClumpFromScene (m_sim_base_clump) ; if (mi_show_3d) { m_base_clump = m_3d_base_clump; RwAddClumpToScene (m_scene, m_base_clump) ; else { m_base_clump = m_flat_base_clump; RwAddClumpToScene (m_scene, m_base_clumρ) ,-

26 APPENDIX E

// Now draw the part. RwInvalidateCameraViewport (m_camera) ;

RwBeginCameraUpdate(m_camera, (void ^♦)m_hwnd) ; RwClearCameraViewport (m_camera) ; i = RwGetSceneNumClu ps (m_scene) ; // Render the scene.

RwRenderScene (m_scene) ; RwEndCameraUpdate (m_camera) ;

RwShowCameralmage (m_camera, (void^) (DWORD)draw_dc) ; // Show the image on the view window.

/* for(i=0; i < ml_num_faces + ml_num_bendlines + ml_num_formings; ++i) { clump = RwFindTaggedClump(m_base_clump, i + 10001) ; if ( ! clump) continue;

RwTransformClump (clump, old_matrix[i] , rwREPLACE) ; RwGetClumpMatrix(clump, old_matrix[i] ) ;

RwGetMatrixElements (old_matrix [i] , rw_array) ; if (old matrix[i] ) RwDestroyMatrix(old_matrix[i] ) ;

} ^♦/

// Put the simulation clump back in the scene so that simulation can see it. if (mi_simulation_is_on) { if (m_sim_base_clump) RwAddClumpToScene(m_scene, m_sim_base_clump) ; i f ( m i _ s h o w _ 3 d ) RwRemoveClumpFromScene (m_3d_base_clump) ; else RwRemoveClumpFromScene(m_flat_base_clump) ; m base clump = NULL; }

// Now set the color of the clumps in the selection set back to their original colors, if (size) { m_selection_set.display(size, selected_items) ; for(i=0; i < size; ++i) { data = (BV_SELECT_DATA ^♦) selected_items [i*2

+1] ; if (data->edge) ( // Edge is available, => not a trivial bendline or body of something. parent__tag = (selected_items [i<<l] /100000) + 10000; p a r e n t _ c l u m p RwFindTaggedClump (m_base_c__mp , parent_tag) ; else (

27 APPENDIX E key = selected_items [i<<l] %100000; // Check if it is a trivial bendline. if ( (key/10000) == 6) { p a r e n t _ t a g (selected_items [i<<l] /100000) + 10000; p a r e n t _ c l u m p RwFindTaggedClump (m_base_c lump, parent_tag) ; else parent_clump = m_base_clump;

} key = selected_items [i<<l] %100000; picked_clump = RwFindTaggedClump (parent_clump, key) ; if( (! data->edge) __ ((key/10000) == 6) ) // Trivial bendline.

RwForAllPolygonsInClumpPomter(picked_clump, SetPolygonColor, (void ♦)green_color) ; else

RwForAllPolygonsInClumpPointer(picked_clump, SetPolygonColor, (void ♦)md_part color);

if (selected_items) delete [] selected_items; selected items = NULL;

} if (mi_simulation_is_on) { // Restore camera parameters to simulation values.

RwSetCameraViewport (m_camera, 0, 0, mi_bitmap_w , mi_bitmap_h) ; glViewport (0, 0, mi_bitmap_w, mi_bitmap_h) ;

R w S e t C a m e r a V i e w w i n d o w ( m _ c a m e r a , (float) (md_sim_view_vol [0] ) , (float) (md_sim_view_vol [1] ) ) ; glMatrixMode (GL_PROJECTION) ; glLoadldentity 0 ; glOrtho (- md_sim_view_vol [0] , md_sim_view_vol [0] , md_ s i m_vi ew_vo l [ l ] , md_sim_view_vol [1] ,

// - 50000.0, 50000.0) ;

- md_current_view_vol [2] , md_current_view vol [2] ) ; glMatrixMode (GL_MODELVIEW) ;

}

28 APPENDIX E

/* if (size) { for(i=0 ,- i < size; ++i) ( BM_3D_BODY ^b3 ; long three_d_type; if (selected_items [i^2 +1]) { // Edge is available, => not a trival bendline or forming. three_d_type = (selected_items [i<<l] %100000) /10000; if ( (three_d_type >= 3) __ (three_d_type < 6)) three_d_type = BM_ENTITY_TYPE_FACE; else if ( (three_d_type >= 6) __ (three_d__type < 9)) three_d_type = BM_ENTITY_TYPE_BENDLINE; else if (three_d_type >= 9) three_d_type = BM_ENTITY_TYPE_FORMING; parent_tag = (selected_items [i<<l] /100000) + 10000; parent_clump = RwFindTaggedClump (m_base_clump, parent tag) ;

^" e^}lse {

CBendCADDoc^ pDoc = GetDocument 0 ;

ASSERT_VALID(pDθC) ;

BM_PART ♦part = pDoc->get_part () ; if ( ! part) return; // Note: we do not expect part to be NULL at this stage.

BM_TOPOLOGY ^♦topology = part->get_topology0 ; if ( ! topology) return; three_d_type = selected_items [i<<l] /100000; //

Terminology caution: three_d_type is not really b3 = topology->map_id_into_pointer(three_d_type) ,- // a three_d_type until b3->get_type 0. three_d_type = b3->get_type () ; parent_clump - m_base_clump; if (three_d_type == BM_ENTITY_TYPE_FACE) { for(j=0; j < 3; ++j) original_color[j] md face color [j] ;

^{" "} e^}lse if (three_d_type == BM_ENTITY_TYPE_BENDLINE) { for(j=0; j < 3; ++j) original_color [j] md_bendline_color [ j ] ; else if (three_d_type == BM_ENTITY_TYPE_FORMING) ( for(j=0; j < 3; ++j) original_color [j] md_f orming color [j]; key = selected_items [i<<l] ,- key = key%100000; picked_clump = RwFindTaggedClump (parent_clump, key) ; RwForAllPolygonsInClumpPointer (picked_clump , SetPolygonColor, (void ♦) original_color) ;

29 APPENDIX E }

} * /

// This callback function performs the necessary transformations on each clump in the scene.

// Transform only parent clumps, since this automatically transforms child clumps as well. static RwClump ♦ RWCALLBACK do_transformations (RwClump ^♦clump, void

♦matrix)

{ int i = RwGetClumpNumChildren(clump) ; if(i) { if ( ! (i = RwGetClumpNumPolygons (clump) ) ) RwSetClu pState(clump, rwOFF) ; // Do not render this clump if it has no polygons.

RwTransformClump(clump, (RwMatrix4d ^♦)matrix, rwREPLACE) ;

} return clump;

}

// This callback function sets the color of the polygon to the required color. static RwPolygon3d ^♦ RWCALLBACK SetPolygonColor(RwPolygon3d ♦polygon, void ^♦color)

{ polygon = RwSetPolygonColor(polygon, CREAL (( (double

♦) color) [0] ) , CREAL (( (double ^♦) color) [1] ) ,

CREAL ( ( (double

♦)color) [2])) ; return polygon;

}

// This function sets the pixel format of the view window to the kind that

// OpenGL likes. Note: Pixel format of a window should be set only once.

BOOL CBendCADViewPart : :SetupPixelFormat (HDC hDC) static PIXELFORMATDESCRIPTOR pfd = { sizeof (PIXELFORMATDESCRIPTOR) , // size of this pfd

! // version number

PFD_DRAW_TO_WINDOW | // support window

PFD_SUPPORT_OPENGL | // support OpenGL

30 APPENDIX E

PFD_DOUBLEBUFFER, // double buffered

PFD_TYPE_RGBA, // RGBA type

24, // 24-bit color depth

0, 0, 0, 0, 0, 0, // color bits ignored

0, // no alpha buffer

0, // shift bit ignored

0, // no accumulation buffer

0, 0, 0, 0, // accum bits ignored

32, // 32-bit z-buffer

0, // no stencil buffer

0, // no auxiliary buffer

PFD_MAIN_PLANE, // main layer

0, // reserved

0, 0, 0 // layer masks ignored

} int pixelformat ; if ( (pixelformat = ChoosePixelFormat (hDC, _pf d) ) == 0 )

MessageBox O' ChoosePixelFormat failed" ) ; return FALSE ;

} if (SetPixelFormat (hDC, pixelformat, _pfd) == FALSE)

{

MessageBox("SetPixelFormat failed") ; return FALSE; }

// don't need this call any more; the function will be called from OnActivateViewO

// and on OnViewWireframe () -- AK //CreateRGBPalette(hDC) ; return TRUE; }

unsigned char CBendCADViewPart: :ComponentFromlndex(int i, UINT nbits, UINT shift)

{ unsigned char val; val = (unsigned char) (i >> shift) ; switch (nbits) { case 1 : val 8 = 0x1; return onetoβ [val] ; case 2 :

31 APPENDIX E val Sc= 0x3; return twotoδ [val] ; case 3 : val S= 0x7; return threetoθ [val] ; default : return 0;

}

void CBendCADViewPart : :CreateRGBPalette (HDC hDC)

{ PIXELFORMATDESCRIPTOR pfd;

LOGPALETTE ♦pPal; int n, i;

CDC dc; n = : :GetPixelFormat (hDC) ;

: :DescribePixelFormat (hDC, n, sizeof (PIXELFORMATDESCRIPTOR) , -pfd) ; if (pfd.dwFlags _ PFD NEED PALETTE)

{ ^{" ~} n - 1 << pfd.cColorBits; pPal = (PLOGPALETTE) LocalAlloc (LMEM_FIXED, sizeof (LOGPALETTE) + n ^♦ sizeof (PALETTEENTRY) ) ; pPal->palVersion = 0x300; pPal->palNumEntries = n; for (i=0; i<n; i++)

{ pPal->palPalEntry [i] .peRed =

ComponentFromlndex (i, pfd.cRedBits, pfd.cRedShift) ; pPal->palPalEntry[i] .peGreen =

ComponentFromlndex (i , pfd. cGreenBits, pfd. cGreenShif t) ; pPal->palPalEntry [i] .peBlue =

ComponentFromlnde (i , pfd. cBlueBits, pfd.cBlueShift) ; pPal->palPalEntry [i] .peFlags = 0 ;

/^♦ fix up the palette to include the default GDI palette ♦/ if ((pfd.cColorBits == 8) &&

(pfd.cRedBits == 3) _& (pfd.cRedShift == 0) ScSc (pfd. cGreenBits == 3) SS (pf d. cGreenShif t == 3) __ (pfd. cBlueBits == 2) ScS< (pfd.cBlueShift == 6)

32 APPENDIX E

)

{ for ( i = 1 ; i <= 12 ; i++) pPal - >pa l Pa l Ent ry [ de f au l tOve rride [ i ] ] def aultPalEntry [i] ,-

} dc . Attach (hDC) ;

CPalette ^♦pOldPal;

// since we call this function not only when creating a view,

// but also each time activating, delete the prev palette (if any) -- AK int resp = m_cPal.DeleteObject () ; m_cPal .CreatePalette (pPal) ;

LocalFree (pPal) ; pOldPal = dc.SelectPalette(_m_cPal, FALSE) ; dc.RealizePalette () ; dc .Detach0 ;

// This function computes the bounding box parameters of the given part -

// the x, y and z extents, in global coordinates, for the orientation specified by orient_matrix.

// The computation is done with respect to the BM_POINT centroid as origin.

// It does so by taking the bbox of each face, projecting each onto the coordinate

// axes, and then finding the extents along each axis. If threeD_flag is non-zero, the

// three-d version's bbox is computed; else, the flat's bbox is computed. The results of the

// computation are put into the arry bbox_size[3] ; if this pointer is zero, no computation

// is performed. The offset, ie. the vector from the centroid' s coords to the bbox center's coords,

// is passed back in the pointer to the BM_VECTOR.

// Returns 1 if successful, 0 otherwise. int CBendCADViewPart: :compute_part_bbox(BM_PART ^♦part, int threeD_flag, RwMatrix4d ♦orient_matrix,

BM_POINT centroid, double ^♦bbox size, BM VECTOR ^♦offset)

{

33 APPENDIX E

BM_FACE ^♦face = NULL; BM_BENDLINE ♦bendline = NULL; BM_2D_BODY ^♦two_d_body = NULL; BM_LOOP ^♦loop = NULL; BM_POINT dotl, dot2, dot3, dot4 ; double left, right, top, bot, front, back; // First two refer to x direction, // last two to z direction. BM_VECTOR vecl; int first_pass = 1; float rot matrix [4] [4] ; if ( ! part) return 0; if ( ( par t - >ge t_number_of _f ace s ( ) 0) ScSc part->get_number_of_bendlines () == 0) ) rreettuurrnn 00; ; if ( ! bbox size) return 0; if ( I part->get_number_of_faces ()) ^• { // In this case, find bendline bboxes. for(bendline = part->get_bendline_list () ; bendline; bendline = (BM_BENDLINE ^♦)bendline->next () ) {

// First check that the bbox of the bloop is up to date; if not compute bbox. if (threeD_flag) two_d_body = bendline->get_3D_version () ; else two_d_body = bendline->get_flat () ; if ⁽two_d_body⁾ loop = two_d_body->get_bloop () ; else continue; if ( ! loop) continue; if ⁽ ! ⁽loop->is_bbox_up_to_date () ) ) loop->recompute_bbox () ;

// Now find the envelope (union) of the projected bboxes. vecl = BM_VECTOR(centroid.X , centroid.Y() , centroid.Z ()) ; dotl = loop->get_bbox_pl () + (-1.0) ♦vecl // The bbox coordinates, dot2 = loop->get_bbox_p2 () + (-1.0) ♦vecl // with respect to the dot3 = loop->get_bbox_p3 () + (-l.0)*vecl // centroid as origin. dot4 = loop->get_bbox_p4 () + (-1.0) ♦vecl

RwGetMatrixElements (orient_matrix, rot_matrix) _; multiply_jpt_by_matrix(_dotl, rot_matrix) ; // Transform coords by the current multiply_pt_by_matrix(_dot2, rot_matrix) ; // rotation matrix.

34 APPENDIX E multiply_pt_by_matrix ( _dot3 , rot_ matrix) mult iply_pt_by_matrix ( Scdot 4 , rot_ matrix) if (first _pass) { left = min(dotl.X() dot2 .XO) ; right = max(dotl.X() dot2 .XO) ; bot = min(dotl.Y() dot2 •Y0 ) ; top = max(dotl.Y() dot2 ■YD⁾ ; back = min(dotl.ZO dot2 •ZO); front = max(dotl.ZO dot2 •ZO) ; first_pass = 0;

} else { left mm (left, dotl.X left min(left, dot2.X right max(right, dotl.X right max(right, dot2.X bot min(bot dotl.Y bot min(bot , dot2.Y top max(top , dotl.Y top max(top dot2.Y back _min(back , dotl.Z back min(back , dot2.Z front max (front, dotl.Z front max(front, dot2.Z

} left _ min (left, dot3.X( left = min(left, dot4.X( right = max(right, dot3.X( right = max(right, dot4.X( bot min(bot , dot3.Y( bot _min(bot dot4.Y( top max(top , dot3. Y( top max(top , dot4.Y( back min(back , dot3.Z( back min(back , dot4.Z( front = max(front, dot3.Z( front = max(front, dot4.Z (

} bbox_size [0] right left; bbox_size [lϊ top bot; // Copy bbox size into the array passed. bbox size [2] = front - back; if (offset) // Copy the bbox center into the pointer passed.

^♦offset = BM_VECT0R( (right + left)/2.0, (top + bot)/2.0, (front + back)/2.0) ; mi_bbox_is_up_to_date = 1;

35 APPENDIX E return 1;

}

// Faces are available, so find bbox of faces. for (face = part->get_face_list () ; face; face = (BM_FACE ♦) face->next 0 ) { if ( ! face) break;

// First check that the bbox of the bloop is up to date; if not compute bbox. if (threeD_flag) two_d_body = face->get_3D_version () ; else two_d__body = face->get_flat () ; if (two_d_body) loop = two_d_body->get_bloop() ; else continue; if ( ! loop) continue; if ( ! (loop->is_bbox_up_to_date 0 ) ) loop->recompute_bbox() ;

// Now find the envelope (union) of the projected bboxes. vecl = BM_VECTOR(centroid.X () , centroid.Y() , centroid.Z ()) ; dotl = loop->get_bbox_pl () + (-1.0) ^♦vecl // The bbox coordinates, dot2 = loop->get_bbox_p2 () + (-1.0) ^♦vecl // with respect to the dot3 = loop->get_bbox_p3 () + (-1.0) ^♦vecl // centroid as origin. dot4 = loop->get_bbox_jp4 () + (-1.0)⁺vecl

RwGetMatrixElements (orient matrix, rot matrix) multiply_pt_by_matrix(_dotl, rot_matrix) // Transform coords by the current multiply_jpt_by_matrix (_dot2 , rot_matrix) // rotation matrix. multiply_pt_by_matrix (_dot3 , rot_matrix) multiply_pt_by_matrix ( _dot4 , rot_matrix) if (f irst_pass) { left = min (dotl.X () , dot2.X() right = max (dotl.X (), dot2.X() bot = min (dotl. YO , dot2.Y() top = max (dotl. Y () , dot2.Y() back = min(dotl.Z(), dot2.Z() front = max (dotl. Z , dot2.Z() first pass = 0;

} else { left = mindeft, dotl.XO); left = mindeft, dot2.X())_;

36 APPENDIX E right ₌ max(right, dotl.XO ) right = max(right, dot2.X() ) bot = min(bot , dotl.YO ) bot = min(bot , dot2.Y() ) top = max(top , dotl.YO ) top = max(top , dot2.Y() ) back = min(back , dotl.Z () ) back = min(back , dot2.Z() ) front = max(front, dotl.Z (] ) front = max(front, dot2.Z () )

} left - _min(left, dot3.X() ) left = _min(left, dot4.X() ) right = _max(right, dot3.X() ) right = jnax(right, dot4.X() ) bot _miή(bot , dot3.Y() ) bot _min(bot , dot4.Y() ) top _max(top , dot3.Y() ) top _max(top , dot4.Y() ) back _min(back , dot3.Z() ) back _min(back , dot4.Z() ) front = _max(front, dot3.Z() ) front = _max(front, dot4.Z() ) } bbox_size [0] = right - left; bbox_size [1] = top bot; / <7/ Copy bbox size into the array passed. bbox size [2] = front - back; if (offset) // Copy the bbox center into the pointer passed.

♦offset = BM_VECTOR( (right left)/2.0, (top + bot)/2.0, (front + back) /2.0) ; mi_bbox_is_up_to_date - 1; return 1;

// This function computes the part's centroid. This is done with respect to

// the original (given) coordinates of the part. If threeD_flag is non-zero, the

// three-d version's centroid is computed; else, the flat's centroid is computed.

// The results are passed back in the pointer to a BM_POINT; if this pointer is NULL, no

// computation is done. Returns 1 if successful, 0 otherwise.

37 APPENDIX E int CBendCADViewPart : :compute_part_centroid(BM_PART ^♦part, int threeD flag, BM_POINT ♦centroid)

{

BM_FACE ^♦face; BM_2 D_BODY ^♦ two_d_body ; BM_LOOP ^♦loop;

BM_POINT dotl, dot2, dot3, dot4 ; double left, right, top, bot, front, back; // First two refer to x direction,

// last two to z direction. int first_pass = 1; if ( ! part) return 0; if ( ( p a r t - > g e t _numbe r _o f _ f a c e s == 0 ) Sc _

(part->get_number_of_bendlines == 0) ) return 0 ,- if ( ! centroid) return 0 ; for (face = part->get_face_list () ; face; face = (BM_FACE ♦) face->next () ) { if ( ! face) break;

// First check that the bbox of the bloop is up to date; if not compute bbox. if (threeD_flag) two_d_body = face->get_3D_version() ; else two_d_body = face->get_flat () ; if (two_d__body) loop - two_d_body->get_bloop () ; else continue; if ( ! loop) continue; if ( ! (loop->is_bbox_up_to_date () ) ) loop->recompute_bbox0 ;

// Now find the envelope (union) of the projected bboxes. dotl = loop->get_bbox_pl () dot2 = loop->get_bbox_p20 dot3 = loop->get__bbox_p3 O dot4 = loop->get_bbox_jp4 () if (first pass) ( left = min(dotl.X() , dot2.X() right = max(dotl.XO , dot2.X() bot = min(dotl.Y() , dot2.Y() top = max(dotl.Y() , dot2.Y() back = min(dotl.Z() , dot2.Z() front = max(dotl.Z () , dot2.Z() first_pass = 0;

38 APPENDIX E else { left _ min deft, dotl.XO ) left = min deft, dot2.X() ) right = max(right ., dotl.XO) right = max(right , dot2.XO) bot = min(bot , dotl.YO) bot = min(bot , dot2.Y0) top = max(top , dotl.YO) top - max(top , dot2.YO) back _ min(back , dotl.Z ()) back = min(back , dot2.ZO ) front _ max(front, dotl.Z ()) front = ma (front, dot2.Z())

} left - _mindeft, dot3.X() ) ; left = _mindeft, dot4.X() ) ; right = _max(right, dot3.X() ) ; right = _max(right, dot4.X0 ) ; bot _min(bot , dot3.Y() ) ; bot _min(bot , dot4.Y() ) ; top = _max(top , dot3.Y() ) ; top _max(top , dot4.Y() ) ; back _min(back , dot3.Z() ) ; back = _min(back , dot4.Z() ) ; front = _max(front, dot3.Z() ) ; front = max(front, dot4. Z() ) ;

}

♦centroid = BM_P0INT ( (right + left)/2.0, (top + bot )/2.0, (front + back)/2.0) ; // Copy centroid into

/*

// pointer supplied if (threeD_flag) { m_3d_part_centroid.set_X( (right + left)/2.0); m_3d_part_centroid.set_Y( (top + bot )/2.0); // The part' s 3d centroid.

_} _3d_part_centroid.set Z((front + back)/2.0) ; else { m_flat_part_centroid.set_X( (right + left)/2.0) m_flat_part_centroid.set_Y( (top + bot )/2.0) // The part's flat centroid. m flat part centroid.set Z( (front + back)/2.0)

} ^{~ " ~} */ return 1;

}

39 APPENDIX E

// This function multiplies a bend model point by a 4x4 RenderWare transform matrix.

// The function is destructive. It is NOT a BMAPPView member function.

// Note that the point is considered to be a row vector; this is

// RenderWare's (and the Bend Model's) convention. static void multiply_pt_by_matrix(BM_P0INT ^♦point, float rot matrix[4] [4] )

{ ^" double x = point ->X() double y = point ->Y() double z = point ->Z() point->set_X( (x^rot_matrix [0] [0] + y^rot_matrix [1] [0] + z^rot_matrix[2] [0] ) ) ; point ->set_Y( (x^Λrot_matrix [0] [1] + y^rot_matrix [1] [1] + z + rot_matrix[2] [1] ) ) ; point->set_Z ( (x^rot_matrix [0] [2] + y⁺rot_matrix [1] [2] + z^rot_matrix[2] [2] ) ) ;

}

// This function outlines the rectangular region selected by the user. void CBendCADViewPart ::outline region(void)

{

CClientDC view_dc (this) ;

CPoint cornerl_curr = m_lt_btn_down_point , corner2_curr m_lt_btn_down_point ;

CPoint cornerl_prev = m_lt_btn_down_point , corner2_prev m_lt_btn_down_point ; cornerl_curr .Offset (m_current_point.x - m It btn down point. x, _{0) ;} _ _ _ co r ne r 2_cur r . O f f s e t ( 0 , m_ c u r r e n t _p o i n t . y m_lt_btn_down_point .y) ; cornerl_prev. Offset (m_previous_point.x - m_lt_btn_down_point .x, u / ; c or ne r 2_pr e v . Of f s e t ( 0 , m_p r e v i ou s_p o i n t . y m_lt_btn_down_point .y) ; view_dc.SetROP2 (R2_NOT) ; view_dc.MoveTo(m_lt_btn_down_point) ; view_dc.LineTo(cornerl_prev) ; view_dc. oveTo(m_lt_btn_down_point) ; view_dc.LineTo(corner2_prev) ;

40 APPENDIX E view_dc.MoveTo(cornerl_prev) ; view_dc.LineTo (m_previous_point) ; view_dc.MoveTo(corner2_prev) ; view_dc.LineTo (m_previous_point) ; view_dc.MoveTo(m_lt_btn_down_point) ; view_dc.LineTo(cornerl_curr) ; view_dc.MoveTo(m_lt_btn_down_point) ; view_dc.LineTo(corner2_curr) ; view_dc. oveTo(cornerl_curr) ; view_dc.LineTo(m_current_point) ; view_dc.MoveTo(corner2_curr) ; view_dc.LineTo (m_current_point) ;

// This function processes the hits record to pick out the edge(s) selected by the

// user. If add_hit_to_list is non-zero, which it is if the control button is held down,

// the current hit, if any, is added to the selection set; else, the selection set

// is cleared and only the current hit is added to the list. If there is no hit, the

// selection set is cleared as well, unless add_to_list is non-zero, in which case

// nothing happens.

// This function recognizes trivial bendlines and formings and modifies the selection set

// appropriately. v _{oid CBendCADViewPart: :process hits recorddnt add-hit-to-list)

RwClump ^♦picked_clump - NULL; RwClump ^♦parent_clump = NULL;

RwV3d vert; long parent_tag, clump_tag, three_d_type = 0, offset = 0, key = 0; int num_hits = -3, j; unsigned long z_max = OUL, z_min = ULONG_MAX, name_id = OUL; unsigned int ^♦ptr = mw_selected_objects;

BM_EDGE ^edge = NULL; long old_data; if (mi_show_solid) { // RenderWare picking. if (m_pick.type == rwNAPICKOBJECT) { // No clump was picked. if (add_hit_to_list) return; else clear_selection_set 0 ;

41 APPENDIX E return; else if (m_pick. type == rwPICKCLUMP) { // Some clump was picked; process it. picked_clump = m_pick. object . clump. clump; vert = m_pick. object .clump. wcpoint; RwTranslateMatrix (RwScratchMatrix () , CREAL (md_x_trans) , CREAL (md_y_trans) , CREAL (0.0) , rwREPLACE) ;

RwTransf ormMatrix (RwScratchMatrix ( ) , md_rot_matrix, rwPRECONCAT) ;

RwTrans la teMat rix ( RwScrat chMat rix ( ) , CREAL ( - m_part_centroid.X() ) , CREAL (- m_part_centroid. Y ( ) ) ,

CREAL(- m_part_centroid .7,0) , rwPRECONCAT) ;

RwMatrix4d ^♦inverse = RwCreateMatrix 0 ;

RwInvertMatrix(RwScratchMatrix() , inverse) ; RwTransformPoint (_vert, inverse) ; RwDestroyMatrix(inverse) ; parent_clump = RwGetClumpParent (picked_clump) ; parent_tag = RwGetClumpTag(parent_clump) ; clump_tag = RwGetClumpTag(picked_clump) ; if (parent_tag == 1) {

// The body of a face/bendline/forming was picked. key = (clump_tag%10000) ♦lOOOOO + clump_tag; if (m_selection_set.find(key, NULL)) { m_selection_set .remove (key, _old_data) ; // Remove clump from selection set (toggle) . if (old_data) delete (BV_SELECT_DATA ^♦) old_data; return;

} else { i f ( ! a d d _ h i t _ t o _ l i s t ) clear_selection_set () ; // Remove everything else in selection set,

BV_SELECT_DATA ^♦data = newBV_SELECT_DATA; data->key = key; data->edge = NULL; ⁽data->world_pt) = BM_POINT(vert.x, vert.y, vert.z); m_se lect ion_se t . insert ( key , ⁽long⁾data) ; // insert this key and data. return;

// Now an edge was picked. key = (parent_tag%10000) ^♦lOOOOO + clump_tag;

42 APPENDIX E if (clump_tag) {

// Safety check; we expect clump_tag to be non-zero, if (m_selection_set.find(key, NULL)) { m_selection_set.remove (key, _old_data) ; // Remove clump from selection set (toggle) . if(old_data) delete (BV_SELECT_D TA ♦)old_data; else { i f ( ! a d d _ h i t _ t o _ l i s t ) clear_selection_set 0 ; // Remove everything else in selection set . e d g e map_list_name_to_bm_edge (parent_tag, clump_tag) ;

BV_SELECT_DATA ^Λdata = new BV_SELECT_DATA; data->key = key; data->edge = edge; (data->world_pt) = BM_POINT (vert.x, vert.y, vert.z) ; m_selection_set . insert (key , (long)data); // insert this key and data.

}

else { / / OpenGL

Picking. num_hits = glRenderMode (GL_RENDER) ; // The number of hits. if (num_hits == -1) { // Selection buffer has overflowed.

AfxMessageBox("Too many entities selected; change view and try again. ") ; return;

} if (num_hits == 0) {

// No hit; only a miss. if (add_hit_to_list) return; else clear_selection_set () ; return;

}

// Process the hits : find the edge with LOWEST z value of all the edges picked. for(j=0; j < num__hits; ++j) { ptr += 1; // Skip the 1st element bee. it will always be 1, if ( (unsigned long) (^♦ptr) < (unsigned long) z_min) { z_min = (unsigned long) (^♦ptr) ;

43 APPENDIX E ptr += 2; name id = (long) (⁺ptr) ; ++ptr;

} else ptr += 3;

}

// Find the bm edge corresp. to the name and add it to the selection set. if (name_id) { //

Something was selected. if (m_selection_set . find(name_id, NULL)) { m_selection_set .remove (name_id, _old_data) ; // Remove edge from selection set (toggle) . if(old data) delete (BV SELECT DATA *)old data;

} el_e { if ( ! add_hit__to_list) clear_selection_set () ; // First, we massage the name_id into the format RenderWare picking uses. parent_tag = name_id/100000 + 10000; clump_tag = name_id%100000; edge = map_list_name_to_bm_edge (parent_tag, clump_tag) ;

BV_SELECT_DATA ^♦data - new BV_SELECT_DATA; data->key = key; data->edge = edge; (data->world_pt) = BM_POINT(0.0, 0.0, 0.0); m_selection_set.insert (name_id, (long)data) ; // insert this key and data.

} } return;

// This function maps a RenderWare parent clump and child clump pair to the

// Bend Model edge which the child clump represents. The pointer to the edge is returned.

// This function only needs to deal with "real" edges, ie. trivial bendlines and

// formings are filtered out by process_hits_records itself.

BM_EDGE ^♦CBendCADViewPart : : ap_list_name_to_bm_edge (long parent_tag, long clump_tag)

44 APPENDIX E long three_d_name, three_d_type, loop nam , edge name, count = 0, offset = 0;

BM_3D_B0DY ^♦three_d_body - NULL; BM_2D_B0DY ^♦two_d_body = NULL; BM_LOOP ^♦loop = NULL; BM_EDGE ^♦edge = NULL; BM_PART ^♦part = NULL; BM_TOPOLOGY ^♦topology = NULL;

CBendCADDoc^ pDoc = GetDocument () ; ASSERT VALID(pDoc) ; part = pDoc->get_part () ; if ( ! part) return NULL; // Note: we do not expect part to be NULL at this stage, topology = part->get_topology() ; if ( ! topology) return NULL; three_d_name = parent_tag%10000; three_d_body = topology->map_id_into_pointer (three_d_name) ; // Now we have a pointer to the face/bendline to which the picked edge belongs . three_d_type = three_d_body->get_type () ; if (three_d_type == BM_ENTITY_TYPE_FACE) offset = 30000; if (three_d_type == BM_ENTITY_TYPE_BENDLINE) offset = 60000; if (three_d_type == BM_ENTITY_TYPE_FORMING) offset = 90000; loop__name = (clump_tag - offset) /100; edge_name = (clump_tag - offset) %100; if (loop_name == 0) return NULL; // This should never be the case.

// Now have the edge_name-th edge in the loop_name-th loop of the face/bendline/forming.

// Now we have the face/bendline to which the edge belongs, if (mi_show_3d) two_d_body = three_d_body->get_3D_version() ; else two_d_body = three_d_body->get_flat 0 ; if ( ! two_d_body) return NULL; if (loop_name == 1) loop = two_d_body->get_bloop 0 ; else { count - 1; for (loop = two_d_body->get_first_hole O ; loop; loop = (BM_L00P ^♦) loop->next () ) { if ( ! loop) return NULL; ++count;

45 APPENDIX E if (count == loop_name) break;

} }

// Now we have the loop to which the edge belongs, if (loop == NULL) return NULL; // A trivial bendline was picked, count = 0; for(edge = loop->get_first_edge () ; edge; edge = (BM_EDGE ♦) edge->next () ) { if ( ! edge) break;

++count; if (count == edge_name) break;

}

// Now we have the edge itself. return edge;

1st three-d-body . if (three_d_name > (ml_num_bendlines + ml_num_f aces) ) return NULL; // The thing drawn is a forming , if ( ! loop_name) return NULL ; // The thing drawn is a trivial bendline . if (three_d_name > ml_num_bendlines) { // The three_d_body is a face . for (three_d_body = part - >get_f ace_list ( ) ; three_d_body; three_d_body = (BM_3D_B0DY ^♦) three_d_body->next () ) { if(! three_d_body) return NULL; ++count; if ((ml num bendlines + count) == three d name) break;

-lse { // The three_d_body is a bendline. for⁽three_d_body = part->get_bendline_list () ; three_d_body; t hree_d_body = ( BM 3 D BODY

^♦ ) three_d_body- >next ( ) ) { ^— if ( ! three_d_body) return NULL; ++count ; if (count == three_d name) break ;

46 APPENDIX E

} // Now we have the face/bendline to which the edge belongs . if (mi_show_3d) two_d_body = three_d_body->get_3D_version() ; else two_d_body = three_d_body->get_flat () ,- if ( ! two_d_body) return NULL; if (loop_name == 1) loop = two_d_body->get_bloop 0 ; else { count = 1; for(loop = two_d_body->get_first_hole () ; loop; loop = (BM_LOOP ♦) loop->next 0 ) { if ( ! loop) return NULL;

++count; if (count == loop_name) break;

} // Now we have the loop to which the edge belongs. count = 0; for(edge = loop->get_first_edge () ; edge; edge = (BM_EDGE ♦) edge->next () ) { if ( ! edge) break;

++count; if (count == edge_name) break;

} return edge; */

// This function returns the number of items currently in the selection set. long CBendCADViewPart::get number of selected items (void)

{ return m selection set. list size () ;

}

/^♦ This allows other classes to get the selection set. The selected items are passed back in the array selected_items__array; it is the responsibility of the user to ensure that this array is of sufficient size. An array of size num_requested^2 is exactly sufficient. The function will pass back up to num_requested items, if that many are available.

47 APPENDIX E The return value is the number of items actually passed back.

The format of the returned items in the array is as follows: selected_items_array[2^i -2] contains the OpenGL Display List id of the edge drawn, and selected_items_array[2^i -1] contains the Bend Model pointer to that edge. Eg. selected_items_array[0] , selected_items_array[1] correspond to the

1st item in the selection set, etc. all the way up to array[2^Λnum_requested -2] , array[2^num_requested -1] for the num_requested-th item in the selection set.

If the edge pointer is NULL, you need to do some further processing:

Let N be the OpenGl display list id of this edge; N is of type long.

Case A) : If N/10000 > (num_bendlines_in_part + num_faces_in_part) then the user has picked the N - (num_bendlines + num_faces) forming.

Eg: num_bendlines = 4, num_faces = 3, num_formings = 5, N/10000 = 9, edge pointer is NULL. Then user has picked the 2nd forming.

Case B) : If (N%10000) /100 == 0 (ie. the hundredths digit of N is zero) then the user has picked the center line of a trivial (unbent) bendline.

The bendline is the (N/10000) -th bendline. Eg: N/10000 = 3, edge pointer is NULL, (N%10000) /100 == 0. The user has picked the 3rd bendline, which happens to be unbent. (In case you were wondering, this part is guaranteed to have at least 3 bendlines) . */ long CBendCADViewPart : : get_selected_items ( long num_requested, long ^♦selected_items_array) if ( selected_items_array == NULL) return 0 ; l ong i = m_s e l e c t i on_s e t . d i sp l ay ( num_re que s t e d , selected_items_array) ; ^— return i ;

}

48 APPENDIX E

// This allows others to clear the selection set. void CBendCADViewPart ::clear selection set (void)

{ int size = m_selection_set . list_size () ; long ^♦array = new long [ (size<<l) + 1] ; m_selection_set .display(size, array) ; for(long i=0; i < size; ++i) { if (array[ (i<<l) + 1]) delete (BV_SELECT_DATA ^♦) array[ (i<<l)

+ 1] ;

} m_selection_set .remove () ; delete [] array; }

// This function writes a RenderWare matrix's rotation part into an array[16] , which can be used by

// OpenGL. int CBendCADViewPart: :convert_rw_matrix_to_gl_array(RwMatrix4d

♦rw matrix, double ^Λarray)

{ ^"

RwReal rw_array[4] [4] , ^♦test; test = RwGetMatrixElements (rw_matrix, rw_array) ; if ( ! test) return 0; int i , j ; for ( i=0 ; i < 4 ; ++i) { for (j =0 ; j < 4 ; ++j ) array [4 ^Λi + j ] = rw_array [i] [j ] ;

} array [15] = 1 . 0 ; array[3] = array[7] = array[11] - 0.0; return 1;

// This function allows you to set the drawing mode: show_solid =

1 for solid, 0 for wireframe.

// This function assumes that the part has not changed. void CBendCADViewPart ::set drawing mode (int show solid)

{ ^" if (show_solid) OnViewShaded() ; else OnViewWireframe () ; }

// This function allows you to set the current view of the part show_3d = 1 for 3d version,

49 APPENDIX E

// 0 for flat.

// This function assumes that the part has not changed, void CBendCADViewPart::set current view(int show 3d)

{ if(show_3d) OnShow3d( else OnShowFlat 0 ;

// This allows you to set both solid/wireframe (show_solid = 0/1) and flat/3d_version (show_3d = 0/1) .

// If flat is asked for, the picture shown is top view, zoomed-all.

// If 3d is asked for and flat is currently in view, the picture shown is 3d, isometric, zoomed-all.

// If 3d is asked for and 3d is already currently in view, the view volume is not changed.

// If part_has_changed is not zero, re-faceting is done. void CBendCADViewPart::set_drawing_view_mode(int show_solid, int show 3d, int part_has_changed)

{ int show_3d_isometric = 0; if ( ( ! mi show 3d) __ (show 3d) show 3d isometric = 1; mi_show_solid = show_solid; mi show 3d = show 3d; int i;

BM_PART ^♦part = NULL;

CBendCADDoc^ pDoc = GetDocument 0 ;

ASSERT_VALID(pDoc) ; part = pDoc->get_part () ; if(! part) { mi_part_is_up_to_date = 0; return;

} mi_part_is_up_to_date = 1; if (part_has_changed) { compute_part_centroid(part, 0, _m_flat_part_centroid) ; / / Compute 3d and flat part centroids. compute_part__centroid(part, 1, &m_3d_part_centroid) ;

50 APPENDIX E

RwRemoveClumpFromScene (m_base_clump) ; m_base_clump = m_3d_base_clump; RwAddClumpToScene (m_scene, m_base_clump) ,- else { if (part_has_changed | | ( ! m_flat_base_clump) ) { i = acet_part_rw(part, _m_flat_base_clump, mi_show_3d, md_part_color) ; if ( (! i) | | ( ! m 3d base clump) ) return;

}

RwRemoveClumpFromScene (m__base_clump) ; m_base_clump = m_flat_base_clump;

RwAddClumpToScene (m scene, m base clump) ;

} else { // OpenGl stuff. i = prepare_display_list (part, mi_show_3d, 1, 0, 1, 1,

0;

;

mi__bbox_is_up_to_date = 0; if (mi_show_3d) m_part_centroid = m_3d_part_centroid; else m_part_centroid = m_flat_part_centroid; if (part_has_changed) { ml_num_faces = part->get_number_of_faces () ; ml_num_bendlines = part->get_number_of_bendlines () ; ml_num_formings = part->get_number_of_formings 0 ; ml_last_function_chosen = RETURN_FROM_DIALOG_MODE; if (show_3d_isometric) { OnZoomAll () ; return; }

51 APPENDIX E if ( ! mi_show_3d) OnViewTop () ; return; } return;

52 APPENDIX F

nι/11111/ιιιιιnιιιuιιιιιιιιιιnιnιιιι/ιιιnιιιιιιιιιιιιιιι ι 1ι/ Example of BMAPI auto dimensioning main include file / u/ // with comments //

// //

////////////////////////////////////////////////////////////// /*

General Notes :

***************

- Auto-Dimensioning package assumes that the part is already up to date, ie. the 3D version is already computed. It takes the current 3D-version of the part and uses it.

- As of 04/02/96 Auto-Dimensioning only generates dimensions for the 3D version of the part.

- As of 04/02/96 Auto-Dimensioning assumes that the part is shown with thickness (ie. it assumes that we are not showing just one side of the sheetmetal) .

Overview :

***************

Auto-Dimension draws two kinds of things :

- flange length for every bendline

- bendline information for every bendline

Bendline information is drawn in a small info-box that points to the bendline it is associated with.

The only problem with drawing those info-boxes is whether the bendline is visible, where to draw the info-box (ie. where on the screen to draw it so that it looks good and does not interfere with other stuff, for instance, does not overlap with other things) , how to associate it with the bendline. This stuff is relatively straightforward.

Flange length of the bendline is the largest outside distance between the bendline and an adjacent flange on the side of the bendline. Note that a bendline can have at most two adjacent APPENDIX F flanges, so we have to draw flange length with respect to both adjacent flanges.

In practice, flange length is drawn (as a distance) between points that we call dimension points.

There are two types of dimension points. The first type is bendline dimension points.

Since flange length is the (largest outside) distance between a bendline and an adjacent flange, these dimension points that are on the side of the bendline are called bendline dimension points.

Their location is defined solely by the bendline.

Note that (at least in the current version of the BendModel/BendCAD) one bendline can have exactly two bendline dimension points with respect to each of the adjacent flanges.

Each of the two bendline dimension points corresponds to the two ends of a bendline.

Also note that usually these pairs of bendline dimension points with respect to both adjacent flanges match. That is - bendline dimension points with respect to adjacent flange A (there are two bendline dimension points with respect to A) and the same as bendline dimension points with respect to another adjacent flange B. That is, they are physically not the same, but they have the same values.

Bendline dimension points are calculated using intersection approximation (if the bend angle is no more that 90 degrees) or tangent approximation (if the bend angle is more than 90 degrees) .

If the bend angle is not more than 90 degrees, pairs of bendline dimension points with respect both adjacent flanges are the same, because the intersection approximation points match.

The other type of dimension points is flange dimension points. They are points on the flange adjacent to the bendline, on the side of the bendline, that are the farthest away from the bendline. Assuming the flange is correct (ie. non-empty) there is at least one flange dimension point. However, theoretically there is no upper bound on the number of flange dimension points.

Note an important point. If a flange A, adjacent to a bendline B, and also is adjacent to a bendline C at its farthest point (ie. at the flange dimension point with respect to B) , the flange dimension points of B with respect to A are really the same bendline dimension points of C with respect to A (ok, assuming that bendlines B and C are parallel, otherwise things get tricky and it is not true any more) .

Flange dimension points are computed at the corners of the flange, because dimension have to be drawn at the side of the flange. APPENDIX F

This is how the Auto-Dimension works : **************************************

It is divided into two different tasks. One is computing all the dimension points (both bendline dimension points and flange dimension points) . At this stage we compute all dimension points at which flange length could be drawn. Potentially this could be done only once and provided that the part does not change we can reuse this information later every time auto-dimensions have to be drawn. (However, the part usually does change - when the user rotates the part, then the viewer actually rotates the part, not the camera (which is usually done) . So the 3D coordinates of the part change. However, we could base our calculations on Wflat of the part which is not affected by this kind of change, but fortunately computing all dimension points is pretty fast and it seems to be ok to recompute dimension points all the time. I just want to point out that we could, if needed, save some time by precomputing these dimension points once and then later reusing them) .

Once we have computed all dimension points, we can draw dimensions. But first we have to figure out which of the dimension points are visible and which are not. When we have a wireframe picture, it is pretty easy. A dimension points is visible if it is in the view window. Checking this is very fast. If we have a solid (shaded) picture things are harder because flanges closer to the viewer usually block the view of some other flanges farther away, blocking some dimension points as well. So we need to do some kind of hidden line removal.

Once we have determined which dimension points are visible, it could be that a flange length can be drawn in a number of different ways, in which case we have to pick one, hopefully the best one. Note : to show flange length, we need a bendline dimension points and a flange dimension point. One heuristic is to pick bendline-flange-dimension points which are closest to the viewer and for which the side-ways distance is the smallest.

Once we have picked dimension points for all bendlines we can draw dimensions.

Main data structures of Auto-Dimension : ****************************************

There are three classes of objects used. BM_AUT0_DIMENSION

3 APPENDIX F

is the top-level Auto-Dimension object and is usully part of a BendCAD document along with a BM_PART.

It contains a pointer to a BM_PART for which auto-dimensions are being drawn, as well as document, view and device context pointers. It also contains high-level information like the view window size, number of bendlines, a list of BM_AD_bendline objects, etc .

In the future (today is 04/03/96) it should store information about the pen and brush properties, like line color, line width, font size, etc.

BM AD bendline

stores all auto-dimension information related to one BM_BENDLINE.

It has built-in two bendline dimension points with respect to each of the adjacent flanges.

It also has pointers to lists of flange dimension points with respect to each of the adjacent flanges.

BM AD dim corner

is a dimension point (either bendline or flange) . Dimension points can have either 1 or 3 BM_POINTs.

One important member of a dimension points is opening direction. It points in the direction in which the acutual dimension with respect to this dimension points has to be drawn. Also, this is the direction from which this dimension point is visible.

Every dimension points also has a flag indicating whether it is visible or not. If it is visible, it stores screen coodinates of all of its BM POINTS. */

#ifndef BMAPI_AUTO_DIM_HXX_INCLUDED #define BMAPI_AUTO_DIM_HXX_INCLUDED

// include some BMAPI include files

#include "POINT.HXX" #include "VECTOR.HXX" #include "LINE.HXX" #include "PLANE.HXX" #include "CYLINDER.HXX" #include "2D BODY.HXX" APPENDIX F

#include "3D_BODY.HXX" #include "BEND_OP.HXX" #include "BENDLINE.HXX" #include "PART.HXX"

// include header for mesh to determine clear regions for drawing

#ifndef AD_MESH_H

#include "AD_Mesh.h" #endif

// some BendCAD classes class CBendCADDoc ; class CBendCADViewPart ;

// Auto-Dimension classes class BM_AD_dim_corner ; class BM_AD_bendline ; class BM_AUTO_DIMENSION ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

Auto-Dimension constants.

********** ********** ********** ********** **********

********** ********** ********** ********** */

// dimension line length is 30 pixels

#define PIXELS_PER_DIMENSION_LINE 30.0

// arrow angle is 30 degrees

//#define ARROW_ANGLE 0.524

// arrow length is 15 pixels

//#define PIXELS_PER__ARROW 15.0

// this is the distance tolerance used in Auto-Dimension #define AUTODIM_DISTANCE_TOLERANCE 0.001

/*

********** ********** ********** ********** **********

********** ********** ********** **********

Objects of this type represent a dimension point in the part for which a dimension could be drawn.

*/ class BM AD dim corner { - - - APPENDIX F friend BM_AUTO_DIMENSION ; friend BM_AD_bendline ;

/* ********** ********** ********** ********** **********

********** ********** ********** **********

All data members are private.

*/ private :

// Auto-Dimension BM_AD_bendline object that owns this dimension point

//

BM_AD_bendline ^♦bend ;

// this is the 3D-body in the part that has to be visible (actually some points on this

// body have to be visible) in order for this dimension point to be drawable.

// note : this is used only when we are in solid mode.

//

// usually, for bend-dim points, this points to the bendline, and for flange-dim points

// points to the flange.

// However, when this dimension point contains only one point (point p) , some special

// considerations come into effect.

//

BM_3D_BODY ^♦body_required_visible ;

// this variable is used only when we are drawing in a solid mode.

// if TRUE, we will ignore this dimension point as if it was not visible.

//

// in solid mode we will be using this heuristic in order to save time.

// when two bendlines are using equivalent flange-dimension points for (potentially)

// displaying flange-length, we will only compute the visibility information only

// for the bendline (of these two bendlines) which has the larger idx.

//

// this is initialized to FALSE, and set to TRUE only after all other dim-point computations // have been done. APPENDIX F

// int ignore_in_solid_mode ;

// Sometimes when the point in the part for which we want to draw dimensions is a

// bendline (either inside or outside) such that the radius and thickness both are not 0,

// we have to draw two short lines (as in the picture following) that connect the

// outsides of adjacent flanges. HoweVer, sometimes (when both radius and thickness

// are 0, or the point is just an outside edge of a face) we need only one point.

// To capture these two cases, we either compute three points between which the

// short line have to be drawn, or just one point (in which case no short lines are needed) .

/ /

// _P2

/ / 1

1 \l/

/ / P bbFFFFFFFFFFFF

// 1 1 bbb F

/ / two 1 bbb F

/ / short - > i b F

/ / lines 1 b bbFFFFFFFFFFFF

// lb b

// |b b

/ / b b

// b b

/ / p FFFFFFFF

/ / F F

// F F

/ / F F

// F F

/ / F F

//

// this is the number of points that we have computed here, it is either 1 or 3.

// int num_of points ;

//

// p is always the centerpoint to which the dimension arrow should be connected to

// (actually it is what we call a 'dimension-line-extension' ) .

// if we have computed more than 1 point, pi and p2 are the side-points.

//

// IMPORTANT : pi is on the plane of the adjacent body for which the flange length is being drawn. // APPENDIX F

// Note : point p is always computed.

//

BM_P0INT p, pi, p2 ;

// this variable is used when this dimension point is a flange dimension point, and

// the edge in the flange, for which this flange dimension points was computed,

// is adjacent to a bendline. adj_bendline points to that bendline.

//

// this is used to avoid drawing some flange-length twice, for example, in this picture

//

// B2-

//

// Bl

//

// we want to draw dimensions for face A in the middle only once. That means that if

// we draw flange length for bendline Bl with respect to A, we should not draw it for

// bendline B2 (with respect to A) .

//

BM_BENDLINE ^♦adj_bendline ;

// this vector points to the direction in which the dimension should be drawn for this dim-point.

//

// IMPORTANT : this vector should be normalized.

//

BM_VECTOR opening_direction ;

//

// if this is TRUE, we can reverse the opening_direction if we need.

// sometimes the dimension point correcponds to an extreme point of an arc. in that case it does

// not matter on which side we are drawing the dimension arc (ie. we can reverse the opening direction) .

// int can_reverse_opening dir ;

//

// these variable are used only when this dimension point is actually drawn.

// pos_len is the length of the dimension-line-extension drawn in the direction of the opening vector.

// neg_len is the length of the dimension-line-extension drawn

8 APPENDIX F in the direction opposite to // the opening vector.

//

// they are initially set to 0 when the visibility of this point is computed.

//

// the purpose is that we can use them to avoid redrawing the same extension line, if several

// other dimension points need it.

//

// Note : as of 04/02/96 these points are not used. KK.

// double pos_len, neg_len ;

// +1, this dimension point is visible and the dimension arc can be drawn for this point .

// 0, this dimension point is not visible.

// -1, we don't know if this dimension points is visible.

//

// initially we set this to -1 and use it to compute the visibility information on demand

// (ie. only when we want to know if this points is visible. sometimes we don't care) .

// int visible ;

// these are the actual view window coordinates of all 3 points.

// int screen_x, screen_xl, screen_x2 ; int screen_y, screen_yl, screen_y2 ;

//

// this is the z-buffer value of the point p when rendered.

// double z_buf_value ;

// this is used to construct lists of dim-points.

//

BM_AD_dim_corner ^♦next ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

Private member-functions . Used only by Auto-Dimension. ********** ********** ********** ********** **********

********** ********** ********** **********

*/

// this function computes the visibility of this dimension point.

// it handles wireframe and solid modes separately. APPENDIX F

// this function checks strict visibility. void compute_visibility(CBendCADViewPart *vιe ) ;

// this function will draw the dim-point in the given device context .

// this is really just a debug function.

// void draw_with_dim_line (CBendCADViewPart ^♦view, CDC ^♦dc, double object_space_per_pixel) ;

// given a bendline dimension point and a list of flange dimension points, this function

// returns TRUE iff there is a way to display the flange-length-dimension using these points

// (ie. there is a pair of points that is visible) or FALSE if not.

// If it returns TRUE, it also returns the best flange dimension point ; as well as

// the distance between the bendline dimension point p and the best flange dimension point p,

// with respect to the bendline dimension point.

//

// Note that the best distance returned could be negative since it is the k-factor with respect

// to the bendline-dimension-opening-direction-vector.

// Remember : bendline-dimension-opening-direction-vector should be normalized.

//

// this function checks strict visibility.

// friend int compute_best_dim_pair(CBendCADViewPart ♦view, BM_AD_dim_corner ^♦bend_dim_point,

BM_AD_dim_corner ^♦ flange_dim_list , double object_space_per_pixel,

// OUTPUT parameters double ^♦best_distance /^♦ distance of the best flange-dim points found ^♦/,

BM_AD_dim corner ^♦♦best_flange_dim_point /♦ the best flange-dim points *J) ;

// this function draws one flange length dimension, the first argument is a bendline dimension point,

// the second is a flange dimension point.

// this function might allow cases where some dimension extension line is a little-bit

// outside the view window.

// however, endpoints of the dimension arrow have to be within the view window.

//

10 APPENDIX F friend void draw_dim_point (BM_AD_dim_corner ^♦bend_dim_point , BM_AD_dim_corner ⁺f lange_dim_point ,

CBendCADVi ewPart ^♦vi ew , CDC ^♦ dc , doub l e obj ect_space_per_pixel) ;

/*

* * * * * *** * * *** * * *** * * * * ** * ***** ***** * ** * * **** * * ****

** ******** ********** ********** **********

Public member-functions.

********** ********** ********** ********** **********

********** ********** ********** **********

*/ public :

BM_AD_dim_corner(void) ;

// this constructor is for creating a dimension point that has only one point (point p) in it .

//

BM_AD_dim_corner (BM_POINT const- p, BM_AD_bendline ^♦bend,

BM_3D_BODY ^♦body_required_visible,

BM_VECTOR constSc opening_direction, int reverse_bit, BM_AD_dim_corner ^♦♦prev, BM_BENDLINE ^♦adj_bendline) ;

// this constructor is for creating a dimension point that has 3 points in it.

//

BM_AD_dim_corner(BM_P0INT const- p, BM_POINT const- pi, BM_POINT const_ p2,

BM_AD_bendline ^♦bend, BM_3D_BODY ^♦body_required_visible, BM_VECTOR const- opening_direction, int reverse_bit, BM_AD_dim_corner ^♦♦prev, BM_BENDLINE ^♦adj^endline) ;

// to assign one dimension point to another dimension point // it copies contents exactly, not sure if it is a very useful function.

// void operator= (BM_AD_dim_corner const Sc) ;

// this function will return TRUE iff the given point is, for the purpose of drawing dimensions,

// the same as this point.

// Two dimension points points are equivalent iff the point p of one of them is on the

// line defined by the line (point p, opening direction) of the other dimension point,

// and their opening directions match.

//

// here we depend on the fact that opening vectors are

11 APPENDIX F normalized.

// it returns 2 iff both points p match, otherwise 1. int operator== (BM_AD_dim_corner const- another_dim_point) ;

} ^;

/*

********** ********** ********** ********** **********

* ******** ********** ********** **********

Objects of this class represent auto-dimensioning data associated with one bendline.

Basic AD operation with respect to a bendline is drawing adjacent flange length.

Technical notes :

- the opening direction vectors of bendline dimension points and flange dimension points are not correlated. That means that when drawing, you should check for two point having the same opening direction.

- algorithm : First we compute all bendline dimension points for all bendlines. Then we compute all flange dimension points for all bendlines. The reason is that when we compute flange dimension points, some edges in the flange and adjacent to some other bendline.

In order to create a flange dimension point, we need some vectors about the other bendline

(ie. bendlines should already be computed) .

If we compute all bendline dimensions first, we can take these vectors from the bendline dimension points of some other BM_AD_bendline structure.

*/ class BM AD bendline

{ ^{~ "} friend BM_AUTO_DI ENSION ; friend BM_AD_dim_corner ;

/*

********** ********** ********** ********** **********

12 APPENDIX F

********** ********** ********** **********

All data members are private.

*/ private :

/*

********** ********** ********** ********** **********

********** ********** ********** **********

These data members are set by the constructor.

********** ********** ********** ********** **********

********** ********** ********** **********

*/

// information stored in this BM_AD_bendline object refers to this bendline.

// ie . this dimension point is drawn to show the flange length of this bendline.

//

BM_BENDLINE ^♦bend ;

// this object belongs to this Auto-Dimension object.

//

BM_AUTO_DIMENSION ^♦parent ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

These data members are computed when the bendline dimension point data structures are computed - in BM_AD_bendline: :compute_BM_AD_bendline ( ... ) . ********** ********** ********** ********** ********** ********** ********** ********** **********

*/

// if FALSE, this data for this object has not been yet computed, or the computation failed.

// this means that this object will always be ignored, no matter what.

// this can be set to TRUE only by BM_AD_bendline: :compute_BM_AD_bendline (...) .

// int data_ alid ;

// number of adjacent bodies for which we have to draw dimensions, either 0, 1 or 2. // int num_of_adjacent__bodies ;

// every bendline is adjacent to at most two bodies.

13 APPENDIX F

// either of these can be NULL, you should always check for that.

// adj_bodyl is the adjacent body that has the smaller idx. // adj_body2 is the adjacent body that has the larger idx.

//

BM_3D_BODY ^♦adj Dodyl ;

BM_3D_BODY ^♦adj_body2 ;

// this is the flange length with respect to each of the adjacent bodies

// double fll ; double fl2 ;

/* These parameters are computed some time durmg the process and are needed later, so for performance reasons we store them so that we don't have to recompute them later.

Each of them is a vector pointing towards the adjacent body (from the bendline) and is on a plane tangent to end end-point of the bend-arc touching the adjacent body.

*/

BM_VECTOR side_vectorl ; BM_VECTOR side_vector2 ;

/*

These are some of the main data structures. For each adjacent body, we have two points, one for each endpoint (ie. left and right) of the bendline.

These points specify the points at which dimensions for flange length can be drawn at the bendline (ie. on the side of the bendline) .

Remark : currently (04/02/96) we consider exactly two possible points for drawing dimensions at every bendline. These two points are two extreme points of the bendline in the direction of the bendline centerline.

Note : these are called bendline dimension points of the bendline.

Note : these points are always computed, even if adj_bodyl and adj_body2 are NULL. */

BM_AD_dim_corner bend_bodyl_point_left, bend_bodyl__j)oint_right

14 APPENDIX F

BM_AD_dim_corner bend_body2_point_left, bend_body2_point_right ;

// these variables indicate whether the left (right) bendline dimension points

// with respect to adj -bodyl and adj-body2 are respectively equivalent.

// these points should be computed right after bend_bodyl_point_left, bend_bodyl_point_right,

// bend_body2_point_left, bend_body2_point_right are computed.

//

// if 0, they are not equivalent.

// if 2, they are equivalent, and points p match. // if 1, they are equivalent, but points p don't match (ie. on the same line, but some distance apart) .

//

// for us, equivalence value 2 is the only really useful one, because in that case we can

// trust that when one is visible, the other is too, without checking, in that case they have the

// same screen coordinates as well.

//

// note the same means not the same pointers, but the contents are equivalent.

// int bodyl2_left_the_same ; int bodyl2_right_the_same ;

/*

Here are two lists of dimension points. One with respect to adj -bodyl, the other with respect to adj-body2. Both lists contain points that are flange-length dimensions points.

Note : these are called flange dimension points of the bendline.

*/

BM_AD_dim_corner ^♦bodyl_points ; BM_AD_dim_corner ^♦body2_points ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

These data members are used when dimensions are being drawn. ********** ********** ********** ********** ********** ********** ********** ********** **********

*/

// if this is TRUE, this bendline structure should be ignored for the purpose of drawing dimensions.

15 APPENDIX F

// this is usually used when we have a simultaneous bendline. in that case we have to draw

// dimensions for only one of them, the rest should be ignored.

// int ignore ;

// if TRUE, the adjacent body already has its flange length drawn.

// if the value is 2, this bendline drew the flange dimension with respect to the adj body.

// if the value is 1, some other bendline drew the flange dimension with respect to the adj body.

// int bodyl_drawn ; int body2_drawn ;

// when we draw flange length for adjacent bodies, we will memorize the point at which the

// dimensions were drawn.

// actually the dimensions are always drawn between two points - one at the bendline,

// the other at the farthest point on the adjacent body (ie. bendline dimension point and

// flange dimension point) .

//

// the reason we need to memorize these points is that sometimes the flange-length-dimension

// with respect to one bendline is also the flange-length dimension with respect to the

// other bendline. in that case we want to avoid drawing the same thing twice.

//

// the first two points are bendline dimension points.

// other two are flange dimension points.

//

// reason #2 : when we draw bendline dimensions with respect to adj-bodyl and adj-body2,

// it depends whether the bend angle is acute or not. if the bend angle is not more than

// 90 degrees, then most likely we will be able to use the same dimension-point at the bendline

// to draw both flange length's. However, if the bend angle is more than 90 degrees, we will

// be using the tangent-dimension-point at the bendline and these points are different for

// each adjacent body. In general, we want to know, before we draw a dimension-point at a

// bendline, if we can use any of the dimension-points already

16 APPENDIX F drawn at this bendline.

//

// note these points do not have to part of this AD_bendline structure.

//

BM_AD_dim_corner ♦drawn_bend_dim_pointl ;

BM_AD_dim_corner ♦drawn_bend_dim_point2 ;

BM_AD_dim_corner ^♦drawn_flange_dim_pointl ;

BM_AD_dim_corner ^Λdrawn_flange_dim_point2 ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

Private member-functions .

********** ********** ********** ********** **********

********** ********** ********** **********

*/

// This function computes bendline dimension points of this

BM_AD_bendline structure.

// Note that bend-dimension points will always be computed, even if adj_bodyl and adj_body2

// are NULL.

// int compute_BM_AD_bendline(void) ;

// This function computes flange dimensions points of this BM_AD bendline structure

7/ with respect to one adjacent body.

// int compute_BM_AD_flange_points (BM_3D_BODY ^♦adj_body) ;

// This function computes the visibility of all bendline dimension points of the bendline.

// note the the visibility of bend-dimension points with respect to adj-bodyl is always

// computed, even if adj_bodyl pointer is NULL.

// This is needed for drawing bendline info-boxes.

// void compute_bendline_dim_points_visibility(void) ;

// this function computes which dimension points are best to draw flange length for this bendline.

//

// this function does it strictly for one bendline.

// void compute_dim_points_to_display_indiv(

BM_AD_bendline ^♦best_AD_bendline [2] /^♦ best AD_bendline structure so far ♦/,

BM_AD_dim_corner ^♦bodies_bend_flange [6] /♦ 0,1 is bend-dim points; 3,4 is flange-dim points; , 5 is flange lists ♦/,

17 APPENDIX F double distance and z [4] /♦ 0,1 is distance; 3,4 is z-buf

7)

// this function does it for a given bendline, taking into account simultaneous bendlines.

// this function uses the previous function.

// void compute_dim_points_to_display (void) ;

// this function draws flange length dimensions for this bendline .

/ / i t us e s the inf ormat ion comput e d by compute_dim_points_to_display ( ) function .

// void draw_dimension_points (void) ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

Public member-functions.

********** ********** ********** ********** **********

********** ********** ********** ********** */ public :

// some trivial constructors, destructors.

BM_AD_bendline(BM_AUTO_DIMENSION ^♦owner, BM_BENDLINE ♦bendline) ;

~BM AD bendline (void) ;

} ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

This object can be used to create automatic dimensions for a part.

*/ class BM AUTO DIMENSION

{ ^{" "} friend BM_AD_dim_corner ; friend BM_AD_bendline ,-

/*

********** ********** ********** ********** **********

********** ********** ********** **********

18 APPENDIX F

All data members are private.

********** ********** ********** ********** **********

********** ********** ********** ********** */ private :

// this is the document to which this auto-dim object is attached to.

// also, this is where we get the part.

//

CBendCADDoc ^♦doc ;

// view in which we draw the dimensions.

// usually the view attached to the document.

//

CBendCADViewPart ^view ; int view_size_x, view_size_y ;

// device context in which we draw the dimensions.

//

CDC ⁺dc ;

// this is the part for which we will create dimensions.

//

BM_PART ^♦part ;

// is TRUE iff 3D view is being displayed, FALSE if flat.

// int view_type ;

// half of the metal thickness

// double half_metal__thickness ;

// if TRUE, we need to recompute auto dimension bendline and point data structures

// int dirty ;

// number of bendlines in the part

// int num_of_bendlines ;

// this is an array of bendlines for which flange length has to be drawn.

//

BM_AD_bendline ^♦♦bends ;

// this is the distance in object space that gets mapped into

19 APPENDIX F one pixel on the screen.

// we use it to draw lines of fixed pixel length (in that case object-space length varies) .

// double object_space_per_pixel ;

// if this variable is TRUE, it will set the flag which will trigger

// auto-dimensions to be drawn to FALSE in the view class, once dimensions are drawn.

// however, once it is done, this variable will be set to

FALSE automatically.

// int reset_view_to_false_once_drawn ;

// every time we construct new dim-point data structures, we set this to FALSE.

// when points which can be ignored in solid mode have been computed, we set it to TRUE.

// int points_to_ignore_in_solid_computed ;

// pointer to mesh for determining clear areas of the screen to draw dimension info

//

CAutoDimRgnMesh ♦pMesh;

// region to keep track of dirty screen areas

//

CRgn^Λ pDirtyRegion; public:

// Booleans for the show state of dimension info

//

BOOL m_bShowFlangeDim;

BOOL m_bShowBendDim;

BOOL m_bShowPartDim;

// COLORREFs for drawing dimensions

//

COLORREF m_crFlange;

COLORREF m_crBend;

COLORREF m_crPart;

// pointers to fonts

//

CFont^ m_pFlangeFont;

CFont^ m_pBendFont;

CFont^ m pPartFont;

20 APPENDIX F

// line and arrow parameters

// int m_nLineStyle; int m_nLineWeight; // in pixels int m nArrowStyle; int m_nArrowAngle; // in degrees int m_nArrowLength; // in pixels

/ *

********** ********** ********** ********** **********

********** ********** ********** **********

Private member-functions.

********** ********** ********** ********** **********

********** ********** ********** ********** */ private :

// This function is used to propagate the effects of drawing a flange length for a given bendline,

// to account for the fact that the flange dimension point is adjacent to the given adj-body.

// it returns TRUE iff the adj -bend will be marked as drawn as well with respect to the given flange.

// int compute_effects_flange_length_drawn(BM_BENDLINE ♦bendline,

BM_3D_BODY ^♦adj_body /^♦ flange length with respect to this body are being drawn ^♦/,

BM_AD_dim_corner ^♦bend_dim_point /^♦ bend dim point for bendline ^♦/,

BM_AD_dim_corner *flange_dim_point /^♦ flange dim points for bendline ^♦/,

BM_BENDLINE ⁺adj_bend /^♦ flange dim points is adjacent to this bendline ^♦/) ;

// this function returns a pointer to the BM_AD_bendline structure that contains data

// about the given bendline.

//

BM_AD_bendline ^♦get_AD_bend_structure (BM_BENDLINE ♦bendline)

II this function will return TRUE iff adj_bend has a flange-dimension point with respect to the

// given flange, that is adjacent to the given bendline.

//

// a handy function.

// int check_two_bendlines_adjacent_to_same_flange (BM_BENDLINE

21 APPENDIX F

♦bendline,

BM_3D_BODY ♦flange, BM_BENDLINE ^♦adj Dend) ;

// just a debug function

// void test_draw(void) ;

// Check if we can ignore some points in solid mode.

//

// in solid mode we will be using this heuristic in order to save time.

// for the bendline (of these two bendlines) which has the higher idx.

// void compute_points_to_ignore_in_solid(void) ;

// this function draws a box with bendline data for every bendline.

// void draw_bendline_data (CDC ^♦pdc) ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

Public member-functions.

********** ********** ********** ********** **********

********** ********** ********** ********** */ public :

BM_AUTO_DIMENSION(CBendCADDoc ^♦BendCAD_doc) ; -BM_AUTO_DIMENSION(void) ;

// this function returns a (p->p2) vector for a given bendline with respect to a given adj-body.

// It also computes a line between the p2 points of the bend dimension points of the adjacent bendline

// with respect to this flange.

//

BM_VECT0R const- compute_bend_direction vector (BM BENDLINE ^♦bendline, BM_3D_B0DY ^♦adjjsody, ^~

BM_LINE ^♦p2_line) ;

// destroy the contents of this Auto-Dimensions object,

22 APPENDIX F

// void delete_contents (void) ;

// To set a new part. This updates the view class pointer as well .

// void set_part (BM_PART *ne j)art) ;

// to force AD to recompute the dimension points.

// inline void invalidate (void) { dirty = 1 ; }

// get current view type when points were computed

// inline int get_view_type (void) const { return view_type ; }

// this function will build auto dimension bendline and point data structures.

//

// This is the main function for computing dimension points.

// int compute_auto_dimension_data (void) ;

// draw flange dimension data

// void draw_dim_point (BM_AD_dim_corner ^♦bend_dim_point, BM_AD_dim_corner ^♦flange_dim_point,

CBendCADViewPart ^♦view, CDC ♦dc, double object_space_per_pixel) ;

// to draw part dimensions in this device context

// int draw_auto_dimensions (CDC ^♦target_dc) ;

// Used with draw_auto_dimensions to construct a geometric pen // with user-defined styles. This is necessary to draw patterned thick lines.

//

CPen^ create_auto_dim_pen(CPen^# pPen, int nLineStyle, int nLineWeight, COLORREF cr) ;

// this will enable us to draw part dimension only once

// inline void disable_view_dimensions_flag_next_time (void) { reset view to false once drawn = 1 ; } } ; ^{~ " " " ""}

#endif // BMAPI AUTO DIM HXX INCLUDED

23 APPENDIX G u i mi /mnu it mum tin nun i iuuiiiiiiiiiiiii ι 1ι/ Example of auto dimension drawing functions. These '/J/, // functions assume that the dimension points are // // already computed. //

'/l 11111111 It 1111111111111 II 111111111 It 11111 II HI 111111111 II 111

I*

Functions to compute dimension points are in AD_P0INTS.CPP.

First we need to detect which dimension points are visible, which are not.

Then we have to decide how to draw flange length dimensions.

*/

#include "stdafx.h"

#include "BendCAD.h"

#include "BendCADDoc.h"

#include "BendCADViewPart.h"

#include "BCAD_CTXT.HXX"

#include "AD_Arrow.h"

#include "AUTODIM.HXX"

/*

********** ********** ********** ********** **********

********** ********** ********** **********

BM_AD_dim_corner class stuff .

*/

/*

This function computes the visibility of a dimension point . */ void BM_AD_dim_corner: :compute_visibility(CBendCADViewPart ^♦view) int i, j, k ,- double z_buf ;

// by default, the points is not visible visible = 0 ; if (num_of_points < l) return ;

// note : when in wireframe mode, we don't have to check if

1 APPENDIX G points pi and p2 actually

// map into the same bendline/flange points, in solid mode we need to check.

// check if we have solid or wireframe mode. if (view->get_drawing_mode() ) goto solid_picture ;

/*

Wireframe, don't check if the 3D-body to which this dimension points is actually in the foreground.

In wireframe mode we assume that everything in the view window is visible.

*/

// get visibility data i = view->map_point_to_closest_edge(p, _screen_x, Scscreen_y, _z_buf_value, NULL) ; if (! i) return ; if (num_of_points > 1) { i = view->map_point_to_closest_edge(pi, _screen_xl, Scscreen_yl, _z_buf, NULL) ; if ( ! i) return ; i = view->map_point_to_closest_edge(p2, _screen_x2, _screen_y2, _z_buf, NULL) ; if ( ! i) return ; }

// process visibility data pos_len = neg_len = 0.0 ; visible = 1 ; return ;

/*

Part shown in solid mode.

*/ solid__picture :

// here we keep pointers to 3D-bodies that have to be visible.

BM_3D_B0DY ♦visiblejoodies [4] ; visible_bodies [0] = body_required_visible ; if (body_required_visible->is(BM_ENTITY_TYPE_BENDLINE) ) { visible_bodies [1] = ( ( ( BM_BENDL INE ^♦) body_required_visible) ->get_adj_list () ) ->f ind_lowest_id_adj_body ( ) visible_bodies [2 ] = ( ( ( BM_BENDL INE ^♦) body_required_visible) ->get_adj_list 0 ) ->f ind_second_lowest_id_ad j_body() ,- visible bodies [3] = NULL ;

} else visible bodies [1] = NULL ; APPENDIX G

// handle cases when there is just one point differently from when there are 3 points.

// if there is only 1 point, we have a complete paper-part - thickness and radius both 0.

// in that case there is really no bendline. we have to check any of the adjacent bodies. if (num_of_points < 2) {

// if this is a bend-dimension point, the body_required__visible pointer should point to the

// adjacent body with respect to this dimension point is computed. i = view->is_3d_body_near_point (p, visible_bodies, _i, _screen_x, _screen_y, _z_buf_value) ; if ( ! i) return ; else {

/*

SOLID HEURISTIC : this here is a new version of handling solid mode. first we check if all three points p, pi and p2 are in the view window. if true, we compute the centerpoint of the side profile of the bend arc, and check if the bendline at this point is actually visible.

*/ i = view->map_point_to_closest_edge (p, _screen_x, _screen_y, Scz_buf_value, NULL) ; if ( ! i) return ;

/*

SOLID HEURISTIC : if the opening direction cannot be reversed, check if it is not pointing away if ( ! can_reverse_opening_dir) {

BM_P0INT p_opening_direction(p + opening_direction) t__edge (p_opening_direction, + 0.001)) return ; // here pointing away

// check if points pi and p2 are in the view window as well. APPENDIX G i - view->map_point_to_closest_edge (pi, _screen__xl, _screen_yl, _z_buf, NULL) ; if ( ! i) return ; i = view->map_point_to_closest_edg (p , _screen_x2, _screen_y2, _z_buf, NULL) ; if ( ! i) return ;

BM_VECTOR v__ppl(p,pl) ;

BM_VECTOR v_pp2(p,p2) ;

BM_POINT p_center ; double angle_alpha = (v_ppl ^A v_pp2)/2.0 ,^• if (angle_alpha < PI_over_2) {

BM_VECTOR v_plp2(0.5^Λ (p2 - pi)) ;

BM_POINT temp p(pi + v_plp2) ;

BM_VECTOR v_tip(p,temp_ ) ; double z - (v_ppl.Len() ) /cos (angle_alpha) ; double y = z+sin(angle_alpha) ; v _ t i p . s e t _ l e n g t h ( ( z - y ) + (bend->parent) ->half_metal_thickness) ; center - p + v_tip .; - else p_center = p ; ody_near_point (p_center, visible_bodies ,

;

/*

This the old version of handling solid mode. Here we check if the point p is in the view window, and if points pi and p2 have the underlying 3D-body visible.

// if we have 3 points, the point p is not on any 3D-body. threfore we only check if it is

// in the view window. view->is_3d_body_near_point (p, NULL, _i, Scscreen_x, _screen_y, 5cZ_buf_val_e) ; if ( ! i) return ;

// points pi and p2 have to be on the 3D-body. i = view->is_3d__body_near_point (pi, visible_bodies, Scvisibility_count , _screen_xl, _screen_yl, _z_buf) ; if ( ! i) return ; i = view->is_3d__body _near_point (p2, visible_bodies, _visibility_count , _screen_x2, _screen_y2 , Scz_buf) ; if (1 i) return ,- */

}

// process visibility data

4 APPENDIX G pos_len = neg_len = 0.0 visible = 1 ;

/* this function will draw the dim-point in the given device context . this is really just a debug function.

*/ void BM_AD_dim_corne : :draw_with_dim_line (CBendCADViewPart ♦view, CDC ♦dc, double object_space_per_pixel)

{ int i, j ; double x ; static POINT points [3] ;

BOOL res ;

// if this point is not known to be visible, return. if (visible < 1) return ; if (num_of_points < 1) return ; points [1] . x = screen_x ; points [1] . y = screen_y ;

// draw dimension line

BM_POINT p_dim_line (p + (PIXELS_PER_DIMENSION_LINE ^♦ obj ect_space_per_j?ixel) ^♦opening_direction) ; view- >map_point_to_closest_edge (p_dim_line , _i , _j , SX, NULL)

; points [2] .x = i ; points [2] .y = j ,- if (can_reverse_opening_dir) { p_dim_line p + (-PIXELS_PER_DIMENSION_LINE ^♦ object_space_per__pixel) ^♦opening_direction ; view->map_point_to_closest_edge (p_dim_line, Sci, Scj , SX, NULL) ; points [1] .x - i ; points [1] .y = j ; res = dc->Polyline(points + 1, 2) ; points [1] .x = screen_x ; points [1] -y = screen_y ; else res = dc->Polyline (points + 1, 2) ; if (num_of_points < 2) return ; points [0] . x = screen_xl ; points [0] . y = screen_yl ; APPENDIX G points [2] .x = screen_x2 ; points [2] . y = screen_y2 ; res = dc->Polyline (points, 3)

This function draws one flange length dimension.

The first argument is a bendline dimension point, the second is a flange dimension point. */ void BM_AUTO_DIMENSION : : draw_dim_point (BM_AD_dim_corner

♦bend_dim_point , BM_AD_dim_corner ^♦f lange_dim_point ,

CBendCADViewPart ^♦view, CDC ^♦dc, double object space_per_pixel)

{ int i, j ; double x, y, z, dotp, half_dim_.line ^■

BOOL res ;

// points with screen coordinates. Windows type, static POINT points [3] ;

// these variable store the coordinates of the dim- line-extension endpoints double bend_end_x, flange_end_x ;

BM_POINT bend_end_p, flange_end_p ;

BM_VECTOR end_v (bend_dim_point->opening_direction) ;

// these variables store arrow head endpoints BM_POINT bend_arrow_ , f lange_arrow_p ;

// compute the dimension line extension

BM_VECTOR v( (f lange_dim_point->p) - (bend_dim_point->p) ) ; x = v % (bend_dim__point->opening_direction) ; if (fabs(x) > AUTODIM_DISTANCE_TOLERANCE) { // we have to extend at least one dimension line half_dim_line = (PIXELS_PER_DIMENSION_LINE ^♦ object_space_per_pixel) /2.0 ;

// here we will first compute extension lines, then the arrow, and then the dimension points themselves if (x < 0.0) {

// here we assume that for bendline, we cannot reverse the opening direction.

// it must be that the flange-dim points has a valid opening direction in the same direction

// as the bend-dim point. // APPENDIX G

// arrow point is to the right of the bend-dim point . bend_arrow_p = (bend_dim_point - >p ) + half_dim_line^end_v ; if (! view->map_point_to_closest_edge (bend_arrow_p, Sci, Scj, Scy, NULL)) return ; points [0] . x = i ; points [0] .y - j ; f lange_arrow_p = (f lange_dim_point->p) + (-x + half dim_line) *end_v ; i f ( ! view-_>map_point_to_closest_edge (flange_arrow_p, &i , Scj, Scy, NULL)) return ; points [1] .x = i ; points [1] .y = j ;

APPENDIX G bend_end_x = z + half_dim_line ; flange_end_x = -x - half_dim_line ; bend_end_p = bend_dim_point - >p bend_end_x^Aend_v ; flange_end_p = f lange_dim_point->p f lange_end_x^end_v goto begin to draw } ^{" ~} try_other_direction : i f ( d o t p > 0 . 0 J | f lange_dim_point- >can_reverse_opening_dir) {

// arrow point is to the right of the flange-dim point. bend_arrow_p = (bend_dim_point->p) + (x + half dim_line) ^♦end_v ; i f ( ! view-_>map_point_to_closest_edge (bend_arrow_p, Sci, _j , Scy, NULL)) return ; points [0] .x = i ; points [0] -y = j ; f lange_arrow_p = (f lange_dim_jpoint->p) + half dim line^#end_v ; i f ( ! view->map_point_to_closest_edge (f lange_arrow_p, _i, Scj, Scy, NULL)) return ; points [1] .x = i ; points [1] .y = j ; half_dim_line ^♦= 2.0 ; bend_end_x = x + half_dim_line ; flange_end_x = half_dim_line ; bend_end_p - bend_dim_point - >p + bend_end_x^Λ end_v flange_end_p = flange_dim_jpoint->p flange_end_x^end_v ;

} else return ; // failure, cannot draw because opening direction don't match or not visible.

else { // no need to extend dimension lines

// check the direction of opening vector if (end_v % (flange_dim_point->opening_direction) < 0.0) if (! flange_dim_point->can_reverse_opening_dir) end v.reverse() ;

} bend_end_x - flange_end_x = (PIXELS_PER_DIMENSION_LINE ^♦ APPENDIX G object_space_per_pixel) ; bend_end_p = bend_dim_point->p + bend_end_x^Λend_v ; flange_end_p = f lange_dim_jpoint->p + f lange_end_x^Aend_v bend_arrow_p = bend_dim_point->p + (bend_end_x/2.0) ^♦end v

; f lange_arrow_p = f 1 ange_d im_po i n t - >p +

( f 1 ange_end_x/ 2.0 ) ♦ end_v ;

// compute arrow endpoints if (! view->map_point_to_closest_edge (bend_arrow_p, Sci, Scj , Scy, NULL) ) return ; points [0] .x = i ; points [0] .y = j ; if ( ! view->map_point_to_closest_edge (f lange_arrow_p, Sci, &j , _y, NULL) ) return ; points [1] .x = i ; points [1] .y = j ;

begin_to_draw :

/*

Draw arrow.

*/

// draw arrow line, points[0] is the bend-dim-points side of the arrow. res = dc->Polyline(points, 2) ;

// draw arrows

BM_VECTOR v_arrow(flange_arrow_p - bend_arrow_p) ; x = v_arrow.Len() ; // this is the real flange length

// change to solid pen for arrow heads

CPen* pArrowPen = new C P e n ( P S _ S O L I D ,

BCAD_context . m_nFlangeLineWeight , BCAD_context . _crFlange ) ; CPen^ pOldPen = dc->SelectObject (pArrowPen) ;

// compute arrow head at bend side CAutoDimArrow arrowBend(

CPoint(points [0] .x, oints[0] .y) ,

CPoint(points[1] .x,points[1] .y) ,

BCAD_context.m_nArrowAngle,

BCAD_context.m_nArrowLength ) ; arrowBend.DrawHead(dc) ;

// compute arrow head at flange side CAutoDimArrow arrowFlange( APPENDIX G

CPoint (points [1] .x,points [1] .y) , CPoint (points [0] .x,points [0] .y) , BCAD_context .m_nArrowAngle, BCAD_context .m_nArrowLength ) ; arrowFlange.DrawHead(dc) ; dc->SelectObject (pOldPen) ; delete pArrowPen;

// write flange length on the screen int center_x = (points [0] .x + points [1] .x) >> 1 ; int center_y = (points [0] .y + points [1] .y) >> 1 ; int font_height = 21 ; double dx = (double) (points [l] .x - points [0] .x) ,- double dy = (double) (points [1] .y - points [0] .y) ; if (fabs(dx) > AUTODIM_DISTANCE_TOLERANCE) ( if (dy/dx > 0.0) center y -= font_height ;

} center_x += 3 ; // move the beginning to the right so that it won't overlap the arrow line.

CString str; str.Format ("%6.4f" ,x) ;

CSize sizeTextExtent = dc->GetTextExtent (str) ; if (pDirtyRegion != NULL)

{

CPoint pntText (center_x, center_y) ;

CRect rectText (pntText, sizeTextExtent) ; if ( !pDirtyRegion->RectInRegion(_rectText) ) dc->TextOut (center_x, center_y, str) ;

CRgn rgn; rgn.CreateRectRgn(center_x, center_y, center__x + sizeTextExtent.ex, center_y + sizeTextExtent.cy) ; pDirtyRegion->CombineRgn(pDirtyRegion, _rgn, RGN OR) ;

} else

{ pDirtyRegion = new CRgn() ; pDirtyRegion->CreateRectRgn(center_x, center_y, center_x + sizeTextExtent .ex, center_y + sizeTextExtent.cy) ; dc->TextOut (center x, center y, str) ; }

/* if necessary draw pl,p2 for bend-dim-points

*/

10 APPENDIX G points [1] .x = bend_dim_point->screen_x ; points [1] .y = bend_dim_point->screen_y ; if (bend_dim_point->num_of_points > 1) { points [0] .x = bend_dim_point->screen_xl points [0] .y = bend_dim_point->screen_yl points [2] .x = bend_dim_j?oint->screen_x2 points [2] .y = bend_dim_point->screen_y2 res = dc->Polyline (points, 3) ;

} // draw dimension line extension view-_>map_point_to_closest_edge (bend_end_ , Sci, Scj , _x, NULL) points [2] .x = i ; points [2] .y = j ; res = dc->Polyline (points + 1, 2) ;

if necessary draw pl,p2 for flange-dim-points points [1] .x = flange_dim_point-.>screen_x ; points [1] .y = flange_dim_j?oint->screen_y ; if (flange_dim__point->num_of_points > 1) { points [0] .x = flange_dim_point->screen_xl points [0] .y = flange_dim_point->screen_yl points [2] .x = flange_dim_point->screen_x2 points [2] .y = flange_dim_point->screen_y2 res = dc->Polyline (points, 3) ;

} // draw dimension line extension view->map_point_to_closest__edge (flange_end_p, Sci , Scj SLX ,

NULL) points [2] .x = i ; points [2] .y = j ; res = dc->Polyline (points + 1, 2)

/* given a bendline dimension point and a list of flange dimension points, this function returns TRUE iff there is a way to display flange-length-dimension using these points

(ie. everything is visible) or FALSE if not.

If it returns TRUE, it also returns the best flange dimension point ; as well as the distance between the bendline dimension point p and the best flange dimension point p, with respect to the bendline dimension point.

Note that the best distance returned could be negative since

11 APPENDIX G it is the k-factor with respect to the bendline-dimension-opening-direction-vector.

Remember : bendline-dimension-opening-direction-vector should be normalized.

Note : right now this function assumes that the dimension-extension-lines are drawn in the following way : one dimension-extension-line is unit-length, and the other is as long as necessary to reach the first one. Ie. this function draws the dimension arrow rigth next to one of the dimension points. Ie. it does not try to draw the dimension arrow at some point in the middle of the line between the two dimension points.

Note : in this function we will not directly check if dim points are visible (we will call another function to do that) . we will check that some dimension-extension-lines be within the viewing window.

Note : in this function we check points where dimension arrows would go. We require that they be strictly within the viewing area.

Note : in this function we really use approximated locations, ie. as a heuristic. */ int compute_best_dim_pair(CBendCADViewPart ^♦view, BM_AD_dim_corner ♦bend_dim_jpoint,

BM_AD_di m_c o rne r ^♦ f 1 ang e_di m_ l i s t , double object space_per_pixel, ll OUTPUT parameters double ♦best_distance /^♦ distance of the best flange-dim points found ^♦/,

BM_AD_dim_corner ^♦♦best_flange_dim_point /^♦ the best flange-dim points ^♦/)

{

BM_VECTOR v ;

BM_POINT dim_arrow_point ; int i, j , k ; double x, y, dotp ;

// if the bendline dimension point is not known to be visible, return FALSE.

// otherwise we might be wasting our time, if (bend_dim_ oint->visible < 1) return 0 ;

// scan the list of flange dimension points BM_AD_dim_corner ^♦temp_point, ^♦best_fdp = NULL ;

12 APPENDIX G double best_distance_so_far ; for (temp_point = f lange_dim_list ; temp_point ,- temp_point = temp_point->next) { if (! temp_j_?oint->visible) continue ;

/*

SOLID HEURISTIC : in solid mode we will be using this heuristic in order to save time. when two bendlines are using equivalent flange-dimension points for (potentially) displaying flange-length, we will only compute the visibility information only for the bendline (of these two bendlines) which has the higher idx.

*/ i f ( v i e w - > g e t _d r a w i n g_m o d e ( ) _ _ temp_point->ignore_in_solid_mode) continue ;

// compute a vector from bend-dim-p to flange-dim-p v = (temp_point->p) - (bend_dim_point->p) ; x = v % (bend_dim_point->opening_direction) ;

// check that opening directions are valid.

// idea : if both vectors points in opposite directions are we cannot reverse them,

// we cannot use these two dimension points to draw dimensions. dotp = (bend_dim_point->opening_direction) % (temp_point->opening_direction) ; if (dotp < 0.0 cSc ! temp_point->can_reverse_opening_dir ScSc. ! bend_dim_point->can_reverse__opening_dir) continue ;

// check that the point where the dimension arrow would go is in the view area

/ / n o t e : w e w i l l u s e half-of-the-minimal-extension-length as the unit separating // zero and non-zero length extension lines. if (fabs(x) > AUTODIM_DISTANCE_TOLERANCE) { k = 0 ; // if k is TRUE at the end, we can draw the dimension arrow point if (x < 0.0) {

// note : add space for a line segment where we will draw the arrow

//

// check if the directions agree. // we check this first, because it would give us a smaller distance. if (dotp > 0.0) { dim_arrow_point = (temp_point->p) +

13 APPENDIX G

(-x) ^♦ (bend_dim_point->opening_direction) ; i f

(view->map_point_to_closest_edge (dim_arrow__point, _i, Scj, Scy, NULL) ) k = 1 ;

} // check that the extension of the bend-dim-line is visible.

// in that case the opening direction of the flange-dime-line does not matter. i f ( ! k - Sc bend_dim_point->can_reverse_opening_dir) { dim_arrow_point = (bend_dim_point->p) + (x) ♦ (bend_dim__point->opening_direction) ; i f

(view->map_point_to_closest_edge (dim_arrow_point, _i, Scj, Scy, NULL) ) k = 1 ;

else { // x > 0.0

// note -. add space for a line segment where we will draw the arrow.

//

// check if the extension line is visible.

// we check this first, because it would give us a smaller distance. i f ( d o t p < 0 . 0 | | temp_point-_>can_reverse_opening_dir) { dim_arrow_point = (temp_point->p) + (-x) ♦ (bend_dim_point->opening_direction) ; i f

(view->map_point_to_closest_edge (diτn_arrow_point, Sci, Scj, _y, NULL) ) k = 1 ;

} // otherwise check if the directions agree. if Ok) { dim_arrow_point = (bend_dim_point->p) + (x) ♦ (bend_dim__point->opening_direction) ; i f

(view->map_jpoint_to_closest_edge(dim_arrow_jpoint, _i, Scj , Scy, NULL) ) k = 1 ;

if (!k) continue ; // cannot draw the dimension arrow point

}

// check if visibility information is already computed. // if not, compute it. if flange dimension points not visible, continue.

14 APPENDIX G if (temp_point->visible < 0) { temp_point->compute_visibility(view) ; if (! temp_point->visible) continue ;

// check if this new distance is the best distance if (best_fdp) { if (fabs(best_distance_so_far) > fabs(x)) { best_fdp = temp_point ; best_distance_so_far = x ;

else { best_fdp = temp__point ; best distance so far = x ;

} ^"

// check if we are done.

// if we have a point with a distance 0, we are done since we cannot find a better point. if (best_fdp Sc_ fabs (best_distance_so_far) <= AUTODIM DISTANCE TOLERANCE) break ; ^" } if (best_fdp) {

♦best_distance = best_distance_so_far ; ♦best_flange_dim_point = best_fdp ; return 1 ; } return 0 ;

}

/ *

********** ********** ********** ********** **********

********** ********** ********** **********

BM_AD_bendline class stuff.

********** ********** ********** ********** **********

********** ********** ********** ********** */

/*

This function computes the visibility of all bendline dimension points of the bendline.

Note : in order to save some time, we will check if bodyl,2 left (right) points are equivalent.

*/

15 APPENDIX G void BM_AD_bendline: :compute_bendline_dim_points_visibility(void) CBendCADViewPart ^♦view = parent->view ;

// here we always compute the visibility of adj-bodyl bend-dimension points,

// even though we don't check if adj_bodyl pointer is NULL or not,

// because we might want to know the visibility even if they don't exist.

//

// in practice this should not be a problem because almost always there are two

// adjacent bodies.

// bend_bodyl_point_left.compute_visibility(view) ; bend_bodyl__point_right.compute_visibility(view) ; if (adj_body2) { if (2 == bodyl2_left_the_same) { if (bend_body2_point_lef t .visible bend_bodyl_point_left .visible) {

// note we are switching points bend_body 2 _po i n t _ 1 e f t . s c r e e n_x bend_bodyl_jpoint_left . screen_x ; be nd_body 2 _po in t _1 e f t . s c r e e n_y bend_bodyl_point_lef t . screen_y ; bend_body2_point_left . screen_xl bend_bodyl_jpoint_lef t . screen_x2 ; bend_body2_point lef t . screen_yl bend_bodyl_jpoint_lef t . screen_y2 ; bend_body2_point lef t . screen_x2 = bend_bodyl_point_lef t . screen_xl ; bend_body2_point lef t . screen_y2 bend_bodyl_point_lef t . screen_yl ; bend_body2_point_lef t . z_buf_value bend bodyl point left . z buf value ;

»^' ' ^{" "} else bend_body2_point_lef t .compute_visibility (view) ; if (2 == bodyl2_right_the_same) { if ( bend_body 2_point_r ight . vi s ible bend_bodyl_point_right .visible) {

// note we are switching points bend_body2_point_right . screen_x = bend_bodyl_point_right . screen_x ; bend_body2_point_right . screen_y bend_bodyl_point_right . screen_y ; bend_body2_point_right . screen_xl

16 APPENDIX G bend_bodyl_point_right . screen_x2 bend_body 2_ p] oint_right . screen_yl bend_bodyl_point_right . screen_y2 bend_body 2_ p] oint_right . screen_x2 bend_bodyl_point_right . screen_xl ; bend_body2_point_right . screen_y2 bend_bodyl_point_right . screen_yl ; bend_body2_point_right . z_buf _value bend_bodyl_point_right . z_buf _value ;

else bend_body2_point_right .compute_visibility (view) ;

// else adj_body2 is NULL and there is nothing to compute

}

/*

This function tries to find the best way for drawing flange length for one bendline.

This function does it strictly for one bendline.

Important : it only determines which way would be the best way, but it does not decide if AD will really do it that way. Now will it do any updating other data structures.

In short, it does not change any of the AD data structures.

Note : no checks in the beginning.

Idea : in this function we use a heuristic to assign values for different bend-flange-dim points combinations. Eventually we choose the best combination.

The heuristic we use is a combination of the distance between the bend and flange dim points, plus the distance of the bend-dim point from the view screen.

A practical trick : first we check if the best known distance values is already 0. If yes, we will try to find a best match for another bend-dim points only if this new bend-dim points has a better z-buf value.

Note : after testing 4 parts, I found that this gives just a very slight improvement, may-be 5-10%. KK. */ void BM_AD_bendline: :compute_dim__points_to_display_indiv(

BM_AD_dim_corner ^♦bodies_bend_flange [6] /♦ 0,1 is bend-dim points; 3,4 is flange-dim points; 4,5 is flange lists ♦/,

17 APPENDIX G double distance and z [4] /^♦ 0,1 is distance; 3,4 is z-buf ♦/)

{ ^{" "}

BM_AD_dim_corner ^♦bes^fdp ; double best_distance ;

// check adjacent bodyl if (adj_bodyl ScSc ! (bodyl_drawn) ) {

// check if we already have a distance 0 bend- flange-dim points, if yes, go in only if the

// new bend-dim points has a strictly better z-buf value, if ((NULL == best_AD_bendline [0] ) |j distance_and_z [2] > (bend_bodyl_point_lef t . z_buf _value + AUTODIM_DISTANCE_TOLERANCE) | | distance_and_z [0] > AUTODIM_DISTANCE_TOLERANCE) { i f ( comp t e_bes t_dim_pair (parent- >view,

_bend_bodyl_point_lef t , bodyl_points , parent ->object_space_per_pixel, _best_distance, Scbest_fdp) ) { best_distance = fabs (best_distance) ; // check if the new point is better if (best_AD_bendline [0] ) {

// heuristic : if the sum of the z-buf value and distance is smaller i f ( ( b e s t _ d i s t a n c e + bend_bodyl_point_left . z_buf_value) < (distance_and_z [0] + distance_and_z [2] ) ) { best_AD_bendline [0] = this ; bodies_bend_f lange [0] =

_bend_bodyl_point_left ; bodies_bend_f lange [2] = best_fdp ; bodies_bend_f lange [4] = bodyl__points distance_and_z [0] = best_distance ; d i s t a n c e _a n d_ z [ 2 ] bend_bodyl_point_lef t . z_buf_value ;

else { best_AD_bendline [0] = this ; bodi e s_be nd_f l ange [ 0 ] _bend_bodyl_point_lef t ; bodies_bend_flange [2] = best_fdp ; bodies_bend_flange [4] = bodyl_points ; distance_and_z [0] = best_distance ; d i s t a n c e _ a n d _ z [ 2 ] = bend bodyl point lef t . z buf value ;

// check if we already have a distance 0 bend-flange-dim

18 APPENDIX G points, if yes, go in only if the

// new bend-dim points has a strictly better z-buf value. i f ( d i s t a n c e _ a n d _ z [ 2 ] >

(bend_bodyl_point_right.z_buf_ alue + AUTODIM_DISTANCE_TOLERANCE)

| j (NULL == best_AD_bendline [0] ) j | distance_and_z [0] >

AUTOD I M_D I S TANCE_TOLERANCE ) { if (compute_best_dim_pair (parent->view,

_bend_bodyl_point_right , bodyl_points , parent - >ob j ect_space_per__pixel , &cbest_distance , _best_fdp) ) { best_distance = fabs (best_distance) ; // check if the new point is better if (best_AD_bendline[0] ) {

// heuristic : if the sum of the z-buf value and distance is smaller i f ( ( b e s t _ d i s t a n c e + bend_bodyl_point_right . z_buf_value) < (distance_and_z [0] + distance_and_z [2] ) ) { best_AD_bendline [0] = this ; bod i-e s_bend_f lange [ 0]

_bend_bodyl_point_right bodies_bend_flange [2] = best_fdp ; bodies bend flange [4] = bodyl_points distance_and_z [0] = best_distance d i s t a n c e _ a n d _ z [ 2 ] bend_bodyl_j?oint_right . z_buf_value ;

}

} else ( best_AD_bendline [0] = this ; bo d i e s_ben d_ f l ang e [ 0 ] Scbend_bodyl_point_right ; bodies_bend_flange [2] = best_fdp ; bodies_bend_flange [4] = bodyl__points distance_and_z [0] = best_distance ; d i s t a n c e _ a n d _ z [ 2 ] bend_bodyl_point_right . z_buf _value ;

}

// check adjacent body2 if (adj_body2 ScSc ! (body2_drawn) ) {

// check if we already have a distance 0 bend-flange-dim points, if yes, go in only if the

// new bend-dim points has a strictly better z-buf value. if ^{( (}NULL == best_AD_bendline[l] ) | | distance and z [3] _>

(bend_body2_point_left.z_buf_value +AUTODIM_DISTANCE TOLERANCE) I I I I

19 APPENDIX G distance_and_z [1] > AUTODIM_DISTANCE_TOLERANCE) { if ( comput e_best_dim_pair (parent ->view,

Scbend_body2_point_lef t , body2_points , parent- >object_space_jρer_pixel , _best_di stance , _best_f dp) ) { best_distance = f abs (best_distance) ; // check if the new point is better if (best_AD_bendline[l] ) {

// heuristic : if the sum of the z-buf value and distance is smaller i f ( ( b e s t _ d i s t a n c e + bend_body2_point_lef t . z_buf_value) < (distance_and_z [1] + distance_and_z [3] ) ) { best_AD_bendline [1] = this ; bodies_bend_f lange [ 1 ] Scbend_body2_point_left ; bodies_bend_f lange [3] = best_fdp ; bodies_bend_f lange [5] = body2_points distance_and_z [1] = best_distance ; d i s t a n c e _ a n d _ z [ 3 ] bend_body2_point_lef t . z_buf_value ;

else { best_AD_bendline [1] = this ; b o d i e s _b e n d_ f l a ng e [ l ] Scbend_body2_point_left ; bodies_bend_flange[3] = best_fdp ; bodies_bend_flange[5] = body2_points ; distance_and_z [1] = best_distance ; d i s t a n c e _ a n d _ z [ 3 ] - bend_body2_point_lef t . z_buf _value ;

// check if we already have a distance 0 bend- flange -dim points, if yes, go in only if the

// new bend-dim points has a strictly better z-buf value. i f ( d i s t a n c e _ a n d _ z [ 3 ] >

(bend_body2_point_right . z_buf _value + AUTODIM_DISTANCE_TOLERANCE)

| | (NULL == best_AD_bendline [1] ) | | distance_and_z [1] >

AUTODIM_DISTANCE_TOLERANCE) { i f ( comput e_be s t_dim_pair (parent ->view,

_bend_body2_point_right , body2_point s , parent - >ob j ect_space_per_pixel , 5cbest_distance , Scbest_fdp) ) { best_distance = f abs (best_distance) ; // check if the new point is better

20 APPENDIX G 44 r

if (best_AD_bendline [1] ) {

// heuristic : if the sum of the z-buf value and distance is smaller i f ( ( b e s t _ d i s t a n c e + bend_body2_point_right . z_buf_value) < (distance_and_z [1] + distance_and_z [3] ) ) { best_AD_bendline [1] = this ; bodies_bend_f lange [1]

Scbend_body2_point_right bodies_bend_f lange [3] = best_fdp ; bodies_bend_f lange [5] = body2_points distance_and_z [1] = best_distance d i s t a n c e _ a n d_ z [ 3 ] bend body2_point right . z buf value ;

- - } else { best_AD_bendline [1] = this ; bod i e s__^b end_f l ange [ l ] Scbend_body2__point_right ,- bodies_bend_flange [3] = best_fdp ; bodies_bend_flange [5] = body2_points ; distance_and_z [1] = best_distance ; d i s t a n c e _ a n d _ z [ 3 ] bend body2 point right . z buf value ;

}

/*

This function determines how to draw flange length for a bendline.

It is different from the previous function in that using the previous function as a subroutine, it decides how to draw flange length dimensions for a bendline, as well as updates all data structures accordingly.

Also, it takes into account simultaneous bendlines. */ void BM_AD_bendline: :compute_dim_points_to_display(void)

// qualification check if (NULL == bend) return ;

21 APPENDIX G

BM_AD_bendline ^♦bl ; long i ;

// these variable store the bendline dimension data that will be used later to draw

// 0,1 is with respect to adj -bodyl; 2,3 is with respect to adj -body2

BM_AD_dim_corner ^♦bodies_bend_flange_list [6] ; // 0,1 is bend-dim points; 2,3 is flange-dim points; 4,5 is flange lists double distance_and_z [4] ; // 0,1 is distance; 2,3 is z-buf

// this is the bendline for which we will be drawing the dimensions.

// this is trivial when there is no simultaneous bendlines.

BM_AD_bendline ♦best_AD_bendline [2] = { NULL, NULL } ;

/*

Compute the best way for drawing flange length. Take simultaneous bendlines into account.

*/

// these variables are for handling simultaneous bendline property things.

BM_BEND_PROPERTY ^♦prop ; BM_LINKED_LIST_NODE ^♦prop_list ; DBM_ARRAY ^♦bendlines ; BM_BENDLINE ^Λbnd ; long size, this_bend_seen = 0 ; for (prop_list = bend->get_properties () ; prop_list ; prop_list = prop_list->next () ) { prop = (BM_BEND_PROPERTY ^♦) prop_list->get_obj 0 ; if (prop- >is (BM_TYPE_BEND_SIMULTANEOUS) ) { bendlines = (DBM_ARRAY ^♦) prop->get_bendlines 0 ; size = bendlines->get_size () ; for (i = 0 ; i < size ; i++) {

// get a bendline in the bend property object if (NULL == (bnd = (BM_BENDLINE ^♦) bendlines->get (i) ) ) continue ; bl = parent->get_AD_bend_structure (bnd) ; if (NULL == bl) continue ; if (bl == this) this bend_seen = 1 ; if (1 bl->data_valid^~] I bl->ignore) continue ; // compute how to draw flange length dimensions for this one bendline. bl- >compute_dim_points_to_display_indiv (best_AD_bendline, bodies bend flange list, distance and z) ;

if ( ! this_bend_seen) { // make sure this bendline gets processed

22 APPENDIX G bl = parent->get_AD_bend_strueture(bend) ; if (bl -Sc bl->data_valid ScSc ! bl->ignore) { bl- >compute_dim_points_to_display_indiv (best_AD_bendline, bodies bend flange list, distance_and_z) ;

^' )^{' ] ~}

I*

Mark all other bendlines in the simultaneous bend as "ignore" .

*/ for (prop_list = bend->get_properties 0 ; prop_list ; prop_list = prop_list->next0 ) { prop - (BM_BEND_PROPERTY ♦) prop_list->get_obj () ; if (prop->is(BM_TYPE_BEND_SIMULTANEOUS) ) { bendlines = (DBM_ARRAY ^♦) prop->get_bendlines() ; size = bendlines->get_size() ; for (i = 0 ; i < size ; i++) {

// get next bendline in the bend property object if (NULL == (bnd = (BM_BENDLINE ^♦) bendlines->get (i) ) ) continue ; bl = parent->get_AD_bend_structure (bnd) ; if (NULL == bl) continue ;

// mark as ignore if (bl->adj_bodyl __ ! (bl->bodyl_drawn) __ bl != best_AD_bendline[0] ) { bl->bodyl_drawn = 1 ; b 1 - > d r a n_b n d_d i m_p o i n t 1 bodies_bend_f lange_list [0] ; b 1 - > dr awn_f 1 ange_dim_po in t 1 bodies bend flange list [2] ;

^{" " "} } if (bl->adj_bodyl __ ! (bl->body2_drawn) &_ bl != best_AD_bendline [1] ) { bl->body2_drawn = 1 ; b 1 - > d r a n_b e n d_d i m_p o i n t 2 bodies_bend_f lange_list [1] ; bl - >drawn_f lange_dim_po int 2 bodies bend flange list [3] ;

/* mark this bendline for drawinσ

*/ if (best_AD_bendline[0] ) {

23 *

APPENDIX G best_AD_bendline [0] - >bodyl_drawn = 2 ; best_AD_bendline [0] - >drawn_bend_dim_point 1 bodies_bend_f lange_list [0] ; best_AD_bendline [0] - >drawn_f lange_dim_point 1 bodies bend flange list [2] ; i ^"f (be rst_AD_bendline [1] ) { best_AD_bendline [1] - >body2_drawn = 2 ; bes t_AD_bendl ine [1] - >drawn_bend_dim_point 2 = bodies_bend_f lange_list [1] ; best_AD_bendline [1] - >drawn_f lange_dim_point2 bodies ^" bend r flange list [3] ;

/*

Check if the flange-dim points is adjacent to a bendline. Here we will not check simultanoues bends, because they should have exactly the same flange-list as the one we have here now.

*/

BM_AD_dim_corner ^♦temp_point ; if (best_AD_bendline[0] ) { for (temp_point = bodies bend_f lange_list [4] ; temp_point ,- temp_point = temp_point->next) ^~[ if (NULL == temp point->adj_bendline) continue ;

(best_AD_bendline [0] ->parent) ->compute_ef fects_f lange_length_draw n(best_AD_bendline [0] ->bend, best_AD_bendline [ 0 ] - >adj _bodyl , bodies_bend_f lange_list [0] , bodies_bend_f lange_list [2] , temp_point->adj_bendline) ;

if (best_AD_bendline[l] ) { for (temp_point = bodies bend_f lange_list [5] ; temp_point ; tempjpoint = temp_point->next) ^~{ if (NULL == temp_point->adj_bendline) continue ;

(best_AD_bendline [1] ->parent) -_>compute_ef f ects_f lange_length_draw n(best_AD_bendline [1] ->bend, best_AD_bendline [ l ] - >adj _body2 , bodies_bend_f lange_list [1] , bodies_bend_f lange_list [3] , temp point->adj bendline) ;

} ' ^"

}

24 APPENDIX G

/*

This function draws flange length for this bendline.

It uses the informat ion computed by compute_dim_points_to_display() f nction.

TODO : in this function we could keep track which bendline dimension points have already been drawn.

Sometimes bendline dimension points with respect to adj -bodyl and adj-body2 match. In that case we should draw it really only once. */ void BM AD_bendline: :draw_dimension_points 0

{

// qualification check if (! data_valid | | ignore) return ; if (adj_bodyl __ 2 == bodyl_drawn) { parent - >draw_dim_point ( dr awn_bend_dim_point 1 , drawn_f lange_dim_point 1 , parent - >view , parent->dc, parent - >obj ect_space_per_j_?ixel ) ,- if (adj_body2 ScSc 2 == body2_drawn) { parent - >draw_dim_point ( dr awn_bend_dim_point 2 , drawn_f lange_dim_point 2 , parent - >view , parent->dc, parent - >ob j ect_space_per_pixel ) ;

}

/*

********** ********** ********** ********** **********

********** ********** ********** **********

BM_AUTO_DIMENSION class stuff.

*/

/*

This is a debug function.

*/ void BM AUTO DIMENSION: :test draw(void)

{ ^{" "}

BM_AD_dim_corner ^♦temp_point ; int i ; for (i = 0 ; i < num_of_bendlines ; i++) { if (! bends [i] ->data_valid) continue ;

25 APPENDIX G if (bends [i] ->adj_bodyl) {

(bends [i] ->bend_bodyl_point_left) .draw_with_dim_line (view,dc,obje ct_space_per_pixel) ;

(bends [i] ->bend_bodyl_point_right) .draw_with_dim_line (view, dc,obj ect_space_per_pixel) ; for (temp_point = bends [i] ->bodyl_points ; temp_point ; temp_ oint = temp_point->next) { temp_point->draw_with_dim_line (view,dc,object_space_per__pixel) ;

if (bends [i] ->adj_body2) {

(bends [i] ->bend_body2_point_left) .draw_with_dim_line (view,dc,obje ct_space_per_pixel) ;

(bends [i] ->bend_body2_point_right) .draw_with_dim_line (view,dc,obj ect_space_jper_pixel) ; for (temp_point = bends [i] ->body2_points ; temp_point ; temp__point = temp_point->next) { temp_point->draw_with_dim_line (view,dc,object_space_per_pixel) ;

}

/*

This function draws an info-box with bendline data for every bendline.

Idea : we will not check the bounding box of the visible portion of the part in the viewing area, because in SOLID mode renderware does not give out that kind of information.

Moreover, very often the user zooms in to see part dimensions, in which case there is no portion of the viewport that in not covered by the part, ie. the bounding box would cover the entire view area and would be practically useless.

Instead we will be drawing bendline information boxes along the edges of the viewing area.

We will divide the viewing area into 4 regions, and place bendline info boxes in those areas,

26 APPENDIX G whichever is closer to the bendline point to which the info-box refers to.

As of right now, we will be drawing those boxes along the left-right vertical edges of the viewport. */ void BM_AUTO_DIMENSION: :draw_bendline_data(CDC ^♦pdc)

{ int i, j ;

// these variables are used to calculate the info-box size and to draw it int font_height = 21 ; int font_avg_char_width = 8 ; // average width of a character int box_height = (font_height << 1) /^♦ 2 rows of text ^♦/ ; int half_box_height = box_height >> 1 ; int box_width = font_avg_char_width ^♦ 15 ; // space for 15 characters int box_x ; int vert_space_per_box ; int offset ;

// compute maximum number of info-boxes we can draw, int half_max_num_info_boxes = view_size_y / box_height ; if (half_max_num_info_boxes < 1) return ; // not enough screen space int max_num_info_boxes = half_max_num_info_boxes << 1 ;

/* compute the number of bends we can display

*/ int num_bends_to_be_drawn = 0 ; for (i = 0 ; i < num_of_bendlines ; i++) { if (! bends [i] ->data_valid | | NULL == bends [i] ->bend) continue ;

// remember, visibility of adj_bodyl bend-dimension points is always computed.

// here we only check visibility with respect to adj_bodyl.

// most likely visibility with respect to adj_bodyl and adj_body2 is the same. if ( (bends [i] ->bend_bodyl_point left) .visible | | (bends [i] ->bend_bodyl_point_right) .visible) (^~ +_}+num bends-to-be-drawn ; else continue ; if ⁽num_bends_to_be_drawn >= max num_info_boxes) break ; // out of view space

27 APPENDIX G

// check if there are any bendlines to be drawn if (0 == num_bends_to_be_drawn) return ;

/* allocate space for storing bendline pointers and y-coordinates of endpoints.

*/

B M_AD_be n d l i n e ^{♦ ♦} be n d l i n e s = n e w

BM_AD_bendline^Λ [num_bends_to_be_drawn] ; if (NULL == bendlines) return ; // out of space double ^♦bend_y_coordinates = new double [num_bends_to_be_drawn] if (NULL == bend_y_coordinates) { // out of space delete [] bendlines ; return ; } BM_AD_dim_corner ^{♦ ♦}bend_points = new

BM_AD_dim_corner^Λ [num_bends_to be_drawn] ; if (NULL == bend_points) delete [] bendlines ; delete [] bend__y_coordinates ; return ; }

/* collect bendline pointers and bend-dim-point pointers

*/ for (i = j = 0 ; i < num_of_bendlines ; i++) { if (! bends [i] ->data_valid | | NULL == bends [i] ->bend) continue ; if (! (bends [i] ->bend_bodyl_point_left) .visible ScSc ! (bends [i] ->bend_bodyl_point_right) .visible) continue ;

// first, decide which of the endpoints we will use bendlines [j] = bends [i] ; if ( (bends [i] ->bend_bodyl_point_left) .visible) { if (! (bends [i] ->bend_bodyl_point_right) .visible) bend_points [j] = _ (bends [i] ->bend_bodyl_point_left) ; else { i f

( (bends [i] ->bend_bodyl_point_left) . z_buf_value <

(bends [i] ->bend__bodyl_point_right) . z_buf_value) b e n d _ p o i n t s [ j ] _ (bends [i] ->bend_bodyl_point_lef t) ; e l s e b e n d _ p o i n t s [ j ]

Sc (bends [i] ->bend_bodyl_point_right) ;

elsebend_points [j] = -(bends [i] ->bend_bodyl_point_right)

if (++j >= num_bends_to_be_drawn) break ; // out of view space

28 APPENDIX G

}

/ * sort points according to which side of the screen the info-box should be right side contains all boxes (at least initially) whose bend-line-dim-point is not on the left side of the screen. */ int half screen = (view_size_x >> 1) ; i = 0 ; 7/ i will be top j = num_bends_to_be_drawn - 1 ; // j will be bottom

// we will use these to temporarily store stuff

BM_AD_bendline ^♦p aend ;

BM_AD_dim_corner ^♦p_bend_point ; while (i < j ) { while (i < j cSc bend_j_?oints [i] - >screen_x < half_screen)

1 + + while ( i < j ScSc bend_points [ j ] - >screen_x >= half_screen) j - - ;

// switch i and j if d < j ) { p_bend = bendlines [i] ; p_bend_point = bend__points [i] ; bendlines [i] = bendlines [j ] ; bend_jpoints [i] = bend_points [j ] ; bendlines [j ] = p_bend ; bend_points [j ] = p_bend_point ;

if (bend_points [i] ->screen_x < half_screen) i++ ; // i points to the first info-box on the right side

/* check that neither side (ie. left-right) has too many info-boxes */ if (i > half_max_num_info_boxes) {

// too many boxes on the left side.

// sort them according to x-coordinates.

// this way we will move points with the largest x-coordiates to the right side. for (j = 0 ; j < i ; j++) bend_y_coordinates [j] = bend_points[j] ->screen_x ;

DBM_QuickSort_double(bend_y_coordinates, (unsigned long) i, (long ^♦) bendlines) ;

29 APPENDIX G i} = half-max-num-info-boxes ; else if ( ( num_bends_to_be_drawn - i) > half max_num_info_boxes) {

// too many boxes on the right side.

// sort them according to x-coordinates.

// this way we will move points with the smallest x-coordiates to the left side. for (j = i ; j < num_bends_to_be_drawn ; j++) bend_y_coordinates [j] = bend_j?oints [j] ->screen_x ;

DBM_QuickSort_double (bend_y_coordinates + i, (unsigned long) (num_bends_to_be_drawn - i) ,

(long ^♦) (bendlines + i) ) ; i} = num"bends to be drawn - half-max-num_info-boxes ;

// it could have been that when we chose a point for a bendline, it is not the best point.

// (although this point was used to place this point on the left-right side) .

// imagine a point on the left, it could be that both endpoints are visible, but

// out point is not the best, because the other endpoint of the bendline is actually to the // left of our point, for (j = 0 ; j < i ; j++) {

// check if we have choice if (! (bendlines [j] ->bend_bodyl_point_left) .visible | | ! (bendlines [j] ->bend_bodyl_point_right) .visible) continue ; if ( (bendlines [j] ->bend_bodyl_point_left) .screen_x < (bendlines [j] ->bend_bodyl_point_right) .screen_x) b e n d _ p o i n t s [ j ] Sc(bendlines [j] ->bend_bodyl_point_left) ; e l s e b e n d _ p o i n t s [ j ] =

-(bendlines [j] ->bend_bodyl_point_right) ; for (j = i ; j < num_bends_to_be_drawn ; j++) { // check if we have choice if (! (bendlines [j] ->bend_bodyl_point_left) .visible | | ! (bendlines [j] ->bend_bodyl_point_right) .visible) continue ; if ( (bendlines [j] ->bend_bodyl_point_left) . screen_x > (bendlines [j] ->bend_bodyl_point_right) .screen_x) b e n d _ p o i n t s [ j ] Sc (bendlines [ j] ->bend_bodyl_point_lef t) ; e l s e b e n d _ p o i n t s [ j ]

Sc (bendlines [j ] ->bend_bodyl_point_right) ;

30 H?Γ

APPENDIX G

// store the y-coordinate for Quicksort for (j = 0 ; j < num_bends_to_be_drawn ; j++) bend_y_coordinates [j] = bend_points [j] ->screen_y ;

// sort both sides according to the screen y-coordinate

// note a glitch : while the bendlines [] array has been permuted, the bend_points [] array has not been.

// in the next paragraph we restore the point array. TODO : may-be there is a better solution.

// we could resort the array, but then we need to get original y-coordinates array.

// we could just compute a separate permutation array, but then we need space for it, and

// the code to compute the permutation, which is non-tirivial .

// anyway, for now we just do it the easy way. if (i > 0) DBM_QuickSort_double (bend_y_coordinates, (unsigned long) i, (long ^♦) bendlines) ; if ( ( num_bends _t o_be_d r awn - i ) > 0 ) DBM_QuickSort_double (bend_y_coordinates + i,

(unsigned long) (num_bends_to_be_drawn - i) , (long ^♦) (bendlines + i) ) ; for (j = 0 ; j < i ; j++) { if ( (bendlines [j] ->bend_bodyl_point_lef t) .visible) { i f ( !

(bendlines [j] ->bend__bodyl_point_right) .visible) bend_points [j ] Sc (bendlines [j] ->bend bodyl_point_lef t) ; else i f

( (bendlines [j] - >bend_body l_po int lef t) .screen_x <

(bendlines [j] ->bend_bodyl_point_right) .screen_x) b e n d _ p o i n t s [ j ] Sc (bendlines [j] ->bend_bodyl_point_lef t) ; e l s e b e n d _ p o i n t s [ j ]

-(bendlines [j] ->bend bodyl point right) ;

. ^T e l s e b e n d _ p o i n t s [ j ]

Sc (bendlines [ j ] ->bend_bodyl_point_right ) ;

for (j = i ; j < num_bends_to_be_drawn ; j++) { if ⁽ (bendlines [j] ->bend__bodyl_point_left) .visible) { i f ( !

⁽bendlines [j] ->bend_bodyl_point_right) .visible) bend_points [j] = -(bendlines [j] ->bend bodyl_point_left) ; else i f

⁽ (bendlines [j] ->bend_bodyl_point_left) . screen_x _>

31 APPENDIX G

(bendlines [j] ->bend_bodyl_point_right) .screen_x) b e n d _ p o i n t s [ j ] -(bendlines [j] ->bend_bodyl_point_lef t) ; e l s e b e n d _ p o i n t s [ j ]

_ (bendlines [j] ->bend bodyl_point_right) ;

} e l s e b e n d _ p o i n t s [ j ]

Sc (bendlines [ j ] ->bend_bodyl_point_right) ,-

/*

Ok, now we are ready for drawing. */ static char info_box_text [256] ; static POINT ■box_corners [5] ; static POINT arrow[3] ; BM_BENDLINE ^♦b ; BM_BEND_OP ^♦bop ; BM_2D_BODY ♦bjsody ;

/* draw left side */ int left = 3 ; int top ; if (0 == i) goto draw_right_side ;

// compute vertical space per info-box vert_space_per_box = view_size_y / i ;

// compute vertical offset for the top-left corner of the info-box offset = (vert_space_per_box - box_height) >> 1 ; box_corners [0] .x = box_corners [3] .x = box_corners [4] .x = left box_corners [1] .x = box_corners [2] .x = left + box_width ; for (j = 0 ; j < i ; j++) { b = bendlines [j] ->bend ; bop = b->get_bend_op () ; if (NULL == bop) continue ; // 043996KK if (view_type) b_body = b->get_current_3D_version0 ; else b_body = b->get_flat() ; top = j^vert_space_jper_box + offset ;

32 APPENDIX G box_corners [0] .y = box_corners [1] .y = box_corners [4] .y = top ; box_corners [2] .y = box_corners [3] .y = top + box_height ; sprintf (info_box_text, "A:%5.If,L:%6.2f" ,

(bop-_>get_bend_angle2 O ) ^♦57.29578, (b_body->get_current_center_lin e 0 ) .get_len() ) ; pdc - >TextOut ( 5 , top , inf o_box_text , strlen(info_box_text) ) ; switch (bop->get_type 0 ) ( case BM_TYPE_BEND_OP_REGULAR : sprintf (info_box_text, "R:%5.2f,D:%5.2f",

( ( B M_ B E ND_O P _R E G U LAR ♦ ) bop) - _>get_bend_radius ( ) , bop- >get_bend_deduction () ) ,- pdc- >TextOut (5 , top + font_height, inf o_box_text , strlen (info_box_text) ) ; break ; case BM_TYPE_BEND_OP_CONIC : // TODO : implement conic bend op. break ;

}

// draw box pdc->Polyline (box_corners, 5) ;

// draw arrow arrow [0] .x = left + box_width ; arrow [0] .y = top + half_box_height ;

// calculate arrow tip at fixed percentage length in from end point

C P o i n t ptLeft (bendlines [j] - >bend_body l_po in t _1 e ft . screen_x, bendlines [j] ->bend_bodyl__point_lef t . screen_y) ;

C P o i n t ptRight (bendlines [j] - >bend_body l_point_r ight . screen_x, bendlines [j] ->bend_bodyl_point_right . screen_y) ;

CSize sizeDiff = ptRight - ptLeft;

CPoint ptArrowTip; i f ( b e n d _ p o i n t s [ j ] = =

Sc (bendlines [j] ->bend_bodyl_point_lef t) ) ptArrowTip.x = ptLeft.x + sizeDiff .cx/8; ptArrowTip.y = ptLeft.y + sizeDiff.cy/8; else {

33 8*

APPENDIX G ptArrowTip. x = ptRight. x - sizeDif f . cx/8 ; ptArrowTip. y = ptRight. y - sizeDif f . cy/8 ; arrow [1] .x = ptArrowTip.x; arro [1] .y = ptArrowTip.y; pdc->Polyl ine (arrow, 2) ;

// change to solid pen for arrow heads

CPen* pArrowPen = new CPen ( PS_SOL I D

BCAD_context . m_nBendLineWeight , BCAD_context . m_crBend) ; CPen^ pOldPen = pdc->SelectObject (pArrowPen) ;

// compute arrow head CAutoDimArrow arrowBend( ptArrowTip,

CPoint(arrow[0] .x,arro [0] .y) ,

BCAD_context.m nArrowAngle,

BCAD_context.m_nArrowLength ) ;

// draw arrow head arrowBend.DrawHead(dc) ; pdc->SelectObject (pOldPen) ; delete pArrowPen;

}

/* draw right side

*/ draw_right_side : if (0 == num_bends_to_be_drawn - i) goto done ; box_x = view_size_x - box_width - 3 ; left = box_x ;

// compute vertical space per info-box vert_space_per_box = view_size_y / (num_bends_to_be_drawn - i)

;

// compute vertical offset for the top-left corner of the info-box offset = (vert_space_per_box - box_height) >> 1 ; box_corners [0] .x = box_corners [3] .x = box_corners [4] .x = left box corners [1] .x = box corners [2] .x = left + box width ;

34 APPENDIX G box_x += 2 ; for (j = i ; j < num_bends_to_be_drawn ; j++) { b = bendlines [j] ->bend ; bop = b->get_bend_op() ; if (NULL == bop) continue ; // 043996KK if (view_type) b_body = b->get_current_3D_version() ; else b_body = b->get_flat 0 ; top = (j - i) ^♦vert_space_per_box + offset ; box_corners [0] .y = box_corners [1] .y = box_corners [4] .y = top box_comers [2] .y = box_corners [3] .y = top + box_height ; sprintf (info_box_text, "A:%5.If,L:%6.2f" ,

(bop->get_bend_angle2 () ) ♦57.29578, (b_body->get_current_center_lin e() ) .get_len() ) ; pdc - >Text Ou t ( box_x , top, i n f o_box_ t ex t , strlen (info_box_text) ) ; switch (bop->get_type 0 ) { case BM_TYPE_BEND_OP_REGULAR : sprintf (info_box_text, "R:%5.2f,D:%5.2f" ,

( ( B M_ B EN D_O P_ R E G U L A R ^♦ ) bop) ->get_bend_radius () ,bop->get_bend_deduction() ) ; pdc->TextOut (box_x, top + font_height, info_box_text, strlen(info_box_text) ) ; break ; case BM_TYPE_BEND_OP_CONIC : // TODO : implement conic bend op. break ; }

// draw box pdc->Polyline (box_corners, 5) ;

// draw arrow arro [0] .x = left ; arrow [0] .y = top + half_box_height ;

// calculate arrow tip at fixed percentage length in from end point

C P o i n t ptLeft ⁽bendl ines [j] ->bend_bodyl_point_left . screen_x, bendlines [j] ->bend_bodyl_point_lef t . screen_y) ; ^~

35 C«5t

APPENDIX G

C P o i n t ptRight (bendlines [j] - >bend_body l_poi t_r ight . screen_x , bendlines [j ] - >bend_bodyl_point_right . screen_y) ;

CSize sizeDif f = ptRight - ptLeft;

CPoint ptArrowTip; i f ( b e n d _ p o i n t s [ j ] = =

-(bendlines [j] ->bend_bodyl_point_left) )

{ ptArrowTip.x = ptLeft.x + sizeDiff.cx/8; ptArrowTip.y = ptLeft.y + sizeDiff.cy/8;

} else

{ ptArrowTip.x = ptRight.x - sizeDiff .cx/8; ptArrowTip.y = ptRight.y - sizeDiff .cy/8;

} arro [1] .x = ptArrowTip.x; arrow[1] .y = ptArrowTip.y; pdc->Polyline(arrow, 2) ;

// change to solid pen for arrow heads CPen* pArrowPen = new CPen ( PS_SOL ID

// compute arrow head CAutoDimArrow arrowBend ( pt rrowTip ,

CPoint (arrow [0] .x, arrow [0] .y) ,

BCAD_context . _nArrowAngle ,

BCAD_context . m_nArrowLength ⁾ ;

// draw arrow head arrowBend . DrawHead ( dc ) ; pdc->SelectObject (pOldPen) ; delete pArrowPen;

}

/*

Done, release memory we allocated. */ done • delete [] bend_points ; delete [] bend_y_coordinates ; delete [] bendlines ;

36 APPENDIX G

/*

To draw part dimensions in the device context.

*/ int BM AUTO_DIMENSION: :draw_auto_dimensions (CDC ^♦target_dc)

{

BM_AD_dim_corner ^♦temp_pomt ; int i ;

// check that we have a valid dc if (NULL == target_dc) return 0 ;

// check if we need to recompute the data structures if (dirty) { if (! compute_auto_dimension_data () ) return 0 ,- if (dirty) return 0 ;

} if (NULL == part) return 1 ; dc = target_dc ;

// get view size view_size_x = view->get_viewport_width() ; view_size_y = view->get_viewport_height () ;

// create a mesh to determine clear spaces to draw dimensions Sc data pMesh = new CAutoDimRgnMesh(view_size_x, view_size_y) ;

// check if we need to compute points to ignore in solid i f ( v i e w - > g e t _d r a w i n g_m o d e ( ) Sc Sc ! points_to_ignore_in_solid_computed) { c_}ompute points-to-ignore-in-solid0 ;

object_space_per_pixel = view->get_object_to_device_ratio 0 ;

// initially set all "ignore" flags to FALSE, we should not ignore anything.

// also, nothing is drawn yet. for (i = 0 ; i < num_of_bendlines ; i++) { bends [i] ->ignore = 0 ; bends [i] ->bodyl_drawn = bends [i] ->body2_drawn = 0 ;

/*

Compute visibility of all bendline dimension points of all bendlines .

37 APPENDIX G continue ; bends [i] ->draw_dimension_points ()

// test display of dirty region

CBrush brRed(RGB(255,0,0) ) ;

CBrush^ pOldBrush = dc->SelectObject (ScbrRed) dc->PaintRgn(pDirtyRegion) ; dc->SelectObject (pOldBrush) ; delete pDirtyRegion; pDirtyRegion = NULL;

/*

Create a pen for drawing the bendline info.

*/

// create the pen : type, (line)width, color.

CPen penBend;

CPen* pBendPen = create_auto_dim_pen ( _penBend ,

BCAD_context . m_nBendLineStyle , BCAD_context . m_nBendLineWeight ,

BCAD_context , m_crBend) ;

// select the new pen dc->SelectObject (pBendPen) ; dc->SetTextColor(BCAD context.m crBend) ;

/*

Draw bendline info. */ if (BCAD_context.m_bShowBendDim) draw_bendline_data (dc) ;

// Delete the mesh now that we are finished with it .

// delete pMesh;

/*

Restore old pen.

'/ dc->SelectObject (pOldPen) ; dc->SetTextColor(oldtextcolor) ; dc->SetBkMode (old_bk_mode) ;

// see if we have to disable auto-dimensioning in the view class if (reset_view_to_false_once_drawn) { reset_view_to_false_once_drawn = o ; view->set_auto_dimensions (0) ;

39 <fS4

APPENDIX G return 1

speed else {

CBrush brushNew(cr) ; LOGBRUSH lb; brushNew.GetLogBrush(Sclb) ;

DWORD dwStyleArray[6] ; int nStyleCount = 2 ;

DWORD dwDashLength = 2 + (nLineWeight ^♦ 4) DWORD dwDotLength = dwDashLength / 4; DWORD dwGapLength = dwDashLength / 2 ; switch (nLineStyle)

{ case PS_DASH: dwStyleArray[0] = dwDashLength; // pixels on dwStyleArray[1] = dwGapLength; // pixels off nStyleCount = 2; break; case PS_DOT: dwStyleArray[0] = dwDotLength; // on dwStyleArray[1] = dwGapLength; // off nStyleCount = 2 ; break; case PS_DASHDOT: dwStyleArray[0] dwDashLength; // on dwStyleArray[1] dwGapLength; // off dwStyleArray[2] dwDotLength; // on dwStyleArray[3] dwGapLength; // off nStyleCount = 4 break; case PS_DASHDOTDOT: dwStyleArray[0] dwDashLength; // on dwStyleArray[1] dwGapLength; // off dwStyleArray[2] dwDotLength; // on dwStyleArray[3] dwGapLength; // off

40 APPENDIX G dwStyleArray[4] = dwDotLength; // on dwStyleArray[5] = dwGapLength; // off nStyleCount = 6; break; pPen->CreatePen(PS_GEOMETRIC | PS_USERSTYLE, nLineWeight, Sclb, nStyleCount, dwStyleArray) ;

} return pPen; }

void CBendCADDoc: :OnDrawAutoDim()

{

// if we have no dimension object, do nothing. if (NULL == auto_dim) { auto dim = new BM AUTO DIMENSION (this) ;

} ^{" " "} if (NULL == auto_dim) return ; ^'

CBendCADViewPart ^♦pActiveView = (CBendCADViewPart ^♦) get_view (BENDCAD_VIEW_TYPE_PART) ; if (NULL == pActiveView) return ; if ( ! pActiveView- >get_auto_dimensions ( ) ) auto_dim->disable_view_dimensions_f lag_next_time 0 ; pActiveView- >set_auto_dimensions ( 1 ) ;

// always recompute dimension points aut o_dim- > inval idat e ( ) ; pActiveView- >Invalidate (0) ; }

41 APPENDIX H umuiutuuuuiuuu ii urn ///////////////////////////// / /// Example of auto dimensioning functions for computing u// // dimension points. Dimension points are computed // // before flange length dimensions are drawn. //

// //

//////////////////////////////////////////////////////////////

#include "stdafx.h"

#include "BendCAD.h"

#include "BendCADDoc.h"

#include "BendCADViewPart .h"

#include "AUTODIM.HXX"

/*

********** ********** ********** ********** **********

********** ********** ********** ***.*******

BM_AD_bendline class stuff.

********** ********** ********** ********** **********

********** ********** ********** ********** */

/*

Thi s i s a l ocal f unc t ion used by

BM_AD_bendline : : comput e_BM_AD_f lange_points ( .. ) to compute the distance between a given reference point and an edge point .

It returns FALSE if the point is one the wrong side of the refrence line.

Vector dir should be a unit vector.

*/ static int compute_distance_from_ref__point (BM_P0INT const- pt /^♦ given reference point ^♦/,

BM_VECTOR const- dir /^♦ direction vector ^♦/,

BM_POINT ⁺p /♦ point whose distance we want to compute ^♦/, double -distance)

{

BM VECTOR v(*p - pt) ; double x - v % dir ; if (x < -AUTODIM_DISTANCE_TOLERANCE) return 0 ; distance = x ; return 1 ;

}

/* APPENDIX H

A l o c a l f u n c t i o n u s e d b y

BM_AD_bendline : : comput e_BM_AD_f lange_points ( . . . ) .

Note : a dimension list we will be referring to is a list of flange dimension points.

A dimension list is open, it the number of dimension points in the list open to the "left" is not equal to the number of dimension point in the list open to the "right" .

Usually this means that the last flange dimension point we added to the list has no matching "closing" flange dimension point.

This whole thing is causes by the possibility that we can have a line (that looks like one line) actually broken into small pieces. In that case the user does not see those small pieces

(to him it looks like one line) and we don't want to consider points in the middle as legimate flange dimension points.

This function returns TRUE iff we can skip the given edge and leave the dimension list open, and FALSE iff we need to close the open dimension by creating a left-side dimension point from one of the end-points of the current edge.

Basically, this function returns TRUE iff the next edge in the loop is a line that will not be ignored, acutally, plus some more conditions.

Note : this function is called when the endpoint of the edge is at flange-length distance and we are considering whether we should draw a (flange) dimension point for this endpoint . */ static int leave_dimension_open(BM_EDGE ♦edge /♦ current edge ♦/, BM_LOOP ^♦bloop /^♦ loop where the current edge belongs to ♦/, BM_VECTOR const- side_vector /^♦ side vector for the current bendline-ajd_flange pair ♦/,

BM_VECTOR const- normal /^♦ normal of the plane underlying the adjacent flange ♦/)

BM_EDGE ^♦nex^edge = (BM_EDGE ^♦) edge->next() ; if ⁽NULL == next_edge) next_edge = bloop-_>get_first_edge () ; if ⁽next_edge->is⁽BM_ENTITY_TYPE_LINE) ) {

// problem, if the next edge (which is a line) will be ignored, we need to APPENDIX H

// close the dimension here.

BM_VECTOR temp_v( ( ( (BM_LINE ^♦) next_edge) ->get_v() ) ^♦ side_vector) ;

// if the next line will not be ignored we can skip the current edge. if (temp_v.Len() > AUTODIM_DISTANCE_TOLERANCE _Sc (temp_v

% normal) < 0.0) {

// next edge is a line that will not be ignored.

// note that there is a problem, if this edge is a line, and one of the two lines

// (reme eber, next edge is a line) is adjacent to a bendline that

// the other is not adjacent to, we should close the dimension. if (! edge->is(BM_ENTITY_TYPE_LINE) ) return 1 ;

BM_3D_B0DY ^ab = edge->get_adj_body() ; BM_3D_BODY ^♦ab^ext = next_edge->get_adj_body() ; if (1 ab) { if ( ! ab_next) return 1 ; if (! ab_next->is(BMJ_NTITY_TYPE_BENDLINE) ) return 1 ; return 0 ;

} if (! ab_next) { if (! ab->is(BM_ENTITY TYPE_BENDLINE) ) return l ; return 0 ;

} if ((! ab->is(BM_ENTITY_TYPE_BENDLINE) ScSc ! ab_next->is(BM_ENTITY_TYPE_BENDLINE) ) | | ab == ab_next) return 1 ;

» '

// if next edge is not a line, we have to close the open dimension return 0 ;

}

/*

A l o c a l f u n c t i o n u s e d b y

BM_AD_bendlin : : comput e_BM_AD_flange_points (...) .

This function creates a dimension point in the most general

3 APPENDIX H form.

We assume that we are creating the dimension point for a line. This line could be adjacent to a bendline.

*/ static BM_AD_dim_corner ♦create_dim_point (double d /^♦ bendline arc outside length projected onto our plane ^♦/,

// these are basic parameters needed to draw any dimension point BM_POINT const- p /♦ point p in the flange dimension point ^♦/, BM VECTOR const- rv /^♦ reference vector to make sure point p in the flange dimension point is on the same plane as defined by the bendline dimension point

*/' BM_VECT0R const- opening_directιon_vector /^♦ for flange dimension point ^♦/, int opening_vector_reverse_bit,

BM_AD_bendline ^♦owner,

BM_3D_BODY ♦Body_required_visible,

BM_AD_dim_corner ♦♦prev /♦ previous flange dimension point in the list of flange dim points ^♦/,

BM_VECT0R const- side_vector, // these parameters are used when the line for which we are computing the dimension point is adjacent // to another bendline. double d_proj_side_vector /^♦ d projected onto the side-vector : angle_factor^d in compute_BM_AD_flange_points ( ... ) ^♦/,

BM_VECTOR Sc bend_arc_direction /♦ direction of the bend arc projected onto our plane ^♦/

/^♦ should be line->v ^♦ plane normal ^♦/,

BM_VECTOR _ bend_direction /* points to the direction in which the bendline "turns" ♦/

/^♦ depends on whether the bend angle is acute or not. ♦/ /^♦ if acute then tangent to the other end of bend arc ♦/ /^♦ if not acute, then an endpoint of the tangent approximation ^♦/

/^♦ basically this is the (p->p2) vector for the other bendline with respect to this flange ♦/,

BM_BENDLINE *adj_bendline /^♦ if the line is adjacent to a bendline this is a pointer to it ♦/

/^♦ this parameter is stored to avoid drawing some di enions twice ♦/,

BM_LINE const- adj_bend_p2_line /^♦ this is the line between p2 points of the bend-dimension ♦/

/^♦ points of the adjacent bendline with respect to this flange ^♦/

/^♦ note that our p2 point has to be on that line. ♦/ /* usually our p->p2 vector is perpendicular to this line, although not always ♦/ ) CΓΌ

APPENDIX H

// this our local copy of the adj-bendline-pointer BM_3D_BODY ♦adj^end = adj_bendline ;

BM_POINT p_local (p + rv) ;

// note : even if this line is adjacent to another bendline,

// but if the line we are creating the flange dimension point for is not perpendicular to

// the side-vector, we should not memorize that this flange dimension point is adjacent to

// a bendline, because the flange length dimension drawn at this flange dimension point

// is not parallel to the side-vector of the adjacent bendline with respect to this flange. if (adj_bendline _SL fabs (bend_arc_direction % side_vector) < (bend_arc_direction.Len0 - AUTODIM_DISTANCE_TOLERANCE) ) { // not perpendicular adj bendline = NULL ;

} ^"

// check if the line is attached to a bendline if (d <= AUTODIM_DISTANCE_TOLERANCE) (

// not attached to a bendline, or both thickness and bendline radius are 0.

// create a 1-point flange-dimension point. // note that body_required_visible points to the flange with respect to which

// the dimension point is computed. return new BM_AD_dim_corner (p local, owner, Body_required_ isible, o p e n i n g _ d i r e c t i o n _ v e c t o r , opening_vector_reverse_bit , prev, adj_bendline) ;

// note that since we have 3 points now, this line is adjacent to another bendline.

// compute pi. p_local is pi.

// compute p

BM_POINT dim_p (p_local + d_proj_side_vector^side_vector) ;

// compute p2 bend_arc_direction.set_length(d) ;

BM_POINT dim_p2 (p_local + bend_arc_direction) ;

// now the problem is that with respect to the adjacent bendline, our flange dimension points

// p and pi might not be on the outside - they could be on the inside. SOI

APPENDIX H

// the distance from p to p2 depends on that.

// however, we know the p2-line of the adjacent bendline (with respect to our flange) ,

// so we will compute the distance between that p2-line and dim_p2.

// we will use that to compute the correct p2 point for our flange dimension (note that

// our p2 point has to be on the p2-line of the adjacent bendline) . double distance ;

BM_distance_between_point_and_line (dim_p2, adj_bend_p2_line, -distance, (BM_P0INT ^♦) NULL) ; bend_direction.set_length(distance) ; dim_p2 = dim_p2 + bend_direction ;

// note that the body_required_visible pointer points not to this flange, but to the

// adjacent-bendline instead, ALWAYS, even if this flange dim point will not be marked as

// adjacent to another bendline. return new BM_AD_dim_corner(dim_p, p_local, dim_j?2, owner, adj_bend ? adj_bend : Body_required_visible, opening_direction_vector, opening_vector_reverse_bit, prev, adj bendline) ;

}

/*

This function computes flange dimensions points with respect to an adjacent flange.

Note : The reference point computed right in the beginning of the function is given in order to make sure we can handle any kind of adj-body surface.

The first problem is that we don't know whether the 2D-body of the 3D-version of the adj-body represents the OUTSIDE or INSIDE. We do know that the bendline dimension points for the bendline are OUTSIDE points. However, it could be that the adj -body represents the INSIDE explicitly.

In that case we use to reference point (which is on the outside surface of the adj-body) to compute a translation vector by which we will shift all point we are going to compute here.

Also we will use this translation vector to make sure the points pi and p2 of the flange dimension point are correct .

⁽if the adj-body represents inside, we compute dimension points really for outside and the line

(p->p2⁾ has to be constructed with care - it is shorter than APPENDIX H line (p->pl) ) .

Note : this function is very much like BMAPI's BM_2D_B0DY: :compute_outside_width( ... ) function, although we make more assumtions here.

*/

// flange dimension points we are go ng to compute will be stored here

BM_AD_dim_corner ^♦♦list_of_points ; // list_of_points is where we write the pointer to the

// next flange dimension point

BM_AD_dim_corner ^♦temp_dp ;

// these parameters are used as input to this function. double flange_length ; /^♦ it is important that the flange length be the correct flange length.^♦/

BM_POINT ref_point ,- /^♦ reference point, defines the right plane for flange dimension points ^♦/

BM_VECTOR side_vector ; /^♦ should be normalized ^♦/

// check which adjacent body we compute if (adj_body == adj_bodyl) { flange_length = fll ; list_of_points = _bodyl_points ; ref_point = bend_bodyl_point_left .p ; side vector = side vectorl ;

^{} " "} else if (adj_body == adj_body2) { flange_length = fl2 ; list_of_points = _body2_points ; ref_point = bend_body2_point_left.p ; side vector = side vector2 ;

} ^" else return 0 ; // not an adjcent body

// variables to store temporary values BM_VECTOR temp_v ; double temp_x ;

// parameters for fixing the reference point

BM_POINT rp (ref_point) ; // this a local copy of the reference point, we will modify it.

BM_VECTOR rv ; // a vector for translating flange dimension points. int ref_vector_valid = 0 ; // if this is TRUE, we need to translate all flange dimension points. 50i

APPENDIX H

// if TRUE, the only really tricky part is to construct the p2 point in the flange dimension point.

// qualification check BM_2D_BODY ♦adj_displayed_version ; if (parent-_>get_view_type () ) adj_displayed_version adj_body->get_current_3D_version() ; else adj_displayed_version = adj_body->get_flat 0 ; if (NULL == adj_displayed_version) return 0 ;

BM_SURFACE ♦surface = adj_displayed_version->get_surface 0 ; if (NULL == surface) return 0 ;

BM_LOOP ♦bloop = adj_displayed_version->get_bloop 0 ; if (NULL == bloop) return 0 ;

BM_EDGE ^Aedge ; // here we will store edges as we scan the loop. double distancel, distance2 ; // here we will store the distance of points we see double dl, d2 ; // here we store distance that has to be added because of a bend arc. for lines adjacent to a bendline. double tdl, td2 ; // here we store the total distance which is the sum of distanceX + dX.

BM_3D_B0DY *ab ; // here we store a pointer to a body that is adjacent to the current edge.

BM_BEND_0P ^bop ; // here we store a pointer to the bend op associated with the bendline that is

// adjacent to the edge we scan (valid only iff the body adjacent to the edge is a bendline) .

BM_VECTOR right_opening_direction ; // this is the right-side opening direction vector.

BM_VECTOR left_opening_direction ; // this is the left-side opening direction vector. int fl_dim_open = 0 ; // if this is TRUE, we have created a right-side flange dimension point, but

// no corresponding left-side dimension point. That is, the dimension is "open" .

BM_VECT0R bend_arc_direction ; // direction of the bend arc projected onto our plane

// should be line->v ^♦ plane_normal. Used when a line is adjacent to a bendline.

BM_VECTOR bend_direction ; // points in the direction in which the bendline "turns"

// depends on whether the bend angle is acute or not. // if acute then tangent to the other end of bend arc. // if not acute, then an endpoint of the tangent approximation.

// basically this is the (p->p2) vector for the other bendline with respect to this flange.

// this is used when a line is adjacent to a bendline, to compute the p2 point for

8 SOW

APPENDIX H

// the flange dimension point. BM_LINE adj_bend_p2_line ; // this is the left-right-p2 line, of the adjacent bendline,

// with respect to this flange.

// the flange dimension point, double bend_arc_outside_length_proj ; // bendline arc outside length projected onto our plane

/*

Handle surface types.

*/ if (surface->is (BM_TYPE_PLANE) ) goto handle_planes ;

/*

********** ********** ********** ********** ********** **********

********** ********** ********** temporary resctiction : we only handle surfaces that are planes.

TODO : handle other surface types.

********** ********** ********** ********** ********** **********

********** ********** **********

*/

// unknown surface return 0 ;

/*

Here we handle planes

*/ handle_ lanes :

BM_PLANE ^♦plane = (BM_PLANE ^♦) surface ;

// we assume that the reference point correctly defines the plane (which is parallel to our plane)

// that is the OUTSIDE plane.

// check if reference point is on our plane. If not (in that case our plane represents the INSIDE

// plane) compute a translation vector for moving from our plane to reference plane,

// and move the reference point to our plane. rv = plane->get_normal 0 ; rv.normalise 0 ; temp_v = rp - (plane->get_pt () ) ; temp_x = temp_v % rv ; if (fabs (temp_x) > AUTODIM_DISTANCE_TOLERANCE) {

// ok, need to translate all dimension points, reference point not on the plane. ref vector valid = 1 ; APPENDIX H rv. set_length(temp_x) ; rp = rp + (-rv) ;

} else rv = BM_null_vector ;

// compute right- and left-side opening vectors. right_opening_direction = side_vector ^♦ (plane->get_normal C right_opening_direction.normalise () ; left_opening_direction = - (right_opening_direction) ;

/ *

********** ********** ********** ********** **********

********** ********** ********** **********

Main portion of this function. Scan the bounding loop of the adj_body and construct dimension points.

Notice that we only need to scan the bloop since the edge that is the farthest away from the given line has to be in the bloop. ********** ********** ********** ********** **********

********** ********** ********** **********

Note : when the dimension list is open, the last right-side dimension point created must be different from the start-point of the current edge, at the beginning of the loop.

*/

// scan all edges in the bloop

// first, we have a problem where to start scanning.

// if we start from the first edge, we basically start from an arbitrary edge.

// this might not be good since we might create dimension points we don't really want.

// therefore we will find an edge that is adjacent to our bendline and start from that and

// finish when we get back to it.

BM_EDGE ^♦star^edge, ^♦stop_edge, ⁺next_edge ; for ⁽start_edge = bloop->get_first_edge () ,- start edge ; start_edge = (BM_EDGE ^♦) start_edge->next () ) { ^~" if ⁽start_edge->get_adj_body() == bend) break ; // edge adjacent to this bendline is a good start-edge else if ⁽! start_edge->is ⁽BM_ENTITY_TYPE_LINE) ) break ; // any non-line is a good start-edge else { // any line that will be ignored is also a good place to start temp_v = ( ^{( (}BM_LINE *) start_edge) -_>get v()) ♦ side_vector ; ^~

10 APPENDIX H

// in these two cases this line will be ignored if (temp_v.Len() <= AUTODIM_DISTANCE_TOLERANCE) break ; // line parallel to the side-vector if ( (temp_v % plane->get_normal () ) > 0.0) break ; // material "above" the line

if (NULL == start_edge) {

♦list_of_points = NULL ; return 1 ;

}

// this is the main loop.

// initially we set stop_edge to NULL, we would like to set it to start-edge, but we cannot,

// because in that case we would never enter the loop, therefore we set it to NULL and right

// after we enter the loop, we set it to start_edge.

// (this is important in the case when the loop consists of only one circle) . stop_edge = NULL ; for (edge = start_edge ; edge != stop_edge ; edge = next_edge)

{ stop_edge = start_edge ;

// compute the next edge we will scan next_edge = (BM_EDGE ^♦) edge->next() ; if (NULL == next_edge) next_edge bloop->get_first_edge () ; switch (edge->get_type 0 ) { case BM_ENTITY_TYPE_LINE : { // compute distance for start- and end-points

// first, we want to ignore lines that are parallel to the side-vector.

// we also want to ignore lines that have material "above" them when looking in the

// side-vector direction (because they cannot have to largest width, since material is

// "above" them, there must be some other edge bounding the body from above) . temp_v = ( ( (BM_LINE ♦) edge) ->get_v() ) ♦ side_vector ; if (temp_v.Len() <= AUTODIM_DISTANCE_TOLERANCE) continue ; // line parallel to the side-vector if ( (temp_v % plane->get normal () ) > 0.0) continue ; // material "above" the line

// note that this leaves only lines whose direction-vector goes to the "left"

11 APPENDIX H // of the side-vector.

// we need to check if this line is adjacent to a bendline ab = edge->get_adj_body() ;

// check if this line is adjacent to a valid bendline (ie. one that is not the

// bendline associated with this BM_AD_bendline obj ect ) . dl = d2 = bend_arc_outside_length__proj = 0 . 0 ; i f ( a b Sc Sc a b ! = b e n d Sc Sc ab- >is (BM_ENTITY_TYPE_BENDLINE ) ) {

// check if the bendline is being bent bop = ( (BM_BENDLINE ^♦ ) ab) - >get_bend_op 0

; if ( (parent->get_view_type () ) /^♦ have to have a 3D view * I ScSc

( ( B M _ B E N D L I N E ^♦ ) ab) ->get_bending_status 0 ScSc bop) {

// a problem : it could be that the bendline is not perpendicular to the side-vector.

// in that case we need to project the bend-arc-approximation-value onto the side-vector.

// for example, if the bendline is parallel to the side-vector, we should not add

// the approximation value at all, since it plays no role in the width computation. temp_x = side_vector ^Λ ( ( (BM_LINE ♦) edge) ->get_v() ) ; // temp_x is "angle_factor" if (temp_x > PI_over_2) temp_x = PI - temp_x ; temp_x = sin(temp_x) ; switch (bop->get_type 0 ) { case BM_TYPE_BEND_OP_REGULAR : bend_arc_outside_length_proj = ( (BM_BEND_OP_REGULAR ♦) bop⁾ ->compute_3D_outside_approx_per_side () ; d 1 temp_x^bend_arc_outside_length_proj ; d2 = dl ; // d2 is the same as dl ! break ; // TODO : implement more bend operations here. default : ; // unknown bend op }

// compute some needed parameters // compute : direction of the bend arc projected onto our plane

12 APPENDIX H *)

dy,

int to anything we don't want to.

// handle two cases separately - depending on whether the dimension list is open or not. if (fl_dim_open) {

// if the dimension is open, we don't need to check the start-point of the line,

// since it is at the flange length distance anyway.

//

// first we will check if the line is perpendicular to the side-vector if (fabs (side_vector % ( (BM_LINE ^♦) edge) ->get_v() ) > AUTODIM_DISTANCE_TOLERANCE) {

// not perpendicular. // this line has to be "coming down" (because it is not perpendicular to the

close the dimension here by creating

// a left-hand-side dimension point

13 APPENDIX H from the end-point of the line .

// if the next edge is a line , we can iust continue . i f

(leave_dimension_open (edge , bloop , side_vector , plane - >get_normal 0 ) ) break ;

) , s , ♦ )

// this is a small trick that will save us some time.

// if the next edge is a line that will be ignored, we can skip is right now here. i f

(next_edge->is (BM_ENTITY_TYPE_LINE) ScSc next_edge != stop_edge) { next_edge = (BM_EDGE ^♦) edge->next() ; if (NULL == next_edge) next_edge = bloop- >get first edge ( ) ;

break ;

} if (! compute_distance_from_ref_j>oint (rp, side_vector, edge->get_edge_startpoint () , distaneel)) continue ,- if ⁽! compute_distance_from_ref_point (rp, side_vector, edge->get_edge_endpoint () , distance2) ) continue ; tdl = distaneel + dl ; td2 = distance2 + d2 ;

// check if the start-point is a flange dimension point if (fabs (tdl - flange length) _<= AUTODIM_DISTANCE_TOLERANCE) {

// check if the second point is a flange dimension point

14 3

APPENDIX H if (fabs(td2 - flange_length) > AUTODIM_DISTANCE_TOLERANCE) {

// only start-point is a flange dimension point.

// create a left-right (ie. in both directions) flange dim point and continue. i f ( t e m p _ d p create_dim_point (bend_arc_outside_length_proj ,

( ( B M _ L I N E ^♦ ) edge) -_>get_startpt () , rv, right_opening_direction, 1, this, adj_body, list_of_points, side_vector, dl, bend_arc_direction, bend_direction, (BM_BENDLINE ^♦) ab, adj_bend_p2_line) ) { l i s t _ o f _ p o i n t s Sc (temp dp->next) ;^■

} break ;

}

// else open the dimension list // create right flange dimension point from start-point i f ( t e m p _ d p create_dim_point (bend_arc_outside_length_proj ,

( (BM_LINE *) edge) ->get_startpt () , rv, right_opening_direction, 0, this, adj_body, list_of_points, side_vector, dl, bend_arc_direction, bend_direction, (BM_BENDLINE ^♦) ab, adj_bend__p2_line) ) { list_of_points = _(temp_dp->next) ;

// check if we can leave the dimension list open and continue i f

(leave_dimension_open(edge,bloop,side_vector,plane->get_normal () ) ) fl_dim_open = 1 ; break ;

} // create left flange dim point from end-point . i f ( t e m p _ d p create_dim_point (bend_arc_outside_length_proj ,

((BM_LINE ^) edge) ->get_endpt 0 , rv, left_opening_direction, 0, this, adj_body, list_of_points, side_vector, dl, bend_arc_direction, bend_direction, (BM_BENDLINE ♦) ab,

15 APPENDIX H adj_bend_p2_line) ) { list_of_points = _(temp_dp->next) ;

// ok, close the dimension list. fl_dim_open - 0 ;

// this is a small trick that will save us some time.

// if the next edge is a line that will be ignored, we can skip is right now here. if (next_edge->is(BM_ENTITY_TYPE_LINE) __ next_edge != stop_edge) { next_edge = (BM_EDGE ^♦) edge->next()

/" if (NULL == next_edge) next_edge = bloop->get_first_edge 0 ; break ;

} else if (fabs(td2 - flange_length) >

AUTODIM_DISTANCE_TOLERANCE) {

// the start-point is not a flange dimension point

// if the end-point is also not a flange dimension point, we can continue. continue ;

}

// ok, only the end-point is a flange dimension point.

// two possibilities, if we can leave the dimension list open (right now dimension is

// closed!) we don't even have to open it, we can just continue. i f

(leave_dimension_open(edge,bloop, side_vector,plane->get normal () ) ) break ;

// second possibility, now we have to create a left-right flange dimension point.

// dimension list stays closed. i f ( t e m p _ d p create_dim__point (bend_arc_outside_length__proj ,

⁽ (BM_LINE ^♦) edge) ->get_endpt 0 , rv, right_opening_direction, 1, this, adj__body, list__of_points, side_vector, dl, bend_arc_direction, bend_direction, (BM_BENDLINE ♦) ab, adj_bend_p2_line) ) { list_of_points = _(temp_dp->next) ;

16 APPENDIX H

// this is a small trick that will save us some time.

// if the next edge is a line that will be ignored, we can skip is right now here. if (next_edge->is(BM_ENTITY_TYPE_LINE) _Sc next_edge != stop_edge) { next_edge = (BM_EDGE ^♦) edge->next() ; if (NULL == next_edge) next_edge bloop->get_first_edge 0 ;

break ; case BM_ENTITY_TYPE_ARC : {

// note : we can always ignore the start-point of the arc !

// because of the way we keep track of the open dimension list

// make sure the -dimension list will be closed. fl_dim_open = 0 ;

// first, compute start- and end-angles of the arc with respect to the side vector.

// ie. convert start- and end-angles so as if the side-vector was the 0-angle vector. double startang = ( (BM_ARC ^♦) edge) ->get_startang() ; double endang = ( (BM_ARC ^♦) edge) ->get_endang()

; temp_v = side_vector* ( ( (BM_ARC ^♦) edge) ->get_vx() ) ; double delta_angle = (temp_v.Len() ) / ( ( (BM_ARC ♦) edge) ->get_rad() ) ; if (delta_angle < -1.0) delta_angle = -1.0 ; else if (delta_angle > 1.0) delta_angle = 1.0 delta_angle = asin(delta_angle) ; i f ( f a b s ( ( ( ( B M _ A R C ^♦ ) edge) ->get_normal O ) %temp_v) > 0.0) { startang += delta_angle ; endang += delta_angle ; if (startang > PI_times_2) startang -= PI_times_2 ; if (endang > PI_times_2) endang PI_times_2 ; else { startang -= delta_angle ; endang -= delta_angle ;

17 APPENDIX H if (startang < 0.0) startang += PI_times_2

; if (endang < 0.0) endang += PI times 2 ;

}

// compute the extreme point

BM_POINT arc_extreme_point ;

// 0-angle point is the extreme point if (endang < startang) arc_extreme_point = ( (BM_ARC ♦) edge) ->get_center () + ( ( (BM_ARC ♦) edge) ->get_rad() ) ^♦side_vector ; else arc_extreme_point = ^♦ ( ( (BM_ARC ♦) edge) ->get_edge_endpoint () ) ;

// check the distance of the extreme point if (! compute_distance_from_ref_point (rp, side_vector, Scarc_extreme_point, distaneel)) continue ; if (fabs (distaneel - flange_length) <= AUTODIM_DISTANCE_TOLERANCE) {

// if we have the 0-angle point, create a left-right flange dimension point.

// otherwise if next line is a non-ignored line, don't do anything,

// otherwise create a left-right flange dimension point. if (endang >= startang _SL leave dimension open(edge,bloop,side vector,plane->get normal 0 ) )

{ continue ;

}

// create a left-right flange dimension point , dimension list stays closed .

// this is a 1 -point flange dimension point . if (temp_dp = new BM_AD_dim_corner (rv + arc_extreme_point , this , adj_body, r ight_opening_di rect ion , 1 , list_of_points , NULL) ) ( list_of_points = Sc (temp_dp- >next) ;

) ' ' break ; case BM_ENTITY_TYPE_CIRCLE :

// circles are a much simpler version of the arc case.

{

BM_P0INT circle_extreme_point = ( (BM CIRCLE ♦)

18 APPENDIX H edge) ->get_center 0 + ( ( (BM_CIRCLE ^♦) edge) ->get_rad() ) ^♦side_vector

// we won't check the distance since it must be a flange dimension point

// create a left-right flange dimension point. // this is a 1-point flange dimension point, if (temp_dp = new BM_AD_dim_corner (rv + circle_extreme_point, this, adj__body, right_opening_direction , 1 , list_of_points, NULL)) { list of points = _(temp dp->next) ;

> ' ^{" "} break ; case BM_ENTITY_TYPE_ELLIPSE : // TODO : implement ellipse here, should be similar to the way arc is handled, continue ; break ; default : continue ; // unknown edge type

}

/*

Ok, all done.

*/

♦list_of_points = NULL ; return 1 ; }

/*

This function computes bendline dimension points for one

BM_AD_bendline structure.

Note : pi is on the plane of adjacent body for which the flange length is drawn. p is the centerpoint . p2 is the point on the side of the bendline.

Note : in this function we always compute bend-dimension points with respect to adj -bodyl and adj-body2. In might that we need to know the visibility of

19 APPENDIX H these points in the future, although this bendline might have 1 or 0 adjacent bodies.

*/ int BM_AD_bendline: :compute_BM_AD_bendline (void) double metal_thickness ; // thickness of the sheetmetal BM_2D_BODY ♦bendline_displayed_version ; // 3D-version of the bendline

BM_VECT0R temp_v ; // vector used for very short periods for storing temporary stuff

// initally mark the bendline data as invalid, in case we fail here. data_valid = 0 ;

// qualification check if (NULL == bend) return 0 ; if (NULL == bend->get_adj_list 0 ) return 0 ; if (NULL == bend->get_part () ) return 0 ,-

// this bendline has to have a 3D-version if (parent->get_view_type () ) { if (NULL == (bendline_displayed__version bend-_>get_current_3D_version0 ) ) return 0 ;

(bendline_displayed_version

metal_thickness = (bend->get_part 0 ) ->get_metal_thickness () ;

/* ********** ********** ********** ********** ********** ********** ********** ********** ********** this is a temporary measure, we only process regular bendlines. ie. ignore conic bendlines. 03/16/96. KK. ^♦/ if (bend->get_bending_status 0 __ bend->get_bend_op 0 _Sc

! (bend->get_bend_op() ) ->is (BM_TYPE_BEND_OP_REGULAR) ) return 0 ; /* ********** ********** ********** ********** ********** ********** ********** ********** ********** *

// get the number of adjacent bodies. adj_bodyl = (bend->get_adj_list () ) ->f ind_lowest_id_adj_body () a d j _ b o d y 2 ⁽bend->get_adj_list () ) ->f ind_second_lowest_id_adj_body () ; num_of _adj acent bodies = (adj bodyl ? 1 : 0) + (adj bodv2 ? 1 : 0) _; ^" -

/*

20 APPENDIX H Compute side vectors.

* / if (adj_bodyl) ( i f ( ! bendline_displayed_version->compute_vector_perp_towards_adj_body( adj_bodyl, side_vectorl) ) return 0 ; side_vectorl.normalise() ; if (adj_body2) { i f ( ! bendline_displayed_version->compute_vector_perp_towards_adj_body( adj_body2, side_vector2) ) return 0 ; side vector2.normalise() ;

} ^"

/*

Compute bendline dimension points.

*/ if (! bend->get_bending_status () /^♦ bendline not bent ♦/ | |

! parent->get_view_type 0 ^■/^♦ flat view ^♦/) {

// note : here we don't know whether the dimension point is OUTSIDE or INSIDE.

// this has to be taken into account later.

// however, it should work either way. bend_bodyl_point_lef t . num_of_points bend_bodyl_point_right .num_of_points = 1 ; bend_bodyl_point_lef t . can_reverse_opening_dir bend_bodyl_point_right .can_reverse_opening_dir = 0 ; b e n d _ b o d y l _ p o i n t _ l e f t . p (bendline_displayed_version->get_current_center_line 0 ) ,get_start pt 0 ; b e n d _ b o d y l _ p o i n t _ r i g h t . p =

(bendline_displayed_version->get_current_center_line () ) .get_endpt ()

// note opening directions, left will get start-point, right will get end-point, ie. vector goes left->right. bend_bodyl_point_right . opening_direction (bendline_displayed_version->get_current_center_line () ) .get_v() ;

(bend_bodyl_point_right . opening_direction) . normalise ( ) ; bend_bodyl_point_left . opening_direction - (bend_bodyl_point_right .opening_direction) ;

// points with respect to both adjacent bodies are the same bend_body2_point_lef t . num_of_points bend_body2_point_right .num_of_points = 1 ;

21 APPENDIX H bend_body2_point_lef t . can_reverse_opening_dir bend_body2_point_right . can_reverse_opening_dir = 0 ,- bend_body2_point_right . opening_direction bend_bodyl_point_right . opening_direction ; bend_body2_point left . opening_direction bend_bodyl_point_lef t .opening_direction ; bend_body2_point_left .p = bend_bodyl_point_lef t .p ; bend_body2_point_right .p = bend_bodyl_point_right .p ; else (

// first, get some bending parameters B M _ S U R F A C E ^♦ s u r f a c e bendline_displayed_version->get_surface () ; if (NULL == surface) return 0 ; BM_BEND_OP ^♦bop = bend->get_bend_op0 ; if (NULL == bop) return 0 ; double bend_angle = bop->get_bend_angle20 ; double bend_arc_angle = PI - bend_angle ; double half_bend_arc_angle = bend_arc_angle/2.0 ; switch (surface->get_type () ) I case BM_TYPE_CYLINDER : { // note that the bend op can be both regular and conic

BM_CYLINDER ^♦cylinder = (BM_CYLINDER ^♦) surface

; double radius = cylinder->get_rad() ; double outside_radius = radius ; if (bendline_displayed_version->get_sense () ) outside_radius += metal_thickness ;

// set some common values bend_bodyl_j?oint_left .can_reverse_opening_dir = bend_bodyl_point_right.can_reverse_opening_dir = 0 ; bend_bodyl_point_right .opening_direction (bendline_displayed_version->get_current_center_line () ) .get_v() ;

(bend_bodyl_point_right.opening_direction) .normalise () ; bend_bodyl_point_left .opening_direction - (bend_bodyl_point_right.opening_direction) ; i f ( o u t s i d e _ r a d i u s < =

AUTODIM_DISTANCE_TOLERANCE) {

// special case : outside radius 0 - trivial dimension point. bend_bodyl_point_lef t . num_of _points bend_bodyl_point_right .num_of_points = 1 ; bend_bodyl_poi nt_l e f t . p ⁽bendline_displayed version- >get current center line()) .qet start pt() ; ^~ - b e n d_b o dy 1 _p o i n t __ r i g h t . p

22 51 Ϋ

APPENDIX H

(bendline_displayed_version->get_current_center_line ) .get_endpt () ; bend_body2_point_lef t .num_of_points bend_body2_point_right .num_of_points = 1 ,- bend_body2_point lef t . can_reverse_opening_dir bend_body2_point_right . can_reverse_opening_dir = 0 ; bend_body2_point_right . opening_direction = bend_bodyl_point_right .opening_direction ; bend_body2_point_lef t . opening_direction= bend_bodyl_point_lef t .opening_direction ; bend_body2_point_le f t . p bend_bodyl_point_lef t . p ; bend_body2_point_right . p = bend_bodyl_point_right . p ; else { bend_bodyl_point_lef t . num_of _points bend_bodyl_ρoint_right .num_of_points = 3 ;

// compute points pi for adj_bodyl bend_bodyl_point_lef t . pl

(bendline_displayed_version->get_current_center_line () ) .get_start pt 0 ; bend_bodyl_point_right . pl

( bendl ine_displayed_version->get_current_center_lme ( ) ) .get_endpt ( )

;

// the problem now is that it could be that these centerline points (which are

// on the surface of the bendline) , could be only inside surface of the bendline. i f

(bendline_displayed_version->get_sense 0 ScSc metal_thickness > AUTODIM_DISTANCE_TOLERANCE) { temp_v = cylinder->get_vx0 ,- temp_v.set_length(metal_thickness) ; bend_bodyl_point_left .pl bend_bodyl_point_left.pl + temp_v ; bend_bodyl_point_right.pl bend bodyl -point right.pl + temp v ;

} ^"

// a break in the computation of adj_bodyl.pi .

// store current pi values for adj_body2 as well, this will save some time.

//

// note that we should not really compute adj_body2 points unless adj_body2 pointer

23 APPENDIX H

// is not NULL. However, this point (adj_body2.pl) is very useful for computing

// adj_bodyl.p2 (if angle is acute) . so we compute it anyway. bend_body2_point_lef t . pl bend_bodyl_point_left.pl ; bend_body2_point_right . pl bend_bodyl_point_right.pl ;

// finish adj_bodyl.pl. rotate points pi into their final correct location.

// note we rotate in the negative direction since adj_bodyl is the smallest-index

/ / adj acent body as i s "to-the-left-of-the-bendline-centerline" , which means

// rotating to the left.

// first, compute the right vector for the rotation line.

BM_VECTOR rot_v(cylinder->get_v() ) ; i f ( r o t _ v % ( (bendline_displayed_version->get_current_center_line () ) .get_v() ) < 0.0) { rot v.reverse () ;

} ^"

BM_rotate_point_around_line (_ (bend_bodyl_point_lef t .pi) , cylinder- >get_ptl () , rot_v,

-half_bend_arc_angle) ;

BM_rotate_point_around_line (Sc (bend_bodyl_point_right .pi) , cylinder- >get_ptl () , rot_v,

-half_bend_arc_angle) ;

// finish also adj_body2.pl. rotate points pi into their final correct location.

BM_rotate_point_around_line (Sc (bend_body2_point_lef t .pi) , cylinder- >get_ptl () , rot_v, half_bend_arc_angle) ;

BM_rotate_point_around_line (Sc(bend_body2_point_right .pi) , cylinder- >get_ptl () , rot_v, half_bend_arc_angle) ;

{ ^♦)

24 APPENDIX H else if (bop->is ( BM_TYPE_BEND OP_CONIC) )

{

// TODO : implement conic bendline here return 0 ;

} temp_v = -x^side_vectorl ; bend_bodyl_po i nt_l e f t . p bend_bodyl_point_left.pl + temp_v ; bend_bodyl_point_right . p bend_bodyl_point_right.pl + temp_v ;

// note : adj_bodyl.p2 and adj_body2.p, adj_body2.p2 are yet to be computed.

// now there are two cases : 1) the bend arc angle is acute (ie. no more than 90 degrees),

// 2) the bend arc angle is more than 90 degrees . if (bend_arc_angle > (PI_over_2 + BM_ANGLE_TOLERANCE) ) { // use tangent approximation

// we will finish computing adj_bodyl . adj_bodyl.p2 is left.

// idea : notice that adj_bodyl is to the left of the bendline-centerline

// because it is the lowest-index neighbor. Therefore the cross-product

// of centerline-v by temp_v points towards the cylinder (actually it is tangent

// to the cylinder) .

B M _ V E C T 0 R tangent_v(bend_bodyl_point_right .opening_direction ♦ temp_v) ; tangent_v.set_length(x) ; bend_body l_po int le f t . p2 bend_bodyl_point_left .p + tangent_v ; bend_bodyl_point_right .p2 bend_bodyl_point_right . p + tangent_v ;

// note adj_body2.pl is already computed, first set some common values. bend_body2_point_lef t . num_of _points = bend_body2_point_right .num_of_points = 3 ; bend_body2_point_left . can_reverse_opening_dir bend_body2_point_right . can_reverse_opening_dir = 0 ; bend_body2_point_right . opening_direct i on bend_bodyl_point_right .opening_direction ; bend_body2_po i nt_l e f . open i ng_d i re c t ion bend_bodyl_point_lef t . opening_direction ;

25 I

APPENDIX H

// now compute adj_body2.p temp_v = -x^side_vector2 ; bend_body 2 _po i n t _ 1 e f t . p bend_body2_point_left.pl + temp_v ; bend_body 2_po int_r ight . p bend_body2_point_right.pl + temp_v ;

// finally, compute adj_body2.p2

// note the order in the cross-product ! tangent_v = temp_v ♦ bend_body2_point_right . opening_direction ; tangent_v.set_length(x) ; bend_body 2_po i n t _ 1 e f t . p2 bend_body2_point_left .p + tangent_v ; bend_body2_point_r ight .p2 bend_body2_point_right .p + tangent_v ; else { // use intersection approximation // note : adj_body2.p2 is the same as asj_bodyl.pl . bend_body l_po in t_l e f t . p 2 bend_body2_point_left.pl ; bend_bodyl_point_right .p2 bend_body2_point_right.pl ;

// when using intersection approximation, dimensions points with respect to

// both adjacent bodies are the same, so we just copy them. bend_body2_point_left .num_of_points = bend_body2_point_right.num_of_points = 3 ; bend_body2_point_left . can_reverse_opening_dir bend_body2_point_right . can_reverse_opening_dir = 0 ; bend_body2_point_right . opening_direction bend_bodyl_point_right . opening_direction ; bend_body2_point_lef t . opening_direction bend_bodyl_point_lef t . opening_direction ;

// note we are switching points bend_body2_point_lef t . p2 bend_bodyl_point_lef t .pi bend_body2_point_r ight .p2 bend_bodyl_point_right .pi bend_body2_point_left . p bend_bodyl_point_left .p ,- bend_body2_point_right .p bend_bodyl_point_right -p

26 APPENDIX H

}

} break ; case BM_TYPE_CONE break ; default : ; // unknown or illegal surface ; }

/*

Compute flange length for each of the adjacent bodies. */ if (adj_bodyl) {

// if the flange length computation fails, pretend the adjacent body does not exist, for simplicity. i f- ( ! bend->compute_flange_length(parent->get_view_type () ,adj_bodyl, fll) )

{ adj_bodyl = NULL ;

--num of adjacent bodies ;

} ^{~ ~} if (adj_body2 ) { i f ( ! bend- _>compute flange length (parent - >get_view_type ( ) , adj_body2 , f 12 ) )

{ adj_body2 = NULL ;

--num of adjacent bodies ;

» ' ^" ~ ^"

/*

Done. Mark data as valid. Once bendline dimension points have been computed for all bendlines, we can compute flange dimension points for all bendlines as well .

*/ data_valid = 1 ;

// before we can return, we have to check if bodyl, 2 left (right) bendline dimension points

// are the same.

// note the same means not the same pointers, but the contents are equivalent . if (adj_bodyl __ adj_body2) { bodyl2_left_the_same = (bend_bodyl_point_left

27 APPENDIX H bend_body2_point_left) ; bodyl2_right_the_same (bend_bodyl_point_right bend_body2 point_right) ;

/*

Important. If any of the bend-dimension points contains only

1 point (point p) , then the body_required_visible pointer should point to an adjacent body because there really is no bendline to see .

/ if (1 == bend_bodyl_point_lef t . num_of _points) bend_bodyl_point_left .body_required_visible = adj_bodyl ; if (1 == bend_bodyl_point_right . num_of _points ) bend_bodyl_point_right .body_required_visible = adj_bodyl ; if (1 == bend_body2_point_lef t . num_of _points ) bend_body2_point_left .body_required_visible = adj_body2 ; if (1 == bend_body2_point_right . num_of _points ) bend_body2_point_right .body_required_visible = adj_body2 ; return 1 ;

BM_AUTO_DIMENSION class stuff.

********** ********** ********** ********** **********

********** ********** ********** ********** */

SOLID HEURISTIC :

Check if we can ignore some points in solid mode. in solid mode we will be using this heuristic in order to save time. when two bendlines are using equivalent flange-dimension points for (potentially) displaying flange-length, we will only compute the visibility information only for the bendline (of these two bendlines) which has the higher idx.

*/ void BM_AUTO_DIMENSION: :compute_points_to_ignore_in_solid (void)

28 APPENDIX H int i ;

BM_AD_dim_corner ^♦temp_point ; for (i = 0 ,- i < num_of_bendlines ; i++) { if (! bends [i] ->data_valid) continue ; if (bends [i] ->adj_bodyl) { for (temp_point = bends [i] ->bodyl_points temp_point ; temp_point = temp_point->next) { if (NULL == temp_point->adj_bendline) continue if ( (temp_point->adj_bendline) ->get_idx() > (bends [i] ->bend) ->get_idx() ) continue ; i f

(check_two_bendlines_adjacent_to_same_flange (bends [i] ->bend,bends [i] ->adj_bodyl, temp_point->adj_bendline) ) { temp_point->ignore_in_solid_mode = 1 ;

if (bends [i] ->adj_body2) { for (temp_point = bends [i] ->body2_points ; temp_point ; temp_point = temp_point->next) { if (NULL == temp_point->adj_bendline) continue if ( (temp_point->adj_bendline) ->get_idx() > (bends [i] ->bend) ->get_idx() ) continue ; i f

(check_two_bendlines_adjacent_to_same_flange (bends [i] ->bend,bends [i] ->adj_body2, temp_point->adj_bendline) ) { temp_point->ignore_in_solid_mode = 1 ;

points to ignore in solid computed = 1 ;

} ^{" "}

/*

This function will build auto dimension bendline and point data structures .

This is the main function for computing dimension points. */ int BM_AUTO_DIMENSION: :compute_auto_dimension_data (void)

29 APPENDIX H

int l ;

// have to have a document if (NULL == doc) return 0 ;

// if no part, try to get a new part from the document if (NULL == part) { set_part (doc->get_part () ) ; if (NULL == part) { dirty = 0 ; return 1 ; } points_to_ignore_in_solid_computed = 0 ;

get view type if (view) { view_type = view->get_current_view0 else {

// we are screwed return 0 ;

}

I*

Compute all bendline dimension points for all bendlines. *I for (i = 0 ; i < num_of_bendlines ; i++) { bends [i] ->compute BM AD bendlineO ;

}

I*

Compute flange dimension points for every bendline.

*/ for (i = 0 ; i < num_of_bendlines ; i++) { if (bends [i] ->data_valid) { i f ⁽ b e n d s [ i ] - > a d j _ b o d y l ) bends [i] ->compute_BM_AD_f lange_points (bends [i] -_>adj_bodyl) ; i f ⁽ b e n d s [ i ] - > a d j _ b o d y 2 ) bends [i] - >compute_BM_AD_f lange_points (bends [i] -_>adj_body2) ;

dirty = 0 ; return 1 ;

30 APPENDIX I

//////////////////////////////////////////////////////////////

// //

// Example of some basic auto-dimensioning functions, //

// constructors, destructors and some other basic //

// functions. //

// //

//////////////////////////////////////////////////////////////

#include "stdafx.h" #include "BendCAD.h" #include "BendCADDoc.h" #include "BendCADViewPart.h"

#include "AUTODIM.HXX"

/*

********** ********** ********** ********** **********

********** ********** ********** **********

BM_AD_dim_corner class stuff.

********** ********** ********** ********** **********

********** ********** ********** ********** */

BM AD dim corner: :BM AD dim_corner (void)

{ ^{~ "} visible = -1 ; num_of_points = 0 ; bend = NULL ; next = NULL ; adj_bendline = NULL ; body_required_visible = NULL ; _} ignore-in-solid-mode = 0 ;

/*

This constructor is for creating a dimension point that has only one point (point p) in it. */

BM_AD_dim_corner : : BM_AD_dim_corner (BM_POINT const- P, BM_AD_bendline ^♦B, BM_3D_BODY ^♦Body_required_visible,

BM_VECTOR constSL OD, int RB, BM_AD_dim_corner ♦♦prev, BM_BENDLINE ^♦adj_bend) visible = -1 ; num_of_points = 1 ; P = P ; APPENDIX I bend = B ; opening_direction = OD ; can_reverse_opening_dir = RB ; next = NULL ; body_required_visible = Body_required_visible ; adj_bendline = adj_bend ; ignore_in_solid_mode = 0 ,- if (prev) ⁺prev = this ;

/*

This constructor is for creating a dimension point that has 3 points in it. */

BM_AD_dim_corner: :BM_AD_dim_corner(BM_P0INT const- P, BM_POINT constSc PI, BM_P0INT constSc P2,

BM_AD_bendline ^♦B, BM_3D_B0DY ^♦Body_required_visible, BM_VECTOR constSc OD, int RB, BM_AD_dim_corner ♦♦prev,

BM BENDLINE ^♦adj bend)

{ ^" visible = -1 ; num_of_points = 3 ; p = P ; pi = PI ; p2 = P2 ; bend = B ; opening_direction = OD ; can_reverse_opening_dir = RB ; next = NULL ; body_required_visible = Body_required_visible ; adj_bendline = adj_bend ; ignore_in_solid_mode = 0 ; if (prev) ^♦prev = this ; 51*

APPENDIX I

To assign one dimension point to another dimension point. This function copies the contents exactly, not sure if it is a very useful function.

7

can_reverse_opening_dir = given_point.can_reverse_opening_dir

; visible = given_point.visible ; screen_x = given_point.screen_x ; screen_y = given_point.screen_y ; z_buf_value = given_point.z_buf_value ; ignore_in_solid_mode = given_point.ignore_in_solid_mode ;

/* this function will return TRUE iff the given point is, for the purpose of drawing dimensions, the same as this point.

Two dimension points points are equivalent iff the point p of one of them is on the line defined by the line (point p, opening direction) of the other dimension point, and their opening directions match. here we depend on the fact that opening vectors are normalized. it returns 2 iff both points p match, otherwise 1. */ int BM_AD_dim_corner: :operator== (BM_AD_dim_corner const- another dim point)

{ ^{" "} int ret_val = 1 ; double k ; if (num_of_points < 1 | | another_dim_point.num_of_points < 1) return 0 ; APPENDIX I

// point has to be on the line if ( ! BM_is_point_on_line (p, opening_direction, another_dim_point.p, _k) ) return 0 ; if (fabs(k) <= BM_PRECISION) ret_val = 2 ; is %

be

| j

return 0 ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

BM_AD_bendline class stuff.

********** ********** ********** ********** **********

********** ********** ********** ********** */

BM_AD_bendline : :BM_AD_bendline (BM_AUTO_DIMENSION ^♦owner, BM BENDLINE ^♦bendline)

{ ^" parent = owner ; bend = bendline ; data_valid = 0 ; // data not valid - no data computed yet bend_bodyl_point_left .bend = bend_bodyl_point_right .bend = this ; bend_body2_point_left .bend = bend_body2_point_right .bend = this ; bend_bodyl_point_left . body_required_visible bend_bodyl_point_right .body_required_visible = bend ; bend_body2_point left . body_required_visible bend_body2_point_right .body_required_visible = bend ; bodyl_points = NULL ; body2_points = NULL ; APPENDIX I }

BM AD bendline : : ~BM AD bendline (void)

{ ^{" "}

BM_AD_dim_corner ^♦temp_point, ^♦temp_p ;

// delete lists of points for (temp_point = bodyl_points ; temp_point ; temp_point = temp_p) ( temp_p = temp_point->next ; delete temp point ;

} for (temp_point = body2_points ; temp_point ; temp_point = temp_p) { temp_p = temp_point->next ; delete temp point ;

}

/*

********** ********** ********** ********** **********

********** ********** ********** **********

BM_AUTO_DIMENSION class stuff.

********** ********** ********** ********** **********

********** ********** ********** ********** */

BM AUTO DIMENSION: :BM AUTO DIMENSION(CBendCADDoc ^♦BendCAD doc)

{ ^" doc = BendCAD_doc ; v i e w = ( C B e n d C A D V i e w P a r t ^♦ ) doc->get_view(BENDCAD_VIEW_TYPE_PART) ; dc = NULL ; part = NULL ; dirty = 1 ; view_type = 1 ; // 3D view reset_view_to_false_once_drawn = 0 ; points_to_ignore_in_solid_computed = 0 ; num_of_bendlines = 0 ; bends = NULL ;

// initialize show states for dimension info

5 APPENDIX I m_bShowFlangeDim = TRUE; m_bShowBendDim = TRUE; m_bShowPartDim = FALSE;

// initialize Auto-Dim colors m_crFlange = RGB (0, 0, 255) ; m_crBend = RGB (0, 0,255) ; m_crPart = RGB (0, 0, 255) ;

// initialize font pointers m_pFlangeFont = NULL; m_pBendFont = NULL; m_pPartFont = NULL;

// initialize line and arrow styles m_nLineStyle = PS_SOLID; m_nLineWeight = 1; // in pixels m_nArrowStyle = 1,- m_nArrowAngle = 30; // in degrees m_nArrowLength = 15; // in pixels

// initialize dirty region pointer pDirtyRegion = NULL;

BM AUTO DIMENSION: :-BM AUTO DIMENSION(void)

{ ^{" " "} delete contents () ;

}

/* destroy the contents of this Auto-Dimension object, */ void BM_AUTO_DIMENSION: :delete_contents (void) int i ; if (bends) { for (i = 0 ; i < num_of_bendlines ; i++) ( if (bends [i] ) delete bends [i] ; delete [] bends ; bends = NULL ;

} num_of_bendlines = 0 ; points_to_ignore_in_solid_computed = 0 ;

6 .

APPENDIX I part = NULL ; dirty = 1 ;

/*

This function returns a pointer to the BM_AD_bendline structure that contains dimensions data associated with the given bendline.

*/

B M _ A D _ b e n d l i n e

♦BM AUTO DIMENSION: :get_AD_bend_structure (BM_BENDLINE ^♦bendline)

{ int i ; for (i = 0 ; i < num_of_bendlines ; i++) ( if (bends [i] ->bend == bendline) return bends [i] ;

} return NULL ;

/*

This function returns a a (p->p2) vector for a given bendline with respect to a given adj -body.

It also computes a line between the p2 points of the bend dimension points of the adjacent bendline with respect to this flange. */

B M _ V E C T O R c o n s t -

BM_AUTO_DI ENSION: : compute_bend_direction_vector (BM_BENDLINE ^♦bendline, BM 3D BODY ^adj body, BM LINE ^p2 line) { ^{" "} - - - static BM_VECTOR v ; int i ; for (i = 0 ; i < num_of_bendlines ; i++) { if (bends [i] ->data valid ScSc bends [i] ->bend == bendline)

{ if (bends [i] ->adj_bodyl == adj_body) { v = (bends [i] ->bend_bodyl_point_left) .p2 (bends [i] ->bend_bodyl_point_left) .p ; p2_line->set ( (bends [i] ->bend_bodyl_point_left) .p2, (bends [i] ->bend _bodyl_point_right) .p2) ; return v ;

} else if (bends [i] ->adj_body2 == adj_body) { APPENDIX I v = (bends[i] ->bend_body2_point_left) .p2 (bends [i] ->bend_body2_point_left) .p ; p2_line->set ( (bends [i] ->bend_body2_point_left) .p2, (bends [i] ->bend _body2_point_right) .p2) ; return v ;

return BM_null_vector ; }

/*

This function is used to propagate the effects of drawing a flange length for a given bendline, to account for the fact that the flange dimension point is adjacent to the given adj-body. it returns TRUE iff the adj-bend will be marked as drawn as well with respect to the given flange. */ i n t BM_AUTO_DIMENSION: :compute_effects_flange_length_drawn(BM_BENDLINE ♦bendline, BM_3D_B0DY ^♦adj_body,

BM_AD dim_corner ^♦bend_dim_point /♦ bend dim point for bendline *^~J ,

BM_AD dim_corner ♦flange_dim_point /♦ flange dim points for bendline *^"] ,

BM BENDLINE ♦adj bend)

{

BM_AD_dim_corner ^♦♦p_body_points ; int ^♦p_body_drawn ;

BM_AD_dim_corner ^♦♦p_drawn_bend_dim_point ;

BM_AD_dim_corner ^♦♦p_drawn_flange_dim_point ; int i ;

// first we have to find a BM_AD_bendline structure for this adj-bend for (i = 0 ; i < num_of_bendlines ; i++) { if (! bends [i] ->data_valid | | bends [i] ->bend != adj_bend

I I bends [i] ->ignore) continue ; if (bends [i] ->adj_bodyl == adj_body) { if (bends[i] ->bodyl_drawn) return 0 ; p_body_points = Sc(bends [i] ->bodyl_points) ; p_body_drawn = Sc(bends [i] ->bodyl drawn) ; APPENDIX I p _ d r a w n _ b e n d _ d i m _ p o i n t Sc (bends [i] ->drawn_bend_dim_pointl) ; p _ d r a w n _ f l a n g e _ d i m _ p o i n t Sc (bends [i] ->drawn_f lange_dim_pointl) ; break ; else if (bends [i] ->adj_body2 == adj_body) { if (bends [i] ->body2_drawn) return 0 ; p_body_points = Sc(bends [i] ->body2_points) ; p_body_drawn = Sc(bends [i] ->body2_drawn) ; p _ d r a w n _ b e n d _ d i m _ p o i n t Sc (bends [i] ->drawn_bend_dim_point2) ; p _ d r a w n _ f l a n g e _ d i m _ p o i n t Sc (bends [i] ->drawn_f lange_dim_point2) ; break ;

} else return 0 ;

} if (i >= num_of_bendlines) return 0 ;

// now we need to check if the list of flange dimension points for this newly found

// BM_AD_bendline structure contains any flange dimension points that are adjacent to our bendline

BM_AD_dim_corner ^♦temp_point ; for (temp_point = ^♦p_body_points ; temp_point ; temp_point = temp_point->next) { if (temp_point->adj_bendline == bendline) break ; if (NULL == temp_point) return 0 ;

// mark the flange length for the adj -bend drawn with respect to this adj-body

♦p_body_drawn = 1 ;

♦p_drawn_bend_dim_point = flange_dim_point ;

♦p_drawn_flange_dim_point = bend_dim_point ; return 1 ; }

/* this function will return TRUE iff adj_bend has a flange-dimension point with respect to the given flange, that is adjacent to the given bendline.

*/ i n t BM_AUTO_DIMENSION: : check_two_bendlines_adjacent_to_same_f lange (BM BENDLINE ^♦bendline, APPENDIX I BM_3D_BODY ♦flange, BM_BENDLINE ^♦adj send) BM_AD_dim_corner ♦♦p_body_points ; int i ;

// first we have to find a BM_AD_bendline structure for this adj -bend for (i = 0 ,- i < num_of_bendlines ; i++) { if (! bends [i] ->data_valid | | bends [i] ->bend != adj_bend) continue ; if (bends [i] ->adj_bodyl == flange) { p_body_points = Sc(bends [i] ->bodyl_points) ; break ;

} else if (bends [i] ->adj_body2 == flange) { p_body_points = Sc (bends [i] ->body2_points) ; break ;

} else return 0 ,-

} if (i >= n_m_of_bendlines) return 0 ;

// BM_AD_bendline structure contains a flange dimension points that is adjacent to our bendline

BM_AD_dim_corner ^♦temp_point ; for (temp_point = ^♦p_body_points ; temp_point ; temp_point = temp_point->next) { if (temp point->adj bendline == bendline) return 1 ,-

} return 0 ;

}

/*

To set a new part. This updates the view class pointer as well and sets some other parameters. */ void BM_AUTO_DIMENSION: :set_part (BM_PART ♦new_part)

BM_BENDLINE ^♦bendlist ; int i ; delete_contents () ; if (NULL == doc) return ;

10 APPENDIX I part = new_part ; if (NULL == part) return ; half_metal_thickness = (part->get_metal_thickness () ) /2.0 ;

// allocate the bendlist if (0 == (num_of_bendlines = part->get_number_of_bendlines () ) ) return ; if (NULL == (bends = new BM_AD_bendline^ [n_rr_of_bendlines] ) ) goto failure ; bendlist = part->get_bendline_list () ; for (i = 0 ; i < num_of_bendlines ; i++) ( if (bendlist) { bends [i] = new BM_AD_bendline (this,bendlist) ; bendlist = (BM BENDLINE ^♦) bendlist->next () ;

} else bends [i] - NULL ;

} // note, dirty is TRUE return ; failure : delete_contents () ;

}

11 APPENDIX J

UUUUUUUUU//U/////UUUUIUUU/U//UUUUIUUU i ItI Example of bend model viewer implementation including I/I/ // entity visibility function for sheet metal part. //

// //

#include m"sitld nafxu.h u"iiu/u/iium iiiiiiu Him i/i/i ut i ut it

// include all BMAPI files #include "ALLBMAPI .HXX"

// RW_DRAW library include file. #include "RW_DRAW.HXX"

#include "BendCADDoc.h" #include "BendCADViewPart .h"

// RenderWare include files; found in \rwwin\include #include "rwlib.h"

// OpenGL include files. #include "gl\gl.h" #include "gl\glu.h"

// GL_LIST library include file. #include "GL_LIST.HXX"

#include <stdlib.h>

/*

Given a BM_POINT, which represents a point in world coordinates, one can project it onto view window coordinates.

A) If the display is showing the SOLID version of the part, this function passes back a pointer to the edge closest to the camera that lies under that projected point. The x_pos, y_pos, and z_depth pointers carry the view window coordinates (with top left corner as (0,0) and the z-buffer value.

If the projected point is not in the viewing volume, the return value is 0; then of course the edge pointer is meaningless, and is set to NULL. If the projected point IS within the viewing volume, then the return value is 1, and the edge pointer points to the edge that lies under the projected point.

If the pointer is NULL, no edge lies under the projected point. APPENDIX J

B) If the display is showing the WIREFRAME version of the part, only the return value: 1 if within view volume, 0 otherwise - should be examined. If return value is 1, x_pos, y_pos and z_depth values have meaning. The edge pointer is always NULL.

*/ int CBendCADViewPart : :map_point_to_closest_edge (BM_P0INT const- three_d_point, int ^♦x_pos, int ^♦y_pos, double ^♦z_depth, BM EDGE ♦♦closest_edge) { ^" int x, y, visible = 1; double z; visible = map_3d_to_screen (three_d_point, _x, Scy, Scz) ; if ( ! visible ) ( // Point is outside viewing volume . if (closest_edge) ♦closest_edge = NULL; if (x_pos) ^♦x_pos = x; if(y_pos) ^♦y_pos = y; if (z_depth) ^♦z_depth = z; return 0;

}

// Now point is inside view volume. if (x_pos) ^♦x_pos = x; if (y_pos) ^♦y_pos = y; if (z_depth) ♦z_depth = z; if ( ! mi_show_solid) { // Wireframe being drawn; just return that point is if (closest_edge) ^♦closest_edge = NULL; // within view vol. return 1;

}

// Now solid is being shown, if ( ! closest_edge) return 1;

// Point is within view volume. Pick that point. pick_point_rw(x, y, closest_edge) ; return 1;

}

// This function returns the view window coordinates of a point, with top left corner as (0,0) , and the

// z-buffer depth, given the three-dimensional world coordinates of the point. It does so by calling either APPENDIX J

// map_3d_to_screen_solid or map_3d_to_screen_wire, depending on whether wireframe or solid is being shown.

// On return, iff ♦x_pos, ^♦y_pos are both zero, the point was not mapped onto the screen at all. int CBendCADViewPart: :map_3d_to_screen(BM_P0INT three_d_point, int

♦x pos, int ♦y pos, double ^♦z_depth)

{ ^" int visible; if (mi_show_solid) visible =map_3d_to_screen_solid(three_d_point, x_pos, y_pos, z_depth) ; else visible = map_3d_to_screen_wire (three_d_point, x_pos, y_pos, z_depth) ; return visible;

}

// z-buffer depth, given the three-dimensional world coordinates of the point, using OpenGL utilities

// (ie. when wireframe is on) . If return value is 0, point is not mapped onto screen, ie. not visible. int CBendCADViewPart: :map_3d_to_screen_wire (BM_POINT three_d_point, int ^♦x_pos, int ^♦y_pos, double ^♦z_depth)

int viewport [4] , visible = 1; double width, height, depth; double model_mat [16 ] , proj_mat [16] , array [16] ;

// Get the current viewport coords. glGetIntegerv(GL_VIEWPORT, viewport) ;

// Now build a modelview matrix out of the current set of transformations. glMatrixMode (GL_MODELVIEW) ; glPushMatrixO ; glTranslated(md_x_trans, md_y_trans, 0.0); convert_rw_matrix_to_gl_array(md_rot_matrix, array) ; glMultMatrixd(array) ; glTranslated(- m_part_centroid.X() , - m_part_centroid.Y() , - m_part_centroid.Z() ) ; glGetDoublev(GL_MODELVIEW_MATRIX, model_mat) ; glGetDoublev(GL_PROJECTION_MATRIX, proj_mat) ; glPopMatrixO ; gluProject ⁽three_d_point .X () , three_d_point .Y( ) , three_d_point.Z () , model_mat, proj_mat, viewport, -width, -height, -depth) ; if^{( ( (}int⁾width > m_old_rect.right) | | ((int)width < 0) ) visible = 0; APPENDIX J if ( (dnt)height > m_old_rect .bottom) | | ((int)height < 0) ) visible = 0; if(x_pos) ^♦x_pos = (int) width; if(y_P°^s) *y_Pθs = (m_old_rect.bottom - (int) height) ; if(z_depth) ^♦z_depth = depth; return visible; }

// z-buffer depth, given the three-dimensional world coordinates of the point, using RenderWare utilities

// (ie. when solid is on) . If return value is 0, point is not mapped onto the screen, ie. not visible.

// Note: At present, same as map_3d_to_screen_wire. int CBendCADViewPart: : map_3d_to_screen_solid (BM_POINT three_d_point, int ^♦x_pos, int *y_pos, double ^♦z_depth)

int viewport [4] , visible = 1; double width, height, depth; double model_mat [16] , proj_mat [16] , array[16] ;

// Get the current viewport coords. glGetIntegerv(GL_VIEWPORT, viewport) ;

// Now build a modelview matrix out of the current set of transformations. glMatrixMode (GL_MODELVIEW) ; glPushMatrixO ; glTranslated(md_x_trans, md_y_trans, 0.0) ; convert_rw_matrix_to_gl_array(md_rot_matrix, array) ; glMultMatrixd(array) ; glTranslated(- m_part_centroid.X() , - m_part_centroid.Y() , - m_part_centroid.Z () ) ; glGetDoublev(GL_MODELVIEW_MATRIX, model_mat) ; glGetDoublev(GL_PROJECTION_MATRIX, proj_mat) ; glPopMatrixO ; gluProject (three_d_point .X () , three_d_point .Y () , three_d_point .Z () , model_mat, proj_mat, viewport, -width, Scheight, Scdepth) ; if ( ((int)width > tr_old_rect .right) | | ((int)width < 0) ) visible = 0; if ( ((int)height > m_old_rect.bottom) ] | ( nt)height < 0) ) visible = 0; if(x_pos) ^♦x_pos = (int) width;

4 APPENDIX J if(y_pos) ♦y_pos = (m_old_rect .bottom - (int) height) ; if(z_depth) ♦z_depth = depth; return visible;

// This function picks the point indicated by map_3d_to_screen, using RenderWare picking, so this is

// called only when solid is being shown. Note that the pick position is with respect to the top left

// corner of the view window being (0,0) . If the pick results in an edge, a pointer to the edge is

// passed back in picked_edge; else, picked_edge is set to null.

This does not modify the selection

// set in any way. void CBendCADViewPart : :pick_point_rw(int x_pos, int y_pos, BM_EDGE

♦♦picked edge)

{ if (picked_edge == NULL) return;

if( (x_pos > m_old_rect.right) j j (y_pos > m_old_rect .bottom) ) { // Make sure point given is within ^♦picked_edge = NULL;

// viewport, return;

}

RwClump ^♦picked_clump = NULL; RwClump ^♦parent_clump = NULL; long parent_tag, clump_tag, key = 0; BM_EDGE ^♦edge = NULL; if ( ! RwPickScene (m_scene, x_pos, y_pos, m_camera, _m_pick) ) { // Picking command was not successful. ^♦picked_edge = NULL; return;

} if (m_pick.type == rwNAPICKOBJECT) { // No clump was picked.

^♦picked_edge = NULL; return;

} else if (m_pick.type == rwPICKCLUMP) { // Some clump was picked; process it. APPENDIX J picked_clump = m_pick.object.clump.clump; parent_clump = RwGetClumpParent (picked_clump) ; parent_tag = RwGetClumpTag(parent_clump) ; clump_tag = RwGetClumpTag(picked_clump) ; if (parent_tag == 1) { ♦picked_edge = NULL; return; // A trivial bendline or forming was picked.

}

// Now some real edge was picked. key = (parent_tag%10000) ^♦lOOOOO + clump_tag; if (clump_tag) /

Safety check; we expect clump_tag to be non-zero.

♦picked_edge = map_list_name_to_bm_edge (parent_tag, clump tag) ;

} ^"

}

// This function returns the ratio of the view volume dimension

(object space dimensions) to

// the viewport dimension (device space dimensions) . double CBendCADViewPart: :get_object_to_device_ratio (void)

{ double ratio; if (m_old_rect.right) ratio = 2.O^md_current_view_vol [0] / (double) (m_old_rect.right) ; else ratio = 0.0; return ratio;

}

/*

Given a BM_POINT, which represents a point in world coordinates, one can project it onto view window coordinates. This function first checks if the projected point is within the viewing volume. If it IS within viewing volume, the pointer "visible" carries value 1; else, 0. If the point is not visible, no other parameters have any meaning; the return value is 0.

Now suppose the point is visible.

A) If the display is showing the SOLID version of the part, this function checks if ANY of the three_d_bodies in the array "three_db_array" lies "approximately under" the projected point . If the answer is yes, return value is 1; else, return value is zero. The x_pos, y_pos, and z_depth pointers carry the view window coordinates (with top left corner as (0,0)) and the z-buffer value. APPENDIX J

The array "thee_db_array" must be terminated by a NULL pointer to indicate end of array.

B) If the display is showing the WIREFRAME version of the part the return value is always zero. If the point is within view volume, ♦visible is set to 1; else, 0. If ^♦visible is 1, x_pos, y pos and z_depth values have the meaning mentioned above.

*7 int CBendCADViewPart : :is_3d_body_near_point (BM_P0INT const- three_d_point, BM_3D_B0DY ^♦♦three_db_array, int ^♦visible, int ♦x_pos, int ♦y pos, double ^♦z_depth)

{ int x, y, is_visible = 1; double z; int found = 0; is_visible = map_3d_to_screen(three_d_point, _x, Scy, Scz) ; if ( ! is_visible) { // Point is outside viewing volume. if (x_pos) ^♦x_pos = x; if(y_pos) ^♦y_pos = y; if (z_depth) ^♦z_depth = z; if (visible) ^♦visible = 0; return 0;

}

// Now point is inside view volume, if (x_pos) ^♦x_pos = x; if (y_pos) ^♦y_pos = y; if(z_depth) ^♦z_depth = z; if (visible) ^♦visible = 1; if ( ! mi_show_solid) { // Wireframe being drawn; just return that point is return 1;

// within view volume. }

// Now solid is being shown, if ( ! three_db_array) return 0;

// Point is within view volume. Pick pixels in a triangle around that point . found = pick_point_for_3db_rw( (x - 2), (y - 2) , three_db_array) ; if (found) return 1; found = pick_point_for_3db_rw( (x - 2) , (y + 2) , three_db_array) ; if (found) return 1; APPENDIX J found = pick_point_for_3db_rw( (x + 2) , (y + 2) , three_db_array) ; if (found) return 1; found = pick_point_for_3db_rw( (x + 2), (y - 2) , three_db_array) ,- if (found) return 1; return 0;

// This function picks the point indicated, using RenderWare picking, so this is called only when solid is

// being shown. Note that the pick position is with respect to the top left corner of the view window being

// (0,0) . If the pick results in a three_d_body' s clump, and the picked clump corresponds to ANY of the

// three_d_bodies in the array passed, return value is 1; if no match or no three_d_body, return is 0.

// This does not modify the selection set in any way. int CBendCADViewPart : :pick_point_for_3db_rw(int x_pos, int y_pos,

BM 3D BODY ^♦♦three db array)

{ if (three_db_array == NULL) return 0;

if ( (x_pos > m_old_rect .right) | | (y_pos > m_old_rect .bottom) ) { // Make sure point given is within return 0; // viewport.

}

RwClump ^♦picked_clump = NULL; RwClump ^♦parent_clump = NULL; long parent_tag, clump_tag, required_id; int found = 0; if ( ! RwPickScene (m_scene, x_pos, y_pos, m_camera, _m_pick) ) ( // Picking command was not successful . return 0;

} if (m_pick.type == rwNAPICKOBJECT) { // No clump was picked. return 0;

} else if (m_pick.type == rwPICKCLUMP) { // Some clump was picked; process it. picked_clump = m_pick.object.clump.clump; APPENDIX J parent_clump = RwGetClumpParent (picked_clump) ; parent_tag = RwGetClumpTag(parent_clump) ; clump_tag = RwGetClumpTag(picked_clump) ; if (parent_tag == 1) {

// Body of face/bendline/forming was picked, for(int i=0; three_db_array[i] ,- ++i) { // Check if it matches any of the ones in array, if (three_db_array[i] == NULL) break; required_id = three_db_array[i] ->get_idx() ; if (clump_tag%10000 == required_id) ( found = 1; break; }

} return found;

}

// Now some edge's clump was picked. Check if it matches any of the ones in array. fordnt i=0; three_db_array[i] ,- ++i) { if (three_db_array[i] == NULL) break; required_id = three_db_array[i] ->get idx() ; if (parent_tag%10000 == required_id) found = 1; break;

, ' ' return found; }

// This allows the user to insert items into the selection set.

"list" is an array of pointers

// to three-d-bodies, terminated by NULL to indicate end of array.

If list is passed as NULL, every

// three-d-body in the part is inserted into the selection set. void CBendCADViewPart : :insert_selection_set (BM_3D_B0DY ^♦♦list)

int tag;

BV_SELECT_DATA ^♦data = NULL; if ( ! list) { // Insert every three_d_body in part into selection set.

BM_PART ^♦part = NULL; CBendCADDoc^ pDoc = GetDocument () ; ASSERT_VALID(pDoc) ; part = pDoc->get_part 0 ; APPENDIX J if ( ! part) return; BM_FACE ^♦face = NULL;

BM_BENDLINE ^♦bendline = NULL; BM_FORMING ^♦forming = NULL; for(face = part->get_face_list 0 ; face; face (BM FACE ♦) face->next 0 ) { if ( ! face) continue; tag = 100000+ (face->get_idx() ) + 10000 face-_>get_idx() ; if ( ! m_selection_set. find(tag, NULL)) { data - new BV_SELECT_DATA; data->key = tag; data->edge = NULL; data->world_pt = BM_POINT (0.0, 0.0, 0.0) ; m_selection_set . insert (tag, (long) data) ;

}

} for (bendline = part->get_bendline_list 0 ; bendline; bendline = (BM_BENDLINE ^♦)bendline->next () ) { if ( ! bendline) continue; tag = 100000+ (bendline->get_idx() ) + 10000 + bendline->get_idx() ; if ( ! m_selection_set . find(tag, NULL)) { data = new BV_SELECT_DATA; data->key = tag; data->edge = NULL; data->world_pt = BM_POINT (0.0, 0.0, 0.0) ; m_selection_set. insert (tag, (long)data) ;

}

} for (forming = part->get_forming_list () ; forming; forming = (BM_FORMING ^♦) forming->next 0 ) { if { ! forming) continue; tag = 100000+ (forming->get_idx() ) + 10000 + forming->get_idx() ; if ( ! m_selection_set. find(tag, NULL)) { data = new BV_SELECT_DATA; data->key = tag; data->edge = NULL; data->world_pt = BM_POINT(0.0, 0.0, 0.0) ; m_selection_set . insert (tag, (long) data) ;

} ^}

DrawPart (m_hdc) ; return;

10 APPENDIX J

BM_3D_BODY *three_d_body; long i; for ( i=0 , three_d_body = list [ 0] ; three_d_body ; ++i , three_d_body = list [ i] ) { if ( ! three_d_body) break; tag = 100000+ (three_d_body->get_idx() ) + 10000 + three_d_body->get_idx() ; if ( ! m_selection_set. find(tag, NULL)) { data = new BV_SELECT_DATA; data->key = tag; data->edge = NULL; data->world_pt = BM_POINT(0.0, 0.0, 0.0) ; m selection set.insert (tag, (long)data) ;

}

DrawPart (m_hdc) ; return;

}

// This function returns the largest of the x, y and z projections of the part's

// bounding box. double CBendCADViewPart: :get_approx_part_size (void)

{ if (md_part_bbox_size) { double size = max ( md_pa t_bbox_s i z e [ 0 ] , md_part_bbox_size [1] ) ; size = max(size, md_part_bbox_size [2] ) ; return size;

} else return 0.0;

}

// This function passes back the bounding box size of the part in its original orientation, in the

// array "dimensions". This array should be at least of size 3.

Returns 1 if successful, 0 otherwise.

// Caution: This is a quick and hence not very accurate calculation; if you want greater precision,

// write your own function. int CBendCADViewPart ::get part dimensions (double *dimensions)

{ ^{" "}

CBendCADDoc* pDoc = GetDocument () ; if ( ! pDoc) return 0;

BM_PART *part = pDoc->get_part () ; if ( ! part) return 0,- if ( ! dimensions) return 0;

11 APPENDIX J

int i;

BM_POINT centroid;

RwMatrix4d *temp = RwCreateMatrix() ; i = compute_part_centroid(part, 1, -centroid) ; if ( ! i) return 0; compute_part_bbox (part, 1, temp, centroid, dimensions, NULL) ; if ( ! i) return 0; RwDestroyMatrix(temp) ; return 1;

12 APPENDIX K minimum mmimuimuimmmiiiiuiuiummiii ι Uι PART.HXX comments on Bend Model and part structure, etc. '//,

// //

////////////////////////////////////////////////////////////////

/ *

This file contains BMAPI BM_PART class definitions.

Most of the part functions are in PART.CPP.

Constructors and destructors are in PART_NEW.CPP.

Serialization functions are in PART_LOAD_x.CPP and

PART_SAVE.CPP.

Folding functions are in FOLD.CPP and 3D_FUNC.CPP.

Unfolding functions are in UNFOLD.CPP.

***** ******** ********* ********* ********* ********* ********* Discussion :

- One of the design principles of BMAPI is that we want to represent the entire part accurately, with thickness and everything. One of the advantages of this is that we can query anything and get an answer quickly. Even if the answer is not explicitly represented, it is easy to compute. Another advantage of that is that it is straightforward to 3D modelling with the part, since we are representing the part "as a solid part".

- One of the special properties of sheetmetal is that it is symmetric. We can take anvantage of this fact when we represent and reason about the part. For example, we can represent only one side of the sheetmetal, since the other side of the sheetmetal is defined once we know the thickness vector.

- One issue is if we are given a flat version of the part (which usually does not include thickness) , how do we construct the part from that. Do we take that this flat represents the neutral line of the part? In this BMAPI we have taken that the flat version of the part that is explicitly represented in the bend model is the BOTTOM side of the sheetmetal when the part is in TOP VIEW. Another important point is that this flat version is on the X-Y plane. This way even if the thickness changes, changes to our model are minimal.

- Although the flat represents the BOTTOM side of sheet-metal, the dimensions of flat represent neutral-line dimensions. That is, when we bent the part, the neutral-line dimensions of the bent part match the flat dimensions (or are off by an amount defined by the APPENDIX K bend deduction) . This is irrelevant for faces since for them inside=neutral-line=outside. However, it is important for bendlines. That means that the underlying flat of a bendline is mapped onto neutral line of the bend arc.

- When a part is folded, BMAPI normally does not change the 2D->3D transformation of the base-body. However, when the base-body is a bendline and the user has changed some of the bend parameters (for instance, bend angle) the 2D->3D transformation matrix is marked as 'not-up-to-date' . If we used this old transformation matrix, the results would be worng. Therefore we cannot really use the old (out-of-date) transformation matrix. But in that case we don't have anything better to do than to reset the transformation matrix to an identity matrix.

- The size of every 3D body can change in two ways. First, if a face has an adjacent bendline whose bend deduction is non-zero, the face is trimmed by an amount equal to half of the bend deduction. Second, every body can have a trimming value associated with it with respect to an adjacent bendline. When the adjacent bendline is being bent, this body is trimmed by this amount.

- BM API can load all known (ie. latest and all earlier) versions of BMAPI files. However, it saves only in the latest file format .

- When a part is folded, only 2D->3D transformations are computed right away. The rest of the computations is done on demand, ie. whenever needed. When we unfold a part, everything is computed right away, ie. 3D-2D transformations and the flat version of the part.

- When a part is unfolded, BMAPI will use the current 3D-version of 3D-bodies, regardless of whether they are marked as up-to-date or not.

- BM_PART class contains several highlevel functions to facilitate part creation. For example, it contains functions that will inpect the flat version of the part and fix any problems found. Also, it contains functions for creating bendlines for the part when only a set of faces is given. These are high-level functions that are common tasks in many applications.

- For bendlines of the part it is absolutely necessary that they have a centerline specified. Without the centerline, a bendline is almost useless, it cannot be folded or unfolded. Bendline centerline is used for three purposes :

- for bendline folding-unfolding.

- for representing the bendline when the bendline is empty. APPENDIX K

- for determining the bendline orientation.

Notice that the centerline is the most "objective" in the 2D (ie. flat) version of the part, since the flat version has a fixed orientation (ie. normal has to be (0,0,1) ) . In 3D version of the part, centerline is somewhat "subjective" since its meaning is defined by the underlying surface, which can change. For example, we can flip the surface (which necessarily does not change the (sheet-metal) part and as a result have to reverse the centerline as well in order to keep the bendline orientation the same.

- Every bendline has a bend deduction associated with it. This bend deduction controls how part dimensions (at this bendline) change when the part is folded or unfolded. The bend deduction value can be both the positive and negative. When it is positive and we fold the part, then the part dimensions will expand, when it is negative then part dimensions will shrink. (I was told that) it is the industry standard that positive bend deduction means expanding part when folding and shrinking part when unfolding, and accordingly, negative bend deduction means shrinking part when folding and expanding part when unfolding.

- A BM_PART has a parameter bendline_dimensions_status which controls the part dimensions. It is added to facilitate part desing and creation. It is the case often, that the user is given the part dimensions, from outside to outside (for example) . These dimensions are fixed and specified by the customer. The goal of a sheet-metal desing process is to produce a flat of this part so that when this flat is folded its dimensions exactly match the given dimensions. In the design process the user can change the part is a number of ways. However, given this, we always want to keep the part dimensions unchanged.

If bendline_dimensions_status is TRUE, BMAPI will try to keep to part dimensions (at every bendline) constant whenever possible, when the user changes the part/bendline parameters dike, metal thickness, or bendline raidus, etc.) . Otherwise we don't care.

- Another issue that comes up in conjunction with the previous issue is that part dimensions are usually specified as OUTSIDE/INSIDE/NEUTRAL dimensions. For example, the user specifies that given a box, the outside dimensions of the box are 100x100. Technically the user is specifying that on the flat every bendline in this box represents outside dimensions. For example, when the user later changes the metal thickness, outside dimensions have to remain the same, ie. all the added thickness "has to go inside the box". In general, the user (or customer) might specify part dimensions as OUTSIDE/INSIDE/NEUTRAL dimensions, even in the same face. For example, the user might say that across the box, the APPENDIX K dimensions are given as from the inside side of the metal of one side to the outside side of the metal on the other side.

We have a variable in the part to keep track whether the user wants to keep the part dimensions constant or not. If this variable is set so that the user does not require that the part dimensions be kept constant, this while business of INSIDE/OUTSIDE/NUETRAL-LINE is ignored.

*/

#ifndef BM_PART_HXX_INCLUDED #define BM_PART_HXX_INCLUDED

#include <stdio.h>

#include "GLOBAL.HXX" #include "ENTITY.HXX" #include "TOPOLOGY.HXX" class BM PART : public BM ENTITY

{ friend BM_FACE ; friend BM_HOLE ; friend BM_BENDLINE ; friend BM_FORMING ; friend BM_TOPOLOGY ; friend BM_TOPOLOGY_RECORD ; friend BM_BEND_PROPERTY ; friend BM_BEND_PROPERTY_SIMULTANEOUS ;

// every part has a (user-defined) name that it inherits from BM_ENTITY.

// also, the type (BM_ENTITY_TYPE_PART) is stored in BM_ENTITY. protected :

/*

These variables let the user associate parameters, like a name, number, material type, with the part.

*/ char material_type [256 ] ; char name [256] ; char number [32 ] ;

/* APPENDIX K

These variables describe material and metal properties.

*/ double metal_thickness ;

/*

Within this part, two points are considered to be the same if the distance between them is no more than this value .

*/ double distance_tolerance ;

/*

Every part has a list of bend properties. If a bendline has a bend property, this property object has to be on the list of bend properties associated with the part.

This is needed when we save or load a part.

These variables are used only internally. The user of BMAPI has no access to these variables.

*/ long number_of_bend_properties ; BM_BEND_PROPERTY *list_of_bend_properties ;

/'

Part topology.

Note that it is not possible to directly add bodies to the lists here.

This is done only when a body is added to the part topology.

7

BM_TOPOLOGY topology ;

// number of different 3D bodies in the part .

// these numbers should be consistent with the part topology. long number_of_faces ; long number_of_holes ; long number_of_bendlines ; long number_of_formings ;

// lists of faces, holes and bendlines of the part // Note that these are lists of BM_3D_B0DY's. BM_FACE *first_face ; BM_HOLE *first_hole ; BM_BENDLINE +first_bendline ; BM_FORMING *first_forming ;

/*

These parameters represent a bend sequence associated with the part APPENDIX K

*/ long bend_sequence_size ; BM_LINKED_LIST_NODE +bend_sequence ;

/*

This parameter controls part dimensions. If it is TRUE, BMAPI will try to keep to part dimensions (at every bendline) constant whenever possible, when the user changes the part/bendline parameters dike, metal thickness, or bendline raidus, etc.) . Otherwise we don't care.

Default value is 0 - no need to keep dimensions constant .

*/ char bendline_dimensions_status ; public : /*

Constructors and destructors. In PART_NE .CPP,

^■I

BM_PART (void) ;

BM_PART (BM_PART *existing_part) ;

~BM PART(void) ;

// use this to erase the content of the part, the part will be empty.

// it will use topology: :erase_content () to destroy the topology and will then

^■ // destroy all bodies in the part. void erase_content (void) ; void get_material_type (char *out_mat_type) ; void set_material_type(char *in_mat_type) ; void get_name (char *out_name) ; void set name (char +in name) ; void get_part_number(char +out_number) ; void set_part_number(char ^♦in_number) ;

// get number of faces, holes, bendlines inline long get_number_of_faces (void) const { return number_of_faces ; } inline long get_number_of_holes (void) const ( return number_of_holes ; } inline long get number_of_bendlines (void) const { return number_of_bendlines ; ^""} APPENDIX K inline long get_number_of_formings (void) const { return number of formings ; }

// get pointers to lists of faces, holes and bendlines inline BM_FACE +get_face_list (void) const { return first_face inline BM_HOLE *get_hole_list (void) const { return first_hole inline BM_BENDLINE +get_bendline_list (void) const { return first_bendline ; } inline BM_FORMING +get_forming_list (void) const { return first_forming ; }

// To get the number of bend properties and the first property object on the list. inline BM_BEND_PROPERTY +get_list_of_bend_properties (void) const { return list_of_bend_properties ; } inline long get_number_of_bend_properties (void) const { return number_of_bend_properties ; }

// this function tries to find a good default base face. // it chooses a face that has the largest bbox. BM_FACE +get_default_base_face (void) ;

/ *

Functions to get and change the distance tolerance associated with the part.

*/ inline double get_distance_tolerance (void) const { return distance_tolerance ; } void set_distance_tolerance (double new_distance_tolerance) ;

/* ^*

To get the current metal thickness of this part.

*/ inline double get_metal_thickness (void) const { return metal_thickness ; }

/*

Changing thickness will cause a lot of changes in the part. Faces might need some trimming since changing thickness will most likely change bendlines. Also 2D->3D transformations of faces will most likely change. This will happen because the 3D shape of the bendline will change and since the flat represents neutral-line dimensions, the 3D dimensions of the part will most likely change. Bendlines are almost certain to change significantly. W3D of all bendlines has to be recomputed. Also, 2D->3D transformation might have to be recomputed. Note that the part itself does not move in space as the result of changing thickness. However, bendlines 2D->3D transformation will be marked as 'not-up-to-date' . One way APPENDIX K to update them is to pick any face as a base-face and re-fold the part (this will work only if the face picked has its 2D->3D transformation up to date) . The part will stay in the same place (since the base-face will stay in place) and bendlines will have their 2D->3D transformations updated.

Changing thickness is really tricky since most likely all bodies in the part will have 2D->3D transformations and 3D versions invalidated. The problem is that the user normally does not want to re-fold the part. Therefore we have to figure out how to get new valid 2D->3D transformations automatically. Note that this function does not do that. This is something that has to be done by some other function. Note that this function will not try to match faces-bendlines exactly. To accomplish that , call match_all_faces_adjacent_to_bendlines ( ... ) after this function. Even better yet, use set_metal_thickness_update_part ( ... ) which will set new thickness, and then match adjacent faces as well as compute new up-to-date 2D->3D transformations.

*/ void set_metal_thickness (double thickness) ; void set_metal_thickness_update_part (double thickness) ;

/*

This function will trim (if necessary) all faces so that in 3D space they will exactly touch bendlines adjacent to them.

This function is good to call after thickness has changed, or after bendline parameters have been changed.

*/ void match_all_faces_adjacent_to_bendlines (void) ;

/*

Add a 3D body (face, hole, forming or bendline) to the part.

Return FALSE if cannot add.

This function basically calls add_body0 member function of the topology.

*/ int add_3d_body(BM_3D_BODY *body) ;

// get part topology inline BM_TOPOLOGY *get_topology(void) const { return (BM_TOPOLOGY *) -topology ; }

/*

These function allow the user to query and specify whether part dimensions at bendline should be kept constant whenever possible, or not.

*/

// to get the current status

8 APPENDIX K inline int get_bendline_dimensions_status (void) const { return bendline_dimensions_status ; }

// to set the new status.

// for example, if we set it to TRUE, then from this point on, BMAPI

// will try to keep to current part dimensions constant. v o i d s e t _ b e n d l i n e _ d i m e n s i o n s _ s t a t u s ( i n t new_bendline_dimensions_status) ;

/*

Bend sequence related fundtions.

*/ void delete_bend_sequence (void) ; inline long get_bend_sequence_size (void) const { return bend_sequence_size ; }

// this function returns the number of bendlines copied into OUT array int get_bend_sequence (BM_BENDLINE +bendlines [] , long bendlines_array_length) ;

// this function sets a new bend sequence for this bendline int set_bend_sequence (BM_BENDLINE ++bendlines, long bendlines_array_length) ;

// this function checks is a bendline is in the current bend sequence of the part .

// if yes, it returns its index in the bend sequence (ie. the number of this bendline

// in the bend sequence order) . This index is from 0 through sequence_lenght - 1.

// if this bendline is not in the bend sequence, it returns -1. int is_bendline_in_bend_sequence (BM_BENDLINE +bendline) ;

// this function removes a bendline from the bend sequence.

// note that if the bendline is in a simultaneous bend property, it will replace the bendline

// with another bendline from the simultaneous bend property that is not already in the bend sequence. void remove_bendline_from_bend_sequence(BM_BENDLINE *bendline)

// this function deletes the current bend sequence (if one exists) and creates a default // bend sequence. void create_default_bend_sequence (void) ;

/*

This function will eliminate the flat version of all bendlines that do not contain holes.

This function can be useful after a part is unfolded. Normally a folded version of the part contains non-trivial bendlines (ie. inside radius and/or thickness not 0) .

When a part is unfolded, BMAPI transforms these bendlines into 55?

APPENDIX K a flat version exactly as they are.

For example, a 3D-version of a bendline with non-zero thickness or raidus has normally a bend arc drawn. This gets mapped onto flat version of the bendline, which means in the flat version of the part, there is a rectangle between the two faces corresponding to the bendline.

However, sometimes we don't want to see this rectangle.

This function processes only those bendlines that have a simple flat (ie. no holes in the flat) , and that are adjacent to exactly two faces.

This function trims all adjacenct faces so that they touch the bendline centerline and deletes all loops in the bendline flat.

Note that since this function is usually used right after UNFOLD. Once the flat is trimmed, it matches adjacent faces as well - this means that is computes and sets some special trimming values so that in the future when we fold the part we get a correct 3D-version where faces exactly meet bendlines . It does not change the bendline.

*/ void eliminate_simple_bendlines_in_flat_version_A(void) ;

/*

This function will scan all 3D-bodies in the part (except formings), and fix their flat versions. Basically it will call BM_2D_BODY: :fix_flat() function on every 3D_body (except formings) in the part. Note that it will only check the current flat version. It will return FAILURE when either some operation failed (ie. memory allocation) or when it discovered a problem that it cannot fix.

* / int fix flat (void)

/ *

This function checks that face normals are consistent with bendline definitions. For example, if facel has normal (0,0,1) and there is a 90 degree FRONT bend between facel and face2, then face2 normal must be (-1,0,0) . If there is a normal that is a reverse of what it should be, it fixes it. Otherwise when it finds a problem, it returns a FAILURE.

If an input face is given, it only checks the connected component centered at this face. Otherwise it checks the entire part . In that case the order in which it checks faces is the order of face indeces.

This function makes some necessary checks (for example, when a bendline is being bent, it has to have a bend op, a 3D-version

10 APPENDIX K

(although it can be empty) and a surface.

This function checks if 3D-versions of all faces and bendline are up to date. If any is not it will return FAILURE. It will just check if the current version is up to date, it will not compute new 3D-version if the current version is not up to date (because this function is usually used right before unfold - we don't want to compute anything new at this point) .

This function ignores holes and formings.

In 3D_FUNC.CPP. */ int fix_3D_face_normals (BM_FACE *given_base_face = NULL) ;

/*

This function will check 3D-bodies in the part and fix problems it finds. Note that it will only check current 3D-versions. Also, it will not invalidate anything.

Right now, it checks : that face normals are consistent with bendline definitions. For example, if facel has normal (0,0,1) and there is a 90 degree FRONT bend between facel and face2, then face2 normal must be (-1,0,0) .

It will return FAILURE when either some operation failed (ie. memory allocation) or when it discovered a problem that it cannot fix.

In 3D_FUNC.CPP.

*/ int fix_3D(void) ;

********** ********** ********** ********** **********

********** ********** ********** **********

This is a major high-level Bend Model function for part design and construction.

Its purpose is :

- to create bendlines between faces so that the part becomes connected (if possible) .

This function was created to facilitate part creation. Normally a third-party CAD program is used to draw a part. Since Bend Model does not have any control over the CAD system, for our purposes this drawing is just a set of edges. Therefore drawing has to be analyzed then in order to detect the structure of the part .

11 APPENDIX K

After that we can create faces and bendlines of the part .

However, the input drawing is often ambigious - faces and specially bendlines are not uniquely defined. Since this problem is common to many applications this function is part of the Bend Model. Moreover, implementing this function in Bend Model will simplify the face detection software that has to be part of the CAD system.

In this function we assume that edges of faces are sorted according to the left-hand side rule, in a way that is consistent with the plane normal (which defines the plane orientation) . Basically, this function assumes that every plane is correct by itself, when viewed in isolation. However, this function does not require that the orientation of adjacent planes be correct with respect to the bendlines between them.

This function probably does not work well if the part thickness is not zero.

This function can be used to generate bendlines for both 3D-version of the part and the flat version of the part. However, the part can have either the flat or 3D version of the part, but not both. In other words, either all 3D-version of 3D-bodies must be NULL (in which case we have a flat version) , or all flat versions must be NULL.

Implementation is in AUTO_BEND.CPP.

*/ int auto_bend(void) ;

/*

This function was designed to facilitate part creation from data send to Bend Model by a CAD system.

The problem is that the part is often constructed piece by piece, as information becomes available.

However, the problem is that as we construct bodies of the part, we don't know in advance whether the input data represents the FLAT or 3D version of the part.

Usually it is a 3D_version. Therefore by default we assume that it is the 3D version.

Later, once all 3D-bodies of the part have been constructed, we can check it this is true. If not, we have to move all 3D-version "into" flat versions.

This function was designed only for this specific purpose and should not be used for any other purpose. Misuse of this function can crash the system.

This function checks first if the FLAT version of a body is

12 APPENDIX K

NULL. If not, it does not move the 3D-version.

TODO : right now, it only moves faces. */ void BM_PART: :move_3D_to_flat (void) ;

/*

Bend Model can handle part manipulations in both directions :

- Given a 3D version of the part, find the 2D (flat) version of the part (ie. unfold) .

- Given a 2D (flat) version of the part, fold it given a set of bending parameters.

Since BMAPI will have two versions (2D and 3D) of the part most of the time, it is important to maintain their consistency.

BMAPI usually does not, for purposes of efficiency, explicitly compute 2D and 3D versions of the part, but instead computes them on demand - only when the user requests them. For most purposes, it is enough if we just know the transformations that produce of version of a body given another version of the body. Whenever the user manipulates the part, he is actually manipulating the transformations.

There is one difference between 2D->3D and 3D->2D transformations. 2D->3D transformation is usually specified by the user and he can do it in any way he wants (sometimes the result might not be valid for some reason, but we will ignore it here for now) . 3D->2D transformation corresponds to unfolding and there is only one way to do it - there is only one final flat version of the part given a fixed 3D-version.

Both fold and unfold use the bend information of the part to compute all 3D bodies separately according to their individual bending parameters. These transformation generally require that all bodies in the part have 'flat' or ' 3D-version' . That means, when you initially have a 3D version of the part, you have to unfold in before you can use 2D->3D transformations.

Unfolding produces 'flat' . Folding produces ' 3D-version' . Initially the user has to specify either the flat or the 3D-version.

13 APPENDIX K */

/*

This function will force BMAPI to compute the 3D version of the part (from flat) . Specifically, BMAPI will check every face-hole-bendline and whenever the 3D version is not up to date, it will compute an up-to-date 3D version from flat using the 2D-_>3D transformation. If the 2D->3D transformation is not up to date, it will try to compute a valid up to date 2D->3D transformation. Also, whenever a new 3D-version is computed for a bendline, it will first match adjacent bendlines.

*/ void bring_3D_up_to_date (void) ;

/ *

2D->3D transformation functions.

*/

/*

This function will reset all transformations to identity transformations.

It will also set 'bending_status' variable in every bendline to FALSE.

It will also mark all 3D versions of all faces as 'not-up-to-date' but won't recompute them.

Note that this function will reset only the 2D->3D transformation since 3D->2D transformation cannot be reset. */ void reset all transformations (void) ;

/ *

Reset part's 2D->3D transformations.

'body' is the face whose transformation will be an identity transformation. other bodies will have their transformations set according to the bend information with respect to this base body.

This function will update all transformation matrices.

Basically this function allows you to undo all rotations and translations you have done on the entire part.

*/ void reset_part_2Dto3D_transformations (BM_3D_BODY *base_body)

14 APPENDIX K

These function let the user to rotate or translate part. For example, when bending the user might want to move the part of flip it over.

These functions will apply to the entire part equally, ie. every face-hole-bendline changes in exactly the same way.

Note that these are 2D->3D transformations, ie. they change only the 3D version of the part, but they require that the flat version of the part be computed.

Basically the only thing these functions do is to change the 2D-3D transformation associated with every body and mark every body's 3D-version as 'not-up-to-date' .

*/

/*

Rotate the entire part around the line by the angle (in radians) .

Positive direction is defined as clockwise when we look in the direction of the line.

*/ void rotate (BM_LINE const- line, double const angle) ;

/*

Translate the entire part by this vector.

*/ void translate (BM_VECTOR constSc translation_vector)

/*

Apply this transformation to the part. */ void transfor (BM_TRANSFORM const- transformation) ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

FOLD functions. In module FOLD.CPP.

*/

15 APPENDIX K

/ *

Compute 3D version of the part.

This function will compute the 3D version of the part with respect to this base body.

That is, 'base_body' will stay in place while the rest of the part moves .

Note that if you have done any rotations and/or translations with the entire part in the past, they will be in effect after this operation. Normally you want to set some bendline parameters and then call this function to update the part. This function will compute only the transformation matrices of those bodies that change (the minimum change) . 3D versions of every body will be computed on demand, ie. only when the user requests them.This function will use the 'bending_status' member variable of every bendline to determine whether a bendline should be bent.

Note that if the 2D->3D transformation of the base-body is not up to date, FOLD will reset it to an identity transformation.

*/ int fold(BM_3D_BODY *base_body) ;

/ *

This function does the same thing, except that it assumes that the 'bending_status' member variable of every bendline is FALSE (if it is not, it sets it to FALSE) and assumes that the list of bendlines to bend is in the array 'bendline_array' (the length of the array is 'number_of_bendlines' ) . At the end of this function, the 'bending_status' member variable of every bendline indicates whether the bendline was bent or not.

*/ int fold (BM_3D_B0DY *base_body, BM_BENDLINE *bendline_array [] , long number_of_bendlines) ;

/*

This function will fold first n bendline in the bend sequence of the part .

*/ int fold_sequence (BM_3D_BODY +base_body, int n) ;

/*

Given a 3D body that already has its 2D->3D transformation matrix computed, this function will compute a transformation matrix for a given bendline.

This function will assume a number of things (these are not checked) :

- this bendline is adjacent to the body

- this body already has an up-to-date transformation matrix

- this bendline has an up-to-date center_line_2D

16 APPENDIX K

Note that these two functions require that bendlines have 2D centerline. This also means that bendlines should have a flat. For performance reasons these functions do not check if bendlines have a flat, but instead assume that they do and reference it directly. This function is used by FOLD.

*/ int bend_bendline_wrt_3D_body(BM_3D_BODY +body /* finished body */,

BM_BENDLINE *current_bendline /* bendline whose transformation we are computing */,

BM TRANSFORM *tf /* start with this transformation ^♦/) ;

/*

The same, except the bendline is already finished and we are computing the 3D body.

This function is used by FOLD.

*/ intbend_3D_body_wrt_bendline (BM_BENDLINE *current_bendline /♦ finished bendline ^♦/,

BM_3D_BODY +body /* body whose transformation we are computing ^♦/,

BM_TRANSFORM +tf /^♦ start with this transformation */) ;

/*

UNFOLD functions. In module UNFOLD.CPP.

*/

/*

This function will unfold the entire part.

It chooses a default base body for every connected component of the part. When the part is unfolded, this function will move it to the X-Y plane, such that the base-body overlaps the origin (0,0,0) .

This function will first compute 3D->2D transformation for every 3D-body, then compute Wflat for every 3D-body, then trim Wflat and finally compute flat.

This function chooses the face with the smallest id that has a non-NULL normal as the base-body.

If that is not possible, it chooses the first bendline that has an underlying surface for which a non-NULL normal can be computed as the base-body.

17 APPENDIX K

This function will use BM_PART: :unfold(BM_3D_BODY +base_body) to unfold all connected components of the part.

This function returns 1 iff everything was fine; 0 iff unfolding of some part of the part failed.

First this function will check that all face normals are correct with respect to current bendlines.

Note that this function will use the current 3D-version of every 3D-body. In order to unfold, all 3D-versions have to be up to date. This function will not check if they are up to date, it will actually mark all 3D-bodies as up to date. Also, it assumes that all bendlines have a valid 3D-centerline. Again, it will not check that, instead it will simply validate the current centerline.

*/ int unfold(void) ;

/*

This function will unfold only a connected component of the part, centered around the base-body.

This function is used by the BM_PART: :unfold(void) function that unfolds the entire part.

This function will compute new 3D->2D transformations for every 3D-body in the connected component, and will then use BM_3D_B0DY: :compute_flat_from_3D() function to compute a new flat version for every 3D-body.

If the base-body is a face and the input flag is set to TRUE, then thefirst thing this function does is to check that all face normals are correct with respect to current bendlines - but only in the connected component centered at this base-body(face) .

Note that this function will use the current 3D-version of every 3D-body.

Make sure that all 3D-versions are up to date before calling this function.

This function will actually mark all 3D-bodies it processes as up to date.

*/ int unfold(BM_3D_B0DY +base_body, int check_face_normals = 0)

;

/*

Given a 3D body that already has its 3D->2D transformation matrix computed, this function will compute a transformation matrix for a given bendline.

This function will assume a number of things (these are not checked) :

18 APPENDIX K

- this bendline is adjacent to the body

- this body already has an up-to-date transformation matrix

- this bendline has an up-to-date center_line_3D

Note that these two functions require that bendlines have a centerline. This also means that bendlines a non-NULL 3D-version. For performance reasons these functions do not check that, b t instead assume that they do and reference it directly.

This function is used by UNFOLD.

Note that this function is really an inverse of bend_bendline_wrt_3D_body(...) .

*/ int unfold_bendline_wrt_3D_body(BM_3D_BODY *body /^♦ finished body ^♦/,

BM_BENDLINE +current_bendline /* bendline whose transformation we are computing ^♦/,

BM_TRANSFORM +tf /^♦ start with this transformation +/) ;

/*

This function is used by UNFOLD.

Note that this function is really an inverse of bend_3D_body_wrt_bendline (...) .

*/ int unfold_3D_body wrt_bendline (BM_BENDLINE +current_bendline /♦ finished bendline */^",

BM_3D_B0DY *body /* body whose transformation we are computing ^♦/,

BM_TRANSFORM +tf /^♦ start with this transformation ♦/) ;

/*

********** ********** ********** ********** **********

********** ********** ********** **********

This function returns a "paper model of this part". A paper model is a part that has thickness 0, bendline radius and bend deductions 0 as well. If required, this function will return a mapping from user-id's associated with bendlines into former values of 'inside radius' and 'bend deduction' .

*/

BM_PART ⁺make_paper_model (double +*inside_radius /♦ OUT */, double ++bend deduction /♦ OUT ♦/) ;

19 APPENDIX K

/*

This function will save the current part in the given file in PGF format.

PGF format is used by CMU's BendCAD program.

Returns : TRUE iff everything is fine and the part is saved.

*/ int save_PGF_file(FILE *file) ; int save PGF file(char +filename) ;

/*

Load functions are in PART_LOAD_x.CPP. */ virtual int load(FILE *file) ; virtual int load(char ^♦filename /^♦ must not be NULL */) int BM PART Load(FILE *file, long version) ;

/*

Save functions are in PART_SAVE.CPP. */ virtual int save(FILE +file) ; virtual int save(char *filename /^♦ must not be NULL */)

} ;

#endif // BM PART HXX INCLUDED

20 APPENDIX L

/////////////////////////////////////////////////////////////////

// //

// Example of 3-D manipulation and navigation with dynamic // // calculation of the rotation axis. //

// //

void CBendCADViewPart : :OnLButtonUp (UINT nFlags, CPoint point)

: :ReleaseCapture () ;

: :ClipCursor (NULL) ; // Release mouse, free cursor. m_lt_btn_up_point = point; double x_extent, y_extent; int add_hit_to_list = 0; int viewport [4] ;

BM_VECT0R offset;

BM_PART +part = GetDocument () ->get_part () ;

RwV3d vert; switch(ml_last_function_chosen) { case(ZOOM_WINDOW_MODE) : if (mi_dynamic_rot_center_on) {

// User has further zoomed in. compute_non_dyn_equivalent () ; _} RwIdentityMatrix(md-dyn-rot-matrix) ; compute_zoom_region_translation ( ) ; x_extent = fabs ( (double) (m_lt_btn_down_point .x - m_lt_btn_up_point .x) ) *

2. O*md_current_view_vol [0] / (double) (m_old_rect .right) ; y_extent = fabs ( (double) (m_lt_btn_down_point .y - m_lt_btn_up_point .y) ) *

2. O*md_current_view_vol [1] / (double) (m_old_rect .bottom) ; set_new_view_volume (x_extent, y_extent) ; compute_part_bbox(part, mi_show_3d, md_rot_matrix, m_part_centroid, md_part_bbox_size, -offset) ; mi_bbox_is_up_to_date = 0; // Even though bbox computed, we mark as not up to date so that zoom-all

DrawPart (m_hdc) ; // will work correctly.

// Now turn on dynamic rotation center if need be. if ( (md_part_bbox_size [0] > APPENDIX L

2.2*md_current_view_vol [0] ) ] | // User has zoomed in so that not all

( m d _ p a r t _ b b o x _ s i z e [ l ] >

2.2*md_current_view_vol [1] ) ) { // of the part is in view volume . if ( ! md_dyn_rot_matrix) md_dyn_rot_matrix = RwCreateMatrix() ; if (mi show_solid) {

7/ Pick the center of the screen to find point on clump to rotate around. i f ( R w P i c k S c e n e ( m _ s c e n e , m_old_rect . right /2 , m_old_rect .bottom/2 , m_camera, _m_pick) ) { if (m_pick. type == rwNAPICKOBJECT) { // No clump was picked. m_dynamic_rot_center = BM_POINT(0.0, 0.0, 0.0) ;

} else if (m_pick. type == rwPICKCLUMP) { // Some clump was picked; process it. v e r t = m_pick . obj ect . clump . wcpoint ,- m_dynamic_rot_center = BM_POINT (vert .x, vert.y, vert.z)

}

} else { m_dynamic_rot_center = BM_P0INT (0.0, 0.0,

0.0) ;

} mi_dynamic_rot_center_on = 1;

} else { compute_non_dyn_equivalent () ; i f ( m d _ d y n _ r o t _ m a t r i x ) RwDestroyMatrix(md_dyn_rot_matrix) ,- md_dyn_rot_matrix = NULL; mi dynamic rot center on = 0;

} ^{" " "} break; case(SELECT_OBJECTS_MODE) : if (nFlags Sc MK_CONTROL) add_hit_to_list = 1; if ( ! mi_show_solid) { // Wireframe being displayed, which means OpenGL picking. glGetIntegerv(GL_VIEWPORT, viewport) ; glMatrixMode (GL_PROJECTION) ; glPushMatrixO ; glLoadldentity0 ; APPENDIX L gluPickMatrix ( ( doubl e ) point . x , (double) (m_old_rect. bottom - point. y) , 3.0, 3.0, viewport) ; glOrtho (- md_current_view_vol [0] , md_current_view_vol [0] ,

- md_current_view_vol [1] , md_current_view_vol [1] ,

DrawPart (m_hdc) ; //

Draw the part in selection mode. glMatrixMode (GL_PROJECTION) ; glPopMatrix() ; glMatrixMode (GL_MODELVIEW) ; process_hits_record(add_hit_to_list) ; ml_last_function_chosen = NO_FUNCTION_MODE; DrawPart (m_hdc) ; reset_selection_buffer() ; else { // Solid being shown, so RenderWare picking. if (RwPickScene (m_scene, point.x, point.y, m_camera, _m_pick) ) process_hits_record(add_hit_to_list) ; ml_last_function_chosen = NO_FUNCTION_MODE; DrawPart (m hdc) ;

} break; case(ZOOM_IN_OUT_MODE) : compute_part_bbox(part, mi_show_3d, md_rot_matrix, m_part_centroid, md_part_bbox_size, -offset) ; mi_bbox_is_up_to_date = 0; if (mi dynamic_rot_center_on) {

7/ User has further zoomed in. compute_non_dyn_equivalent () ; RwIdentityMatrix(md_dyn_rot_matrix) ;

// Now turn on dynamic rotation center if need be. i f ( ( m d p a r t b b o x _ s i z e [ 0 ] >

2.2*md_current_view_vol [0] ) | //^"User has zoomed in so that not all

( m d _ p a r t b b o x _ s i z e [ l ] >

2.2⁺md_current_view_vol [1] ) ) { ll oi the part is in view volume . mi_bbox_is_up_to_date = 0; // Even though bbox computed, we mark as not up to date so that zoom-all if(! md_dyn_rot_matrix) md_dyn_rot_matrix = RwCreateMatrixO ; // _will work APPENDIX L correctly. if (mi_show_solid) {

// Pick the center of the screen to find point on clump to rotate around. i f ( R w P i c k S c e n e ( m_ s c e n e , m_old_rect . right /2 , m_old_rect .bottom/2, m_camera, _m_pick) ) { if (m_pick.type == rwNAPICKOBJECT) { // No clump was picked. m_dynamic_rot_center = BM_POINT(0.0, 0.0, 0.0) ;

} else if (m_pick.type == rwPICKCLUMP) { // Some clump was picked; process it. v e r t m_pick.object .clump.wcpoint; m_dynamic_rot_center = BM_P0INT(vert .x, vert.y, vert.z)

» > else { m_dynamic_rot_center = BM_POINT (0.0, 0.0,

0.0)

} mi dynamic rot center on = 1;

} else { if (mi_dynamic_rot_center_on) { compute_non_dyn_equivalent () ; i f ( m d _ d y n _ r o t _ m a t r i x ) RwDestroyMatrix(md_dyn_rot_matrix) ; md_dyn_rot_matrix = NULL; mi dynamic rot center on = 0;

break; default: break;

}^; ml_last_function_chosen = NO_FUNCTION_MODE; CBendCADView: :OnLButtonUp (nFlags, point) ;

Claims

5 «WHAT IS CLAIMED:

1. A system for developing a bend model of a part to be produced in an

intelligent production facility, said part including a plurality of faces and at least one

bendline, said system comprising:

a receiving system for receiving initial part information relating to said part, said

initial part information including data relating to a representation of said part within a

first predetermined coordinate space;

a face detecting system for detecting, within said first predetermined coordinate

space, said faces of said part based on said initial part information;

a bendline identification system for identifying at least one bendline of said part

based on said faces detected by said face detecting system; and

a system for generating additional part information, relating to said part and

including data relating to a representation of said part within a second predetermined

coordinate space, by performing a predetermined operation on each of said faces detected by said face detecting system, said operation being performed based upon at least said

initial part information and at least one bendline identified by said bendline identification

system.

2. A system according to claim 1, wherein said first predetermined

coordinate space comprises a 2-D coordinate space and said second predetermined

coordinate space comprises a 3-D coordinate space, said predetermined operation

comprising a folding operation performed on said faces detected by said face detecting system.

3. A system according to claim 2, said folding operation including rotating _- <"v ι_ / 7473

5* 1

and translating each of said faces detected by said face detecting system relative to at

least one bendline identified by said bendline identification system.

4. A system according to claim 2, wherein said initial part information

further comprises a bend angle amount relating to at least one bendline of said part, said

folding operation being performed based on said bend angle amount.

5. A system according to claim 4, wherein said folding operation comprises

rotating and translating each of said faces around at least one bendline of said part based

on said bend angle amount.

6. A system according to claim 1, wherein said first predetermined

coordinate space comprises a 3-D coordinate space and said second predetermined

coordinate space comprises a 2-D coordinate space, said predetermined operation

comprising an unfolding operation performed on said faces detected by said face

detecting system.

7. A system according to claim 6, said unfolding operation including rotating

least one bendline identified by said bendline identification system.

8. A system according to claim 6, wherein said initial part information

unfolding operation being performed based on said bend angle amount.

9. A system according to claim 8, wherein said unfolding operation

comprises rotating and translating each of said faces around at least one bendline of said

part based on said bend angle amount.

10. A system according to claim 1, wherein said data relating to a _■» «* -_,

representation of said part within a first predetermined coordinate space comprises

coordinate data.

11. A system according to claim 1, wherein said data relating to a

representation of said part within a first predetermined coordinate space comprises vector

data.

12. A system according to claim 1 , wherein said part comprises a sheet metal

part.

13. A system according to claim 1 , wherein said faces comprise base surfaces

of said part and folded surfaces of said part.

14. A system according to claim 1 , further comprising a system for detecting additional features of said part, said predetermined operation being performed on said

additional features of said part.

15. A system according to claim 14, said additional features comprising holes,

openings and special formings of said part.

16. A system according to claim 1, wherein said initial part information

comprises data relating to a representation of said part in a 3-D coordinate space, said

data including thickness data of said part in said 3-D coordinate space.

17. A system according to claim 16, further comprising a system for

eliminating said thickness data to provide modified part data relating to a representation

of said part in said 3-D coordinate space with no thickness.

18. A system according to claim 17, wherein said face detecting system detects said faces of said part based on said modified part data and said representation of

said part in said 3-D coordinate space with no thickness.

19. A system according to claim 18, wherein said second predetermined

coordinate space comprises a 2-D coordinate space, said predetermined operation

comprising an unfolding operation performed on said faces detected by said face

detecting system.

20. A system according to claim 19, said unfolding operation including

rotating and translating each of said faces detected by said face detecting system relative

to at least one bendline identified by said bendline identification system.

21. A system according to claim 1 , further comprising an auto-trimming and

cleanup system for performing an auto-trimming and cleanup operation on said data

relating to said initial part information to prepare said data for said face detecting system

and said bendline identification system.

22. A system according to claim 21 , wherein said data comprises part entity

data representing, at least, line entities and bendline entities of said part, said auto-

trimming and cleanup system comprising a system for detecting intersection points of

said entities and for selectively breaking said entities at detected intersection points, and

a system for assigning the resultant broken entities to have a common endpoint based on

said detected intersection points.

23. A system according to claim 21 , wherein said data comprises part entity

data representing, at least, line entities of said part, said auto-trimming and cleanup

system comprising a system for detecting open intersection areas between adjacent

entities and for selectively connecting said adjacent entities by assigning a common

endpoint to said adjacent entities.

24. A system according to claim 23, wherein an open intersection area is detected, by said system for detecting open intersection areas, if the endpoints of said

adjacent entities are within a predetermined distance from one another.

25. A system according to claim 1 , wherein said data relating to said initial

part information comprises part entity data representing, at least, line entities of said part,

said face detecting system performing a loop and entity analysis of said part based on

said part entity data to detect said faces of said part.

26. A system according to claim 25, wherein said loop and entity analysis is

initially performed on an outside boundary of said part and then performed on inner

boundaries and areas of said part.

27. A system according to claim 26, wherein said face detecting system

generates an initial linked list of entities when performing said loop and entity analysis

on the outside boundary of said part, said initial linked list of entities defining an outward

loop and boundary of said part.

28. A system according to claim 27, wherein said face detecting system

further generates an additional linked list of entities when performing said loop and entity

analysis on the inside boundaries and areas of said part, said additional linked list of

entities defining inner loops and boundaries of said part.

29. A system according to claim 28, wherein said face detecting system

further comprises a system for generating a loop tree based on said outward loop defined

by said initial linked list of entities and said inner loops defined by said additional linked

list of entities.

30. A system according to claim 29, wherein said face detecting system

detects said faces of said part based on said loop tree and the sequence of boundaries 7473

defined by said initial linked list entities and said additional linked list of entities. »

31. A system according to claim 28, wherein said bendline identification

system comprises a system for analyzing said initial linked list of entities and said

additional linked list of entities

to determine common line entities between said faces detected by said face detecting

system.

32. A system according to claim 31 , said at least one bendline being identified

based on the detection of one of said faces having only one common line entity with another one of said faces.

33. A system according to claim 32, wherein said at least one bendline is

identified and defined at said one common line entity detected between said faces.

34. A system according to claim 32, wherein said bendline identification

system applies predetermined heuristics to identify bendlines of said part when said

system for determining common line entities detects that there are more than one

common line entity between said faces.

35. A system according to claim 34, wherein said heuristics comprise

identifying bendlines of said part such that a minimum number of total bendlines are

identified for said part.

36. A system according to claim 34, wherein said heuristics comprise

identifying bendlines of said part, when one of said faces have more than one common

line entity with another one of said faces, based on the common line entity that has the

longest length.

37. A system according to claim 1, wherein said bendline identification system comprises a system for detecting common edges between said faces detected by

said face detecting system, said at least one bendline being identified based on the

detection of one of said faces having only one common edge with another one of said

faces.

38. A system according to claim 37, wherein said at least one bendline is

identified and defined at said one common edge detected

between said faces.

39. A system according to claim 37, wherein said bendline identification

system for detecting common edges detects that there are more than one common edge

between said faces.

40. A system according to claim 39, wherein said heuristics comprise

identified for said part.

41. A system according to claim 39, wherein said heuristics comprise

edge with another one of said faces, based on the common edge that has the longest length.

42. A system according to claim 1 , further comprising a system for receiving

a deduction amount related to said part and a system for compensating for bend

deduction, when performing said predetermined operation on said faces, based on said deduction amount.

43. A system according to claim 42, wherein said predetermined operation comprises a folding operation performed on said faces detected by said face detecting

system, said system for compensating for bend deduction increasing a dimensional length

of said faces by one half of said deduction amount on each side of said bendline of said

part when performing said folding operation.

44. A system according to claim 42, wherein said predetermined operation

comprises an unfolding operation performed on said faces detected by said face detecting

system, said system for compensating for bend deduction decreasing a dimensional

length of said faces by one half of said deduction amount on each side of said bendline of said part when performing said unfolding operation.

45. A method for developing a bend model of a part to be produced in an

bendline, said method comprising the steps of:

receiving initial part information relating to said part, said initial part information

including data relating to a representation of said part within a first predetermined

coordinate space; detecting, within said first predetermined coordinate space, said faces of said part

based on said initial part information;

identifying at least one bendline of said part based on said faces detected by said

detecting; and generating additional part information, including data relating to a representation

of said part within a second predetermined coordinate space, by performing a

predetermined operation on each of said faces detected by said detecting, said operation

being performed based upon said initial part information and at least one bendline 5* l

identified by said identifying.

46. A method according to claim 45, wherein said first predetermined

coordinate space comprises a 2-D coordinate space and said second predetermined

coordinate space comprises a 3-D coordinate space, said method further comprising

performing a folding operation on said faces detected by said detecting.

47. A method according to claim 46, said folding operation including rotating

and translating each of said faces relative to at least one bendline identified by said

identifying.

48. A method according to claim 46, wherein said initial part information

folding operation being performed based on said bend angle amount.

49. A method according to claim 48, wherein said folding operation comprises

on said bend angle amount.

50. A method according to claim 45, wherein said first predetermined

coordinate space comprises a 3-D coordinate space and said second predetermined

coordinate space comprises a 2-D coordinate space, said method further comprising

performing an unfolding operation on said faces detected by said detecting.

51. A method according to claim 50, said unfolding operation including

rotating and translating each of said faces relative to at least one bendline identified by

said identifying.

52. A method according to claim 50, wherein said initial part information

further comprises a bend angle amount relating to at least one bendline of said part, said unfolding operation being performed based on said bend angle amount.

53. A method according to claim 52, wherein said unfolding operation

part based on said bend angle amount.

54. A method according to claim 45, further comprising detecting additional

features of said part, said predetermined operation being performed on said additional

features of said part.

55. A method according to claim 54, said additional features comprising

holes, openings and special formings of said part.

56. A method according to claim 45, wherein said initial part information

data including thickness data of said part in said 3-D coordinate space.

57. A method according to claim 56, further comprising eliminating said

thickness data to provide modified part data relating to a representation of said part in

said 3-D coordinate space with no thickness.

58. A method according to claim 57, wherein said detecting said faces of said

part is adapted to be performed based on said modified part data and said representation

of said part in said 3-D coordinate space with no thickness.

59. A method according to claim 58, wherein said second predetermined

performing an unfolding operation on said faces detected by said detecting.

60. A method according to claim 45, further comprising performing an auto-

trimming and cleanup operation on said data of said initial part information before 5*3

detecting said faces and identifying said at least one bendline.

61. A method according to claim 45, wherein said data of said initial part

information comprises part entity data representing, at least, line entities of said part, said

detecting of said faces comprising performing a loop and entity analysis of said part

based on said part entity data to detect said faces of said part.

62. A method according to claim 61 , wherein said loop and entity analysis is

boundaries and areas of said part.

63. A method according to claim 45, wherein said identifying said at least one

bendline comprises detecting common edges between said faces detected by said

detecting, said at least one bendline

being identified based on the detection of one of said faces having only one common

edge with another one of said faces.

64. A method according to claim 63, wherein said identifying of said at least one bendline further comprises applying predetermined heuristics to identify bendlines

of said part when more than one common edge between said faces is detected.

65. A method according to claim 64, wherein said heuristics comprise

identified for said part.

66. A method according to claim 64, wherein said heuristics comprise

identifying bendlines of said part, when one of said faces have more than one common edge with another one of said faces, based on the common edge that has the longest length.

67. A method according to claim 45, further comprising receiving a deduction

amount related to said part and compensating for bend deduction, when performing said

predetermined operation on said faces, based on said deduction amount.

68. A method according to claim 67, further comprising performing a folding

operation on said faces, said compensating for bend deduction comprising increasing a

dimensional length of said faces by one half of said deduction amount on each side of

said bendline of said part when said folding operation is performed.

69. A method according to claim 67, further comprising performing an unfolding operation on said faces, said compensating for bend deduction comprising

decreasing a dimensional length of said faces by one half of said deduction amount on

each side of said bendline of said part when said unfolding operation is performed.

70. A system for developing a bend model of a part to be produced in a

production facility, said part including a plurality of faces and at least one bendline, said

system comprising:

means for receiving initial part information relating to said part, said initial part

information including data relating to a representation of said part within a first

predetermined coordinate space;

means for detecting, within said first predetermined coordinate space, said faces

of said part based on said initial part information;

means for identifying at least one bendline of said part based on said faces

detected by said face detecting means; and means for generating additional part information, including data relating to a representation of said part within a second predetermined coordinate space, by

performing a predetermined operation on said faces detected by said detecting means,

said predetermined operation being performed based on, at least in part, said bendline

identified by said bendline determining means.

71. A system according to claim 70, wherein said first predetermined

coordinate space comprises a 2-D coordinate space and said second predetermined

coordinate space comprises a 3-D coordinate space, said predetermined operation

comprising a folding operation performed on said faces detected by said face detecting

means.

72. A system according to claim 71 , said folding operation including rotating

and translating each of said faces detected by said face detecting means relative to at least

one bendline identified by said bendline identification means.

73. A system according to claim 71, wherein said initial part information

folding operation being performed based on said bend angle amount.

74. A system according to claim 70, wherein said first predetermined

coordinate space comprises a 3-D coordinate space and said second predetermined

coordinate space comprises a 2-D coordinate space, said predetermined operation

comprising an unfolding operation performed on said faces detected by said face

detecting means.

75. A system according to claim 74, said unfolding operation including

rotating and translating each of said faces detected by said face detecting means relative to at least one bendline identified by said bendline identification means.

76. A system according to claim 74, wherein said initial part information

unfolding operation being performed based on said bend angle amount.

77. A system according to claim 70, further comprising means for detecting

additional features of said part, said predetermined operation being performed on said

additional features of said part.

78. A system according to claim 77, said additional features comprising holes,

openings and special formings of said part.

79. A system according to claim 70, wherein said initial part information comprises data relating to a representation of said part in a 3-D coordinate space, said

data including thickness data of said part in said 3-D coordinate space.

80. A system according to claim 79, further comprising means for eliminating

said thickness data to provide modified part data relating to a representation of said part

in said 3-D coordinate space with no thickness.

81. A system according to claim 80, wherein said face detecting means detects

said faces of said part based on said modified part data and said representation of said

part in said 3-D coordinate space with no thickness.

82. A system according to claim 81, wherein said second predetermined

coordinate space comprises a 2-D coordinate space, said predetermined operation

comprising an unfolding operation performed on said faces detected by said face

detecting means.

83. A system according to claim 70, further comprising auto-trimming and

cleanup means for performing an auto-trimming and cleanup operation on said data 5

relating to said initial part information to prepare said data for said face detecting means

and said bendline identification means.

84. A system according to claim 83, wherein said data comprises part entity

trimming and cleanup means comprising means for detecting intersection points of said

entities and for selectively breaking said entities at detected intersection points, and

means for assigning the resultant broken entities to have a common endpoint based on

said detected intersection points.

85. A system according to claim 83, wherein said data comprises part entity

means comprising means for detecting open intersection areas between adjacent entities and for selectively connecting said adjacent entities by assigning a common endpoint to

said adjacent entities.

86. A system according to claim 85, wherein an open intersection area is

detected, by said means for detecting open intersection areas, if the endpoints of said

adjacent entities are within a predetermined distance from one another.

87. A system according to claim 70, wherein said data relating to said initial

said face detecting means performing a loop and entity analysis of said part based on said

part entity data to detect said faces of said part.

88. A system according to claim 87, wherein said loop and entity analysis is

initially performed on an outside boundary of said part and then performed on inner boundaries and areas of said part. 5 P

89. A system according to claim 88, wherein said face detecting means

loop and boundary of said part.

90. A system accordingto claim 89, wherein said face detecting means further

generates an additional linked list of entities when performing said loop and entity

entities defining inner loops and boundaries of said part.

91. A system according to claim 90, wherein said face detecting means further

comprises means for generating a loop tree based on said outward loop defined by said

initial linked list of entities and said inner loops defined by said additional linked list of

entities.

92. A system according to claim 91 , wherein said face detecting means detects

said faces of said part based on said loop tree and the sequence of boundaries defined by

said initial linked list entities and said additional linked list of entities.

93. A system according to claim 70, wherein said bendline identification

means comprises means for detecting common edges between said faces detected by said

face detecting means, said at least one bendline being identified based on the detection

of one of said faces having only one common edge with another one of said faces.

94. A system according to claim 93, wherein said bendline identification

means applies predetermined heuristics to identify bendlines of said part when said

means for detecting common edges detects that there are more than one common edge

between said faces.

95. A system according to claim 94, wherein said heuristics comprise

identified for said part.

96. A system according to claim 94, wherein said heuristics comprise

edge with another one of said faces, based on the common edge that has the longest

length.

97. A system according to claim 70, further comprising means for receiving

a deduction amount related to said part and means for compensating for bend deduction,

when performing said predetermined operation on said faces, based on said deduction

amount.

98. A system according to claim 97, wherein said predetermined operation

comprises a folding operation performed on said faces detected by said face detecting

means, said means for compensating for bend deduction increasing a dimensional length

of

said faces by one half of said deduction amount on each side of said bendline of said part

when performing said folding operation.

99. A system according to claim 97, wherein said predetermined operation

means, said means for compensating for bend deduction decreasing a dimensional length

part when performing said unfolding operation.

100. A system for developing a bend model of a part to be produced in an 5 S^O

intelligent production facility, said system comprising:

initial part information comprising respective representations of a plurality of views said

part in 2-D coordinate space, each of said representations including part thickness

representations of said part, said initial part information further comprising extraneous information;

a cleanup operation system that performs a 2-D cleanup operation on said initial

part information to eliminate said extraneous information and to identify each of said

representations;

a part thickness elimination system for selectively eliminating said part thickness

representations in each of said identified representations to provide modified

representations of said views of said part in 2-D coordinate space with no thickness; and

a system for developing a representation of said part in 3-D coordinate space

based on said modified representations of said part in 2-D coordinate space with no

thickness.

101. A system according to claim 100, wherein said initial part information

comprises part entity data representing, at least, line entities of said part, said cleanup

operation system comprising a break and trimming system for detecting intersection

points of said entities and for selectively breaking said entities at detected intersection

points, said break and trimming system assigning the resultant broken entities to have a

common endpoint based on said detected intersection points.

102. A system according to claim 101, said break and trimming system further

comprising a system for detecting open intersection areas between adjacent entities and for selectively connecting said adjacent entities by assigning a common endpoint to said

adjacent entities.

103. A system according to claim 102, wherein an open intersection area is

detected, by said system for detecting open intersection areas, if the endpoints of said

adjacent entities are within a predetermined distance from one another.

104. A system according to claim 100, wherein said cleanup operation system

comprises a system for developing a connectivity graph structure based on said initial

part information, said cleanup operation system eliminating said extraneous information

based on said connectivity graph structure.

105. A system according to claim 104, wherein said extraneous information

comprises unconnected line entities, said unconnected line entities relating to, at least,

dimension lines.

106. A system according to claim 100, wherein said initial part information

comprises type field dala for identifying part entity data and extraneous information

relating to text, said cleanup operation system eliminating said extraneous information

relating to text based on said type field data.

107. A system according to claim 100, wherein said plurality of views comprise

a top view, a front view, and a right side view of said part in 2-D coordinate space.

108. A system according to claim 107, wherein said cleanup operation system

comprises a system for detecting, based on said initial part information, said

representations of said top view, said front view, and said right side view of said part.

109. A system according to claim 100, wherein said part comprises a sheet

metal part.

1 10. A system according to claim 100, wherein said part thickness elimination

system comprises a system for prompting a user to identify said part thickness

representations to be eliminated in each of said plurality of views and to identify a

dimension of said part to be retained in each of said plurality of views.

1 1 1. A system according to claim 1 10, wherein said dimension of said part

comprises one of an outside dimension or an inside dimension of said part.

112. A system according to claim 100, wherein said developing system

comprises a system for performing a projection operation to develop said representation

of said part in 3-D coordinate space based on said modified representations of said part

in 2-D coordinate space.

113. A system according to claim 112, wherein said projection operation

comprises detecting the relative depths of each of said plurality of views and projecting

each of said plurality of views into 3-D coordinate space.

114. A method for developing a bend model of a part to be produced in an

intelligent production facility, said system comprising the steps of:

comprising respective representations of a plurality of views said part in 2-D coordinate

space, each of said representations including part thickness representations of said part;

performing a 2-D cleanup operation on said initial part information to eliminate

extraneous information and to identify each of said representations;

selectively eliminating said part thickness representations in each of said

identified representations to provide modified representations of said views of said part

in 2-D coordinate space with no thickness; and developing a representation of said part in 3-D coordinate space based on said

modified representations of said part in 2-D coordinate space with no thickness.

115. A method according to claim 114, wherein said initial part information

comprises part entity data representing, at least, line entities of said part, said performing

comprising detecting intersection points of said entities and selectively breaking said

entities at detected intersection points, said resultant broken entities being assigned to

have a common endpoint based on said detected intersection points.

116. A method according to claim 114, said performing comprising detecting

open intersection areas between adjacent entities and selectively connecting said adjacent

entities by assigning a common endpoint to said adjacent entities.

1 17. A method according to claim 1 16, wherein an open intersection area is

detected, by said step of detecting open intersection areas, if the endpoints of said

adjacent entities are within a predetermined distance from one another.

118. A method according to claim 1 14, said performing comprising developing

a connectivity graph structure based on said initial part information and eliminating said

extraneous information based on said connectivity graph structure.

119. A method according to claim 1 18, wherein said extraneous information

dimension lines.

120. A method according to claim 114, wherein said initial part information

comprises type field data for identifying part entity data and extraneous information

relating to text, said performing eliminating said extraneous information relating to text

based on said type field data.

121. A method according to claim 1 14, wherein said plurality of views comprise

122. A method according to claim 121, wherein said performing comprises

detecting, based on said initial part information, said representations of said top view,

said front view, and said right side view of said part.

123. A method according to claim 1 14, wherein said part comprises a sheet metal

part.

124. A method according to claim 1 14, wherein said selectively eliminating said

part thickness comprises prompting a user to identify said part thickness representations

to be eliminated in each of said plurality of views and to identify a dimension of said part

to be retained in each of said plurality of views.

125. A method according to claim 124, wherein said dimension of said part

comprises one of an outside dimension or an inside dimension of said part.

126. A method according to claim 1 14, wherein said developing comprises

performing a projection operation to develop said representation of said part in 3-D

coordinate space based on said modified representations of said part in 2-D coordinate

space.

127. A method according to claim 126, wherein said projection operation

each of said plurality of views into 3-D coordinate space.

128. A system for developing a bend model of a part to be produced in an

intelligent production facility, said system comprising: a receiving system for receiving initial part information relating to said part, said initial part information comprising respective representations of a plurality of views said

representations of said part, said initial part information further comprising extraneous

information; a cleanup operation system that performs a 2-D cleanup operation on said initial

representations; and

a system for developing a representation of said part in 3-D coordinate space

based on said identified representations of said part in 2-D coordinate space.

129. A system according to claim 128, wherein said initial part information

common endpoint based on said detected intersection points.

130. A system according to claim 129, said break and trimming system further

comprising a system for detecting open intersection areas between adjacent entities and

for selectively connecting said adjacent entities by assigning a common endpoint to said

adjacent entities.

131. A system according to claim 130, wherein an open intersection area is

adjacent entities are within a predetermined distance from one another.

132. A system according to claim 128, wherein said plurality of views comprise a top view, a front view, and a right side view of said part in 2-D coordinate space.

133. A system according to claim 132, wherein said cleanup operation system

comprises a system for detecting, based on said initial part information, said

134. A system according to claim 128, wherein said part comprises a sheet metal

part.

135. A system according to claim 128, further comprising

representations in each of said identified representations to provide modified

representations of said views of said part in 2-D coordinate space with no thickness, said developing system being adapted to develop said representation of said part in 3-D

coordinate space based on each of said identified representations of said part.

136. A system according to claim 135, wherein said part thickness elimination

system comprises a system for prompting a user to identify said part thickness

dimension of said part to be retained in each of said plurality of views.

137. A system according to claim 128, wherein said developing system

of said part in 3-D coordinate space based on said identified representations of said part

in 2-D coordinate space.

138. A system according to claim 137, wherein said projection operation comprises detecting the relative depths of each of said plurality of views and projecting

each of said plurality of views into 3-D coordinate space.

139. A system according to claim 128, further comprising a system for

performing a 3-D clean-up operation on said representafionof said part in 3-D coordinate

space to further clean-up and eliminate extraneous information in said representation of

said part.

140. A system according to claim 139, wherein said 3-D clean-up operation

includes a process for identifying and eliminating one sided open lines that are present

in said representation of said part in 3-D coordinate space.

141. A system according to claim 139, wherein said 3-D clean-up operation

includes a process for identifying and cleaning each bendline that is present in said

representation of said part in 3-D coordinate space.

142. A system according to claim 139, wherein said 3-D clean-up operation

includes a process for trimming each face that is present in said representation of said part

in 3-D coordinate space.

143. A system for developing a bend model of a part to be produced in an intelligent production facility, said system comprising:

representations of said part;

representations in each of said representations to provide modified representations of said

views of said part in 2-D coordinate space with no thickness; and

a system for developing a representation of said part in 3-D coordinate space based on said modified representations of said part in 2-D coordinate space with no

thickness.

144. A system according to claim 143, wherein said initial part information

further comprises extraneous information, said system further comprising a cleanup

operation system that performs a 2-D cleanup operation on said initial part information

to eliminate said extraneous information and to identify each of said representations.

145. A system according to claim 144, wherein said initial part information

further comprises part entity data representing, at least, line entities of said part, said

cleanup operation system comprising a break and trimming system for detecting

intersection points of said entities and for selectively breaking said entities at detected

intersection points, said break and trimming system assigning the resultant broken

entities to have a common endpoint based on said detected intersection points.

146. A system according to claim 143, wherein said developing system

in 2-D coordinate space.

147. A system according to claim 146, wherein said projection operation

each of said plurality of views into 3-D coordinate space.

148. A system according to claim 143, wherein said part thickness elimination

system comprises a system for prompting a user to identify said part thickness

dimension of said part to be retained in each of said plurality of views.

149. A system according to claim 143, further comprising a system for

performing a 3-D clean-up operation on said representation of said part in 3-D coordinate

said part.

150. A system according to claim 149, wherein said 3-D clean-up operation

in said representation of said part in 3-D coordinate space.

151. A system according to claim 149, wherein said 3-D clean-up operation

includes a process for identifying and cleaning each bendline that is present in said representation of said part in 3-D coordinate space.

152. A system according to claim 149, wherein said 3-D clean-up operation

in 3-D coordinate space.