WO1997046951A1

WO1997046951A1 - Device for word processing

Info

Publication number: WO1997046951A1
Application number: PCT/SE1997/000984
Authority: WO
Inventors: Håkan LENNERSTAD
Original assignee: Lennerstad Idé
Priority date: 1996-06-04
Filing date: 1997-06-04
Publication date: 1997-12-11

Abstract

A device for word processing is designed, when inputting characters, to automatically and in real time suggest a probable continuation of the formed combination of characters to a user. The word-processing device comprises an input unit (I), a first memory unit (M1) for storing of input characters, a display unit (D), a second memory unit (M2) and a comparator unit (C). When inputting a character in the first memory unit (M1), the comparator unit (C) compares the combination of characters created in the first memory unit (M1) with a word bank which is stored in the second memory unit (M2). The comparator unit (C) selects a word from the word bank, which is suggested to the user via the display unit (D). The comparator unit (C) is activatable to create or update, in the second memory unit (M2), the word bank based on the words in the first memory unit (M1), or in some other memory unit, whereupon the words in the second memory unit (M2) are sorted according to their priority value. The priority value of a word in the second memory unit (M2) is calculated on the basis of the number of presences of this word in the first memory unit (M1), or in some other memory unit.

Description

DEVICE FOR WORD PROCESSING

Field of the Invention

The present invention relates to a device and a com¬ puter program for word processing. Background Art Normal word-processing apparatus have an input unit with a keyboard, a data-processing unit, a memory unit and a display unit with a screen. When the user presses a key on the keyboard, an electric signal is generated which is forwarded to the data processing unit, in which the signal is processed. Then the signal is stored in the memory unit and presented on the screen of the display unit in the form of a character. For each character which is to be shown on the screen, at least one keystroke thus is required. Such electronic word-processing apparatus often have many other functions which are intended to facilitate the user's writing work and increase the speed of writing.

Among such functions, mention can be made of a cut- and-paste function which makes it possible for the user to select a text portion shown on the screen and activate the device to copy the corresponding characters in the memory unit to a second memory unit, whereupon the user on a later occasion can select to again copy this text portion in an arbitrary position in the text presented on the screen.

There is also the possibility of using an electro¬ nic dictionary, which is a bank of words stored in a second memory unit. The user may select to perform a spellcheck in which a word written on the screen is com- pared with the words included in the dictionary. The dictionary may also be adapted to the user by his adding desired words to the word bank.

Some word-processing apparatus also permit automa¬ tic correction of the written text, which implies that the user, in a second memory unit, may store a word or a text portion to which a combination of letters is as¬ signed. When the user writes the combination of letters, the associated text portion is automatically copied into the written text. This function may be used to reduce both the number of spelling mistakes and the number of required keystrokes. It requires, however, that the user himself create a bank of words or text portions with the assigned combinations of letters.

Even if the word-processing apparatus are provided with the above-mentioned functions, it takes a compara¬ tively long time to write a text, also for experienced typists. Any spelling mistakes must be checked afterwards by comparing it with the electronic dictionary, which the user must continuously update with new words. Unexperienced typists may find it very stressful to write long texts, .and the written text may also contain many wrongly spelt words, especially if the text consists of a large number of long words.

EP-A-0 643 357 discloses a method for input of text, in which a word which the typed combination of characters can form in continued typing is suggested to the user in real time, i.e. when a character has been input. The sug¬ gested word is selected from an assigned dictionary. The user may either accept the suggested word, for instance by pressing a special key on the keyboard, or continue to type a character, whereupon a new word is suggested etc. Thus, the number of keystrokes that is required to type the commonest words is reduced.

A drawback of this known method resides in the man- ner in which the search and the selection take place in the assigned dictionary before a word is suggested to the user. The search, a so-called binary search, in fact pro¬ ceeds such that the input combination of characters is first compared with the central word in the dictionary, which is sorted in alphabetical order. Depending on the outcome, half of the words can be excluded. Then a new comparison is effected, and the procedure is repeated until only one word remains. This is a simple method of searching which, however, is relatively slow, restricting the possibilities of using the text input process in a word-processing apparatus, above all when using large dictionaries, to suggest, in real time, to the user con¬ tinuations of the input combination of characters. If the search method is so slow that the user must wait for the suggestions, the advantage of the text input process dis¬ appears completely. An effective search therefore is of vital importance.

A further restriction of the prior-art text input process is that the input combination of characters must, for a continuation to be suggested to the user, have a unique equivalence in the dictionary, i.e. the diction- ary must not contain a number of alternative continua¬ tions of the input combination of characters. This re¬ sults in general in the necessity of the user inputting several characters of a word before a continuation can be suggested at all. A similar technique is disclosed in US-A-4,330, 845 and US-A-4, 954,956. However, these references do not de¬ scribe methods for searching and selecting proposed words in the assigned dictionary. Summary of the Invention An object of the present invention is to obviate, wholly or partly, the above-described drawbacks, in par¬ ticular to provide a device and a computer program which reduce the number of keystrokes required to write a given text, and which reduces the number of spelling mistakes in the written text, by automatically suggesting words that can be formed of the input combination of charac¬ ters.

These and other objects, which will appear from the following description, have now been achieved by a device and a computer program as defined in appended claims 1 and 7, respectively. The subclaims define preferred embo¬ diments. By data item is meant a character. However, it should be emphasised that the invention is not restrict¬ ed to such characters as form words, but may also com¬ prise other characters, such as symbols, mathematical signs etc. By group of data items is meant a group of characters which may form a word or part of a word, i.e. a combination of characters. It is also conceivable that the group of characters forms a combination of symbols, mathematical signs etc. For reasons of clarity, the fol- lowing description will, however, be restricted to char¬ acters and words.

The fact that each word in the second memory unit has been given a priority value permits a quick and ef¬ fective search in the second memory unit for words that can be formed in the continued input of characters.

According to the invention, the priority value of each word in the second memory unit, i.e. in the word bank, is calculated on the basis of the number of pre¬ sences of this word in a reference unit. The reference unit may be the first memory unit, i.e. the text that is being prepared, or another memory unit having a similar text content. Since the priority values are based on the presence of the words in the reference unit, the inven¬ tive device will suggest to the user the word that the input combination of characters, based on the user's actually employed vocabulary, has the greatest probabi¬ lity or at least a great probability of forming in the continued input of characters. If a non-desired sugges¬ tion is obtained, the user may obtain new suggestions by further typing the word. For each new character that is typed, a new suggestion is obtained. With this device, the number of keystrokes that is required to type the commonest words is reduced to a considerable extent.

It should be emphasised that one more advantage of the invention resides in the fact that the user need not care about creating a special dictionary in the second memory unit, but the word bank in the second memory unit can be created by command or automatically based on the text which is presently being prepared or based on some other data file having a similar content of words. Tech¬ nical terms that are frequent in the text and names of people and places will thus also be automatically sug¬ gested to the user. It would be time-consuming and, besides, difficult to manually compile a dictionary which has an efficient priority of words for the purpose of minimising the typing work. According to a preferred embodiment, the priority value of each word in the second memory unit is calcu¬ lated on the basis of the presence of the word as an ini¬ tial partial amount of the words in the reference unit. This has been found to increase the probability of the most frequent word which the input combination of charac¬ ters can form in continued typing being suggested to the user.

According to a further preferred embodiment, the words in the second memory unit are organised in a tree structure. This results in extremely short search times for finding the next suggestion for the continuation of the input combination of characters. Furthermore, this structure permits a fast update of the second memory unit, and therefore the words that the user himself uses most frequently can be continuously suggested to the user. Brief Description of the Drawings

The invention and its advantages will be described in more detail below with reference to the drawings, which for the purpose of exemplification illustrate a presently preferred embodiment.

Fig. 1 illustrates very schematically the basic structure of the device in the form of interconnected units. Fig. 2 shows an example of part of a preferred tree structure in the second memory unit of the device. Figs 3 and 4 show flow diagrams for illustrating the function of an inventive computer program. Description of Preferred Embodiments

Fig. 1 illustrates the basic structure of an embodi- ment of a device according to the invention. The device has an input unit ( I ) which comprises a keyboard and which communicates with a data processing unit (DP). The data processing unit (DP) is connected to a first memory unit (Ml), a comparator unit (C) and a display unit (D) which comprises a screen. The device also has a second memory unit (M2), which via the comparator unit (C) com¬ municates with the first memory unit (Ml). The units are monitored and controlled by means of a central controller (not shown) . When the user presses a key on the keyboard of the input unit (I), an electric signal is generated, which is transmitted to and processed in the data processing unit (DP) which then forwards the processed signal on the one hand to the display unit (D) which presents a correspond- ing character on the screen and, on the other hand, to the first memory unit (Ml) in which the character is stored. Several successive characters which are terminat¬ ed with a space or some other punctuation mark, form a word, which is stored in the memory unit (Ml). The second memory unit (M2) contains a set of diffe¬ rent words, to each of which a priority value is assign¬ ed. As will be described below, this set is formed on the basis of the text which is stored in a reference unit, which can be the first memory unit (Ml) or some other data file which preferably has a content of words corre¬ sponding to the text that is to be typed.

The priority value of a certain word in the second memory unit (M2) is calculated, according to a preferred embodiment, as the number of times this word appears as an initial partial amount of the words in the reference unit. With a view to calculating, for instance, the pri¬ ority value of the word "common", the number of presences of "common", "commonplace", "commonweal", "commonwealth" etc. is summed up. Each word appears only once in the word bank.

The comparator unit (C) is designed to compare, when a character is input via the keyboard, the combination of characters formed in the first memory unit (Ml) with the words stored in the second memory unit (M2). The compara¬ tor unit (C) then selects a suggestion word among the words in the second memory unit (M2) . The suggestion word is the word that has the greatest priority value and that begins with the combination of characters formed in the first memory unit (Ml). This suggestion word is presented on the screen, suitably as a continuation of the recent¬ ly typed combination of characters, for instance in a deviating colour or in italics, and the user may select to accept the suggestion word by inputting a special character, such as a tabulator character ("TAB"), thereby activating the comparator unit (C) to replace the combi¬ nation of characters in the first memory unit (Ml) with the suggestion word. If the user selects to ignore the suggestion word and instead inputs a new character, the procedure is repeated, whereupon the comparator unit (C) selects a new suggestion word from the words in the second memory unit (M2) and presents this to the user. Below follows an example of this procedure, keystroke by keystroke.

Keystroke Accepted word on screen Suggestion word c c can o co common

"TAB" common commonhold w commonw commonweal

"TAB" commonweal commonwealth

"space" commonweal

In this example, five keystrokes are effected in- stead of normally 11 keystrokes, which obviously saves the user typing work and increases the probability of the word being correctly spelt.

In a particularly preferred embodiment, the words in the second memory unit (M2 ) are organised in a tree structure, which is illustrated for the above example in Fig. 2. The tree structure has a root which branches off in main nodes, the number of which corresponds to the number of conceivable characters, and therefore an Eng¬ lish device generally has at least 26 branches (the let- ters "a"-"z" and the hyphen "-"). In other applications, the alphabet may be different. The letters of the lan¬ guage and the hyphen should always be included. For instance, when typing a mathematical text in the program LATEX^®, the alphabet should be increased by the six char- acters \{}$_^".

Each of the main nodes branches off in an equally large number of subnodes, which also branch off in the same number of subnodes etc. However, there are no branches if words are missing, and therefore the number of nodes is equal to the number of different words in the word bank.

To each node an initial sequence is assigned, which for each main node is a character and which for each sub- node is the initial sequence of the superordinated node increased by a character. All nodes in a branch support different initial sequences.

The words are sorted in the tree such that the words having the greatest priority values are connected to the nodes closest to the root. The words are also arranged such that their initial character conforms with the ini¬ tial sequence of the respective nodes.

When the user inputs a character, the comparator unit (C) reads the word that is connected to the node which is assigned to this initial sequence. This word is suggested to the user. If the user inputs one more char¬ acter and thus forms a combination of characters in the first memory unit (Ml), the comparator unit (C) reads the word that is connected to the subnode, whose initial sequence is identical with said combination of charac¬ ters, and then this word is suggested to the user. The comparator unit (C) thus proceeds a step downwards in the tree, along the branch indicated by the input character, and collects the word that is positioned in this place in the tree, if available. This sorting of the word bank in the second memory unit (M2 ) results in extremely short search times for the comparator unit (C). When a suggested word is accepted by the user, the comparator unit (C) suitably reads the word that is con¬ nected to the subnode, whose initial sequence is iden¬ tical with the accepted word, whereupon this read word is suggested to the user. In the above example, this pro- cedure is used and further reduces the number of neces¬ sary keystrokes since the user can accept parts of the words that are to be typed.

It will be appreciated that the inventive device can be used in exactly the same manner as usual electro- nic word-processing apparatus, apart from the fact the user must press a special key to accept a word suggested by the device.

On each occasion, the user may select to update the word bank on the basis of the text stored in the first memory unit (Ml). In such update, new priority values are calculated for the words included in this text, which are placed in their intended place in the tree structure.

The word bank can also be updated automatically and in real time, i.e. while the text is being typed. This update suitably is effected in such manner that a word which is accepted by the user and which is already avail¬ able in the tree is advanced a step closer to the root. If the word was not previously available in the tree, it is placed in its natural position in the tree, i.e. closest outside the word which has the longest beginning in common with the accepted word. Before the user begins to input his text into the device, he conveniently creates a word bank in the second memory unit (M2) on the basis of some other data file having a similar content of words. In this fashion, words will be suggested to the user even from the beginning. Alternatively, the use may select to begin typing his text without first creating such a word bank, and there¬ fore the second memory unit (M2) is in this case empty when starting. The word bank is suitably updated in real time, i.e. words are stored in the second memory unit (M2) as the user inputs his text.

It should also be possible for the user to shut off the update of the tree for words close to the root. As a result, the user may learn to rely on one-letter and two- letter abbreviations of common words and learn them by heart. This may be of great importance to the typing speed. Suitably, the user may have the possibility of selecting how many levels from the root that are to be locked. The device should also provide the user with a pos¬ sibility of setting a limit to the number of characters that must distinguish a suggestion word from an input combination of characters. If this limit is set at, for instance, 3 characters, words that only differ by two characters from the input combination of characters will thus not be suggested to the user. This increases the search speed in the second memory unit (M2) since the words thus, to be stored at all in a node in this unit, must have a length exceeding the initial sequence of the node by more than two characters.

It should be pointed out that the described method of calculating the priority value of a word can be modi¬ fied within the scope of the invention. For example, the priority value need not reflect the presence of the word as a partial amount of other words in the reference unit, but can instead reflect the presence of merely the word itself in the reference unit. It is also possible to let a word move closer to the root quicker than one step when using the word. Besides, if the user completes a word that is not available in the search tree, this word can be placed closer to the root than farthest away. Thus, the words appearing in the last typed passages, or on the last typed page, will be suggested more often to the user. In texts with a less homogeneous use of words, if different vocabularies appear in different parts of the text, this priority may be preferred in order to minimise the typing work. It is also conceivable that the priority value is in some manner weighted with the length of the word. The important thing is that the device is so com¬ posed that the word that minimises the total amount of work in typing a text is suggested to the user. From the practical point of view, the most important thing is to minimise the number of keystrokes when typing the text. The most frequent word should therefore be suggested to the user, except for such words as form too short a con¬ tinuation of the already typed part of the word. The inventive device is generally suited for users who type so slowly that they have time to read the words suggested on the screen. These users can thus work at a slower speed without their typing speed decreasing. The number of spelling mistakes in the text will also de- crease since the user often need not type each word in its entirety.

The device is also to a high extent suited for users who type fast and use much time with word processing. For these users, the word proposals will sometimes succeed each other so quickly on the screen that they cannot be read. By observing the suggestions, these users can learn the commonest suggestions by heart and can then accept the suggested word even before they have had time to per¬ ceive the text on the screen. Long words which are typi- cal of the document can then, without any preparatory work at all, be typed regularly with two or three key¬ strokes. Moreover, very common words such as "and", "that", "but" can be regularly typed as "a", "t", "b" . These changes result in a great saving of work in word processing and also reduce the number of typing errors. It will be appreciated that the device according to the invention can be accomplished by means of a software- controlled computer, the comparator unit (C) being accom¬ plished as a program module, which performs the above- described functions of the comparator unit (C), is stored in a memory of the computer and works against a conven- tional word-processing program. Figs 3 and 4 show a block diagram of such a comparator module included in an inven¬ tive computer program.

The skilled person realises that the above-mentioned memory units (Ml, M2) can either be two physically sepa- rated units or two parts of an overall memory, such as a ROM or RAM memory.

Claims

1. A device for word processing, comprising an input unit (I), a first memory unit (Ml), which is adapted to store groups of data items which are supplied via the input unit (I), a display unit (D), which is adapted to present the group of data items created in the first memory unit (Ml), a second memory unit (M2), which is adapted to store a bank of groups of data items, and a comparator unit (C), which when inputting a data item in the first memory unit (Ml), compares the group of data items created in the first memory unit (Ml) with the groups of data items stored in the second memory unit ( 2 ) , the comparator unit (C) being activatable to replace the group. of data items in question in the first memory unit (Ml) with a group of data items selected by the comparator unit (C) in the second memory unit ( 2) , the display unit (D) being adapted to automatically pre- sent, when inputting a data item in the first memory unit (Ml), the group of data items which is selected by the comparator unit (C) in the second memory unit (M2) , c h a r a c t e r i s e d in that the comparator unit (C) is activatable to create or update, in the second memory unit (M2), said bank on the basis of the groups of data items in the first memory unit (Ml), or in some other memory unit, each group of data items in the second memory unit (M2) after said update being given a priority value, that the priority value of a group of data items in the second memory unit (M2) is calculated on the basis of the number of presences of said group of data items in the first memory unit (Ml), or in some other memory unit, and that the comparator unit (C) on the basis of said priority values selects the group of data items in the second memory unit (M2), whose initial data item conforms with the entire group of data items in question in the first memory unit (Ml).

2. A device as claimed in claim 1, wherein the pri¬ ority value of a group of data items in the second memory unit (M2) is calculated on the basis of the number of presences of this group of data items as an initial par¬ tial amount of the groups of data items in the first memory unit (Ml), or in some other memory unit.

3. A device as claimed in claim 1 or 2, wherein the groups of data items in the second memory unit (M2) are organised in a tree structure.

4. A device as claimed in claim 3, wherein said tree structure has a root, which branches off in a number of main nodes, each branching off in a number of subnodes, each in turn branching off in a number of subnodes etc, the number of possible branches being equal to the number of different individual data items which may be included in said groups of data items, each node being given an initial sequence, which for each main node is a data item and which for each subnode is the initial sequence of the superordinated node in¬ creased by a data item, all nodes within a branch being given different initial sequences, and the comparator unit (C) being designed to sort, in said update, the groups of data items according to said priority values and connect the groups of data items to the nodes, each group of data items being connected to an associated node whose initial sequence conforms with the initial data item of the group of data items.

5. A device as claimed in claim 4, wherein the com¬ parator unit (C) is designed to select the group of data items which is connected to the node in the tree whose initial sequence conforms with the entire group of data items in question in the first memory unit (Ml).

6. A device as claimed in any one of claims 1-5, wherein the group of data items which is selected by the comparator unit (C), after being preliminarily suggested to a user via the display unit (D), can be accepted by the user and be caused to replace the group of data items in question in the first memory unit (Ml) by the user activating the comparator unit (C) via the input unit (I).

7. A computer program for storing in a word-process¬ ing apparatus comprising an input unit (I), a first memo¬ ry unit (Ml), which is adapted to store groups of data items supplied via the input unit (I), and a second me o- ry unit (M2), which is adapted to store a bank of groups of data items, the computer program comprising a comparator module (C), which is designed to compare, when inputting a data item in the first memory unit (Ml), the group of data items created in the first memory unit (Ml) with the groups of data items stored in the second memory unit (M2), and which is activatable to replace the group of data items in question in the first memory unit (Ml) with a group of data items selected by the comparator module ( C ) in the second memory unit ( M2 ) , c h a r a c t e r i s e d in that the comparator module (C) is activatable to create or update in said second memory unit (M2) said bank on the basis of groups of data items in a reference unit, the comparator module (C) assigning to each group of data items in the second memory unit (M2) an asso¬ ciated priority value, which is calculated on the basis of the number of presences of this group of data items in said reference unit, and that the comparator module (C) on the basis of said priority values selects that group of data items in the second memory unit (M2) whose initial data item conforms with the entire group of data items in question in the first memory unit (Ml).

8. A computer program as claimed in claim 7, wherein the priority value of a group of data items in a second memory unit (M2) is calculated on the basis of the number of presences of this group of data items as an initial partial amount of the groups of data items in the refe¬ rence unit.

9. A computer program as claimed in claim 7 or 8, wherein the comparator module (C) is designed to organise the groups of data items in the second memory unit (M2) in a tree structure.

10. A computer program as claimed^, in claim 9, where¬ in said tree structure has a root, which branches off in a number of main nodes, each branching off in a number of subnodes, each in turn branching off in a number of sub¬ nodes etc, the number of possible branches being equal to the number of different individual data items which may be included in said groups of data items, each node being given an initial sequence, which for each main node is a data item and which for each subnode is the initial sequence of the superordinated node increased by a data item, all nodes within a branch being given different initial sequences, the comparator module (C) being designed to sort, during said update, the groups of data items according to said priority values and connect the groups of data items to the nodes, each group of data items being connected to an associated node, whose initial sequence conforms with the initial data item of the group of data items.

11. A computer program as claimed in claim 10, wherein the comparator module (C) is designed to select the group of data items which is connected to that node in the tree whose initial sequence conforms with the entire group of data items in question in the first memo¬ ry unit (Ml ) .