WO1991015818A1

WO1991015818A1 - System for searching and retrieving data from data bases via audio access with automatic faxing of results

Info

Publication number: WO1991015818A1
Application number: PCT/US1991/002377
Authority: WO
Inventors: Richard Kollin; Robert C. Miller; James E. Shea
Original assignee: Richard Kollin; Miller Robert C; Shea James E
Priority date: 1990-04-09
Filing date: 1991-04-08
Publication date: 1991-10-17
Also published as: AU7699391A

Abstract

This invention permits any user to access one or more database (12) by using a telephone (1). Upon successful connection, the caller hears a set of instructions from a voice board (4) which the caller responds to by hitting required numbers on the keypad of his touchtone telephone (1) and initiating a search of one or more databases (12), by spelling out a term, phrase, or proper name using the keypad and a coding technique. The system transmits the search request through a microcomputer (5) and modem (11) in the language required by the database (12), establishes telecommunication, downloads data, and either disconnects with the database (12) or enters an intermediate stage of communication. The system then converts the digital data received from the database (12) into electronic speech by a voice board (4) and reads the data to the caller, who can elect to have further data either read or re-typeset and FAXed (9, 10).

Description

SYSTEM FOR SEARCHING AND RETRIEVING DATA FROM DffEA BASES VIA AUDIO ACCESS WITH AUTOMATIC FAXING OF RESULTS

BACKGROUND OF THE INVENTION

This invention is an advance in the field of information science, which is the collection, classification, storage, re¬ trieval and dissemination of recorded knowledge. Information retrieval has advanced markedly as a technology over the last decade but there still exists an oppressive number of restric¬ tions to the fast and efficient access to and the retrieval of information from over 5000 databases now commercially available to the public . This invention makes such access fast, simple, and Inexpensive. It is also an advantage of this invention that a caller can obtain Instantaneously, via FAX, full-length, re- typeset documents without the need for a computer or printer.

Many thousands of databases covering virtually all aspects of knowledge are available to the public. To access them however, a user must open an account with a database host (an entity that makes one or more databases avalable for searching) , obtain a password, learn a communications language and possess a personal computer, modem, and dedicated telephone link in order to connect with the host. Aspects of telephonic retrieval transmission of data are represented In U.S. Pat. Nps. 4,450,477; 4,337,485; 4,303,941; and 3,751,670. None of these cited patents, except the first, addresses the features Inherent In the access of multiple databases. The most advanced of the current systems, in terms of ease-of-use for the inexperienced searcher, which access multiple databases is shown in U.S. Pat. No. 4,774,655. This system still requires that the user have available a personal computer, modem, printer, and dedicated telephone line (or cen¬ tralized equivalent) and personal prearranged ID or account number, however.

Several important advances have taken place in the field of information retrieval over the last 3-6 years. Until approximate¬ ly 1984, publicly available, commercial databases, produced by hundreds of small private publishers as well as the U.S. govern¬ ment, were gathered together and mounted on the large computers of less than a dozen "vendors" , the largest of which are Dialog Information Services of California and BRS Technologies of Wash¬ ington, D.C. The databases offered by these vendors were searcha¬ ble mainly by search specialists, however, because each vendor required a knowledge of a unique search language to obtain data, as well as an account with each. These barriers effectively excluded the public (students, teachers, and professionals) from access to the data stored with the vendors. In 1984 Telebase Systems, Inc. of Pennsylvania created a system (EasyNet) previ¬ ously cited as U.S. Pat. No. 4,774,655 for the menu-driven access to a plurality of databases hosted by multiple vendors. This advance was followed in 1985 and 1986 by similar videotex systems created by the vendors themselves, most notable of which are Dialog's "Knowledge Index" and BRS's "BRS After Dark." The major thrust of these "user friendly" systems was to eliminate the need for knowledge of search languages and to open up the stored data to access by unsophisticated users. All of the current systems are limited by computer technology of the user, however, in that a personal computer or terminal is required to gain access to them, a printer is required to download retrieved data, and an account is still required to to do business with each system provider. This invention solves all of these problems by making access to databases available via standard telephone and permit

ting the downloading of data either audibly or by FAX machine. The system is available to anyone with access to a touchtone telephone and therefore not restricted to computer owners. Re¬ trieved data may either be listened to over the telephone or received via FAX machine and therefore not restricted to owners of printers. The FAXed data is re-typeset by the system computer in order to make it more usable and efficient.

SUMMARY OF THE INVENTION

This invention permits a caller to obtain information from one or more databases using only a telephone and a FAX machine, if desired. The caller dials an 800 or 900 series telephone number and thereby gains access to a specially programmed micro¬ computer. The caller then hears a preprogrammed voice operated by digitally stored commands, instructing the caller in the use of the system, and makes suitable choices using the numbers on the touchtone keypad in order to select the service required. The system then asks for a search term or phrase to be run against millions of computer records for relevant material at which point the caller "spells" the term or phrase for the computer again using the keypad and a simple code. The computer then dials out to either a remote database (or set of databases) or an inhouse database, depending on the caller's needs and choices. An example of a remote database would be any of the databases available through one or more of the vendors. An example of an inhouse database would be a private file or one or more publicly avail¬ able files stored inhouse on a medium such as CD-ROM, which is a large read-only optical disk with attendant hardware.

Once the system computer has retrieved the relevant material, the caller may elect to hear the material read to him by a text- to-speech translation device or have it re-typeset and FAXed or mailed to him immediately. The caller may also request the system computer to go back for more information related to the search term or phrase.

It is therefore an object of the present invention to provide an information retrieval system which allows a telephone caller to search one or more databases.

It is another object of the invention to provide a system as described above to retrieve data audibly via the telephone or graphically by re-typeset FAX or mail delivery.

It is another object of the invention to provide a system as described above that avoids the requirement on the part of the user to own or have access to a personal computer, terminal, modem, and printer.

It is another object of the invention as described above to provide a system wherein a caller is not required to have estab¬ lished an account with a system provider or to possess a credit card in order to access a database.

It is another object of the invention to make the searching of database free of the restictions of specialized hardware and therefore increase the ease, effectiveness, and cost of database searching.

It is another object of the invention to provide an improved, efficient and more powerful graphic presentation of retrieved data over that delivered by the standard printer.

Other objects and advantages of the invention will be appar¬ ent to those skilled in the art from a reading of the following brief description of the drawings, the detailed description of the invention, and the appended claims.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 is block diagram showing a possible configuration of the present invention.

Figures 2 and 2b are a flowchart of the invention showing the programming of the system and user interaction with it. US91/02377

5 DETAILED DESCRIPTION OF THE INVENTION

In a first aspect , the invention is a method of supplying telephone access to one or more electronic data bases which method comprises the step of entering a search request either at the keypad of a touchtone telephone or the mouthpiece of a telephone ^• receiver .

In another aspect , it is a system for processing information entered as key hits at the keypad of a touchtone telephone or equivalent DTMF hardware said system comprising a programmed computer connected to the telephone or equivalent hardware , said computer capable, according to a code that equates the number of successive key hits of a particular key with a letter of the alphabet , of converting the number of successive key hits into digitally stored information representing that letter of the alphabet , said computer further capable of converting successively entered letters into digitally stored information representing words or phrases .

Optionally, said system ' s computer is capable of converting said digitally stored words or phrases into language that can be recognized as a search request by a data base; and/or is capable of transmitting, via a modem board, the search request to the data base in the language of the data base; and/or is capable of receiving, via said modem board, letters, words and phrases coded as digital data by the data base in response to the search request ; and/or is capable of processing the digital data received from the data base into one or more of the following formats : a format that can be transmitted via a FAX board to a FAX machine for printing of the letters, words and phrases by said FAX machine, a format that can be transmitted over a voice board to a telephone receiver so that the letters words and phrases can b«- heard by said caller, a format that can be transmitted for printing the words, letters and phrases in a format appropriate for mailing.

In a related aspect, the invention is a method for processing information entered as key hits at the keypad of a touchtone telephone or equivalent DTMF hardware said method comprising the steps of

(1) converting, according to a code that equates the number of successive key hits of a particular key with a letter of the alphabet, the number of successive key hits into digitally stored information representing that letter of the alphabet, and

(2) converting successively entered letters into digitally stored information representing words or phrases.

Optionally, the method further comprises the step of converting said digitally stored words or phrases into language that can be recognized as a search request by a data base and/or the step of transmitting, via a modem board, the search request to the data base in the language of the data base, and subsequently receiving, via said modem board, letters, words and phrases coded as digital data by the data base in response to the search request and/or the step of processing the digital data received from the data base to one or more of the following formats: a format that can be transmitted via a FAX board to a FAX machine for printing of the letters, words and phrases by said FAX machine, a format that can be transmitted over a voice board to a telephone receiver so that the letters words and phrases can be heard by said caller, a format that can be transmitted for printing the words, letters and phrases in a format appropriate for mailing.

In yet another aspect, the invention is a system for processing information entered by a voice at the mouthpiece of a telephone said system comprising a programmed computer connected to a telephone, said computer capable of converting words and phrases entered by the voice into digitally stored information representing said words and phrases, said computer capable of converting said digitally stored words or phrases into language that can be recognized as a search request by a data base, said computer capable of transmitting, via a modem board, the search^* request to the data base in the language of the data base, and said computer capable of receiving, via said modem board, letters, words and phrases coded as. digital data by the data base in response to the search request, optionally the system being further capable of processing the digital data received from the data base to one or more of the following formats: a format that can be transmitted via a FAX board to a FAX machine for printing of the letters, words and phrases by said FAX machine, a format that can be transmitted over a voice board to a telephone receiver so that the letters words and phrases can be heard by said caller, a format that can be transmitted for printing the words, letters and phrases in a format appropriate for mailing.

In a related aspect, the invention is a method for processing information entered by a voice at the mouthpiece of a telephone said method comprising the steps of

(1) converting words and phrases entered by the voice into digitally stored information representing said words and phrases,

(2) converting said digitally stored words or phrases into language that can be recognized as a search request by a data base,

(3) transmitting, via a modem board, the search request to the data base in the language of the data base, and

(4) receiving, via said modem board, letters, words and phrases coded as digital data by the data base in response to the search request; and optionally the step of processing the 8 digital data received from the data base to one or more of the following formats: a format that can be transmitted via a FAX board to a FAX machine for printing of the letters, words and phrases by said FAX machine, a format that can be transmitted over a voice board to a telephone receiver so that the letters words and phrases can be heard by said caller, a format that can be transmitted for printing the words, letters and phrases in a format appropriate for mailing.

In a further related aspect, the invention is a system for converting a succession of key hits made at the key pad of a touchtone telephone or equivalent DTMF hardware into digitally stored information representing a letter of the alphabet, said system comprising means for detecting successive key hits at a particular key, means for measuring the times interval between each pair of successive hits, means for determining whether each time interval was less than a predetermined time limit, means for determining how many successive time intervals each within the time limit occurred, and means for converting a particular number of successive time intervals each within the time limit to digitally stored information representing a letter of the alphabet; and, optionally, a means for converting digitally stored words or phrases into language that can be recognized as a search request by a data base and/or means for transmitting, in an audible manner via a voice board, said letter to a telephone receiver.

In a related aspect, the invention is a method for converting a succession of key hits made at the key pad of a touchtone telephone or equivalent DTMF hardware into digitally stored information representing a letter of the alphabet, said method comprising the steps of

(1) detecting successive key hits at a particular key, (2) measuring the times interval between each pair of successive hits,

(3) determining whether each time interval was less than a predetermined time limit,

(4) determining how many successive time intervals each within the time limit occurred, and

(5) converting a particular number of successive time intervals each within the time limit to digitally stored information representing a letter of the alphabet; and optionally the step of converting digitally stored words or phrases into language that can be recognized as a search request by a data base and/or the step of transmitting, in an audible manner via a voice board, said letter to a telephone receiver.

In an additional aspect, the invention is a system for reformatting a text file and transmitting it to a FAX machine for transmission which system comprises a computer or network of computers said computer or network comprising

(a) means for receiving a text file in computer code such as ASCII;

(b) means for identifying one or more elements of the text file, said elements selected from the group, title, author, bibliographic source, date of publication, and text of the article;

(c) means for assigning one or more printing specifications to each element, said specifications selected from the group, type face, type size, lead in, number of columns, left justification, right justification, framing, margin size, flush left, flush right, and center;

(d) means for processing each element such that each element is sent via a FAX board over a connecting line to a FAX machine where it is printed according to its printing specifications.

In a further related aspect, the invention is a method for reformatting a text file and transmitting it to a FAX machine for transmission which method comprises (a) receiving a text file in computer code such as ASCII; 10

(b) identifying one or more elements of the text file, said elements selected from the group, title, author, bibliographic source, date of publication, and text of the article;

(c) assigning one or more printing specifications to each element, said speci ications selected from the group, type face, type size, lead in, number of columns, left justification, right justification, framing, margin size, flush left, flush right, and center;

(d) processing each element such that each element is sent via a FAX board over a connecting line to a FAX machine where it is printed according to its printing specifications.

In another aspect, the invention is a system for accessing records in a group of at least two data bases in response to a search request specified by a user, said system comprising

(1) means for specifying a batch size defined as the number of records for which the contents of a specified field is to be transmitted to the user in response to the search request;

(2) means for specifying which field or fields in a record are to be searched;

(3) means for transmitting the search request to said data bases;

(4) means for determining the number N for a data base where N is the number of records in the data base that meet the specifications of the search request in a field to be searched;

(5) means for calculating a weighting factor for a data base, said factor equal to N/T where T is the total number of records, in the group of data bases, that meet the specifications of the search request in a field to be searched;

(6) means for calculating R for a data base where R is the integer closest to the product obtained by multiplying the weighting factor for that data base by the batch size; and

(7) means for selecting a total of R records from among the N records in a data base that meet the specifications of the search request in a field to be searched;

(8) means for transmitting to the user, from each of the R 11 records selected from a data base, the contents of a specified field.

In a related aspect, the invention is a method for accessing records in a group of at least two data bases in response to a search request specified by a user, said method comprising the steps of:

(1) specifying a batch size defined as the number of records for which ^"the contents of a specified field is to be transmitted to the user in response to the search request;

(2) specifying which field or fields in a record are to be searched;

(3) transmitting the search request to said data bases;

(4) determining the number N for a data base where N is the number of records in the data base that meet the specifications of the search request in a field to be searched;

(5) calculating a weighting factor for a data base, said factor equal to N/T where T is the total number of records, in the group of data bases, that meet the specifications of the search request in a field to be searched;

(6) calculating R for a data base where R is the integer closest to the product obtained by multiplying the weighting factor for that data base by the batch size; and

(7) selecting a total of R records from among the N records in a data base that meet the specifications of the search request in a field to be searched;

(8) transmitting to the user, from each of the R records selected from a data base, the contents of a specified field.

In another related aspect, the invention is a system for accessing one or more data bases which system comprises a first telephone, a voice board, a receiving means for receiving a telephone message and converting it to audible form, a keyboard for typing information into a computer, a computer and a-data base; said first telephone connected by a transmission line to the voice board and to the receiving means, said keyboard connected 12 to said computer, said computer connected to said data base; said computer comprising means for converting a word or phrase typed in at the keyboard into digitally stored information that if received by the data base can be recognized by said data base as a search request including the word or phrase; optionally, the system further comprises a second telephone connected to said first telephone by a transmission line, said second telephone optionally containing the receiving means and/or a message storage means for storing voice messages, said message storage means connected to by a transmission line to the first telephone.

In a related aspect, the invention is a method for accessing one or more data bases which method comprises the steps of

(1) transmitting a word or phrase over a first telephone to a remote receiving means that converts said word or phrase to audible form,

(2) transmitting via digitized voice the word or phrase back to the first telephone,

(3) transmitting over the first telephone a word or other indication that the digitized voice had correctly spoken the word or phrase,

(4) transmitting, via a keyboard located in the proximity of the receiving means, the word or phrase to a data base base as part of a search request; and,, optionally, the method, after step (3) but before step (4), further comprises the step of transmitting to the first telephone via digitized voice a sentence in the form of a question.

In each case, where a system is comprised of multiple means or capablities, it is possible to define a useful system comprised of fewer than all of said means. Similarly, where a method is comprised of multiple steps, it is possible to define a useful process comprised of fewer than all of said steps.

It is also clear that the instant application describes a variety of systems that can be Interconnected so as to provide an overall system with all or many of the capabilities described T/US91/02377

13 herein. Similarly, the instant application describes a variety of processes that can be combined.

Figure 1 is a block diagram showing the major components of the invention. Entry to the system is initiated by a caller from telephone 1 using an 800 or 900 series telephone number. The call goes out over line 2 to switch 3 of the system, which is of standard design. The switch determines which of the dedicated microcomputer boards 5 are available and sends the incoming call to the voice board 4 of the chosen microcomputer board. Both the microcomputer and the voice boards are of standard design. In one design the dedicated microcomputer boards which service an incom¬ ing caller are accessed by a dedicated or shared voice board 4 which converts retrieved digital data to voice for the caller and all boards are operated by a central controller board 6. This does not preclude other arrangements from being used, however. Microcomputer board 5 is connected to an outgoing modem board 11 which is in turn connected to a standard outgoing public tele¬ phone line, public data transmission line, or dedicated private line . The outgoing line is used by the microcomputer board 5 to make contact with one or more remote databases or inhouse data- basesl2. The database or databases accessed by the microcomputer boards 5 may be publicly available through one or more vendors, may be private files stored at the site of the system or at a remote location, or may be copies of publicly available files stored privately on a mass storage medium, such as CD-ROM.

Data retrieved from databases 12 is returned in digital form to the microcomputer boards 5. The microcomputer board and con¬ troller 6 process the data through voice board 4 to be read to the caller. Certain of the digital data from database 12 is sent to a central storage device 7 of standard design and forwarded to the FAX board 9 to be sent then to the caller's FAX machine 10 or mailed.

The modem boards in Figure 1, being of standard design, are replacable by equivalent technology, such as large scale communi- cations links betwee computers utilizing packet assemblers and disassemblers. Such alternative arrangements are a natural evolu tion in the standard technology of computers and are anticipated within the scope of this invention.

Figure 2 is a flowchart showing the operation of one embodi ment of the present invention. The caller uses a standard touch tone telephone to dial an 800 or 900 series telphone number to become connected with the system as indicated in block 13.

In decision 14 the caller is asked whether he wants to listen to preprogrammed subject matter or perform a search of database(s). If the caller chooses to listen only, he may then stop and exit the system as indicated in block 15 or he may elect to perform a search. In decision 17 the caller is asked whether he wants retrieved abstracts and text to be read to him by the system or FAXed to him automatically. Once the caller has elected to perform a search, he is asked by the system to enter a search term or phrase, as indicated in input 18, using the touchtone keypad to spell out the terms. Spelling is accomplished by se lecting letters from the keypad based on the following algorithm: hit the key once if the letter is in the first position on the key, twice if the letter is in the second position, and three times if it is in the third position. In an alternative embodi ment, the caller may speak the search term or phrase and the .pa system will recognize the speech. Such an alternative arrangement is within the scope of this invention.

Once the caller has indicated that he is finished spelling a search term or phrase, the system then takes the search phrase and translates it into the requirements of the database being searched, submits it to one or more databases and automatically retrieves relevant information, as indicated in block 19. Titles retrieved from multiple databases are weighted in block 20 according to number of titles in each database. The databases searched may be a combination of one or more public or private, and remote or inhouse files. This invention is not limited to the type of file searched. The system then reads the retrieved titles I⁵ to the caller, as indicated in block 21. The system computer either stays connected in an intermediate phase with the vendor of the database or disconnects.

The caller then determines which of the titles to retrieve abstracts, block 26, or full text for, 29 or whether to retrieve more titles, as indicated in block 25. Upon redialing the system, block 31, the caller can elect to have previously-chosen full text automatically FAXed, block 32, or mailed to him, block 33. The advantage of the present invention is that the caller is able to receive immediately full-text data re-typeset in such a manner as to resemble the original document before it was input into the database thereby making the data more comprehensible and effi¬ cient to use. The caller does not need to have access to computer and printer to receive data. The re-typeset data may also be mailed for next-day delivery, as indicated in block 30.

The following is a list of features of the present invention, one or more of which can be included in a particular embodiment:

1. Telephone access to the system

2. Automatic selection of one or more databases

3. Automatic translation of standardized commands into vendor language.

4. Automatic re-typesetting and FAXing of retrieved text to caller.

Touchtone typing

Each key in a subset of the typewriter keyboard maps to a certain number of successive presses of a touchtone keypad button. This mapping is logical and intuitive. Since most touchtone phones manufactured in the U.S. have three letters and a numeral on each button (e.g., 'A¹, *B', 'C, and '2' on button 2), one press of a button produces the first letter ('A'), two presses in quick succession produces the second letter ('B'), three presses produces the third letter ( 'C' ) , and four presses produces the numeral ( '2' ) .

The program written to recognize this form of input waits in a loop until the user presses a touchtone button, which it detects with DTMF-detection hardware. This first button-press triggers a timer that counts out a certain number of milliseconds (the default is 500 — half a second), waiting for the user to press the same button again. If the same touchtone is pressed a second time within the half-second time limit, the program assumes that the two button-presses should be taken together as part of a single letter (thus, if button 2 was pressed, the button-pres^'ses represent a single 'B' instead of two *A's) . The second button-press restarts the timer, and the program waits for a third press of the same button. This process continues until" (1) four presses of the same button have been registered; or (2) the user waits more than a half-second between two presses; or (3) the user presses a different button (clearly starting to enter a new letter).

The program then consults a table, using the number of the touchtone button that was pressed and the number of times it was pressed, to determine what the user's button-presses mean. If they signify a special function (e.g., button 1, the pound sign, the star button — see below) , that function is performed as described below. If they signify a character (e.g., 'A'), a recording of that character is echoed back to the user, and the character is added to the end of a buffer that represents the user's input line.

The letters A-Z (except Q and Z) and the numerals 0-9 are all printed on touchtone buttons, and map as described above.

Button 1, which has no letters on it, causes the program to go into '1-menu mode¹. In ' 1-menu mode', it reads a recorded menu of additional characters and functions: 2 for Q, 3 for ^'Z, 4 for parenthesis, 5 for truncation (a punctuation mark used in bibliographic searching), 6 for period ('.'), 7 for dash ('-'), 8 for ampersand ('&')» or 9 to toggle high-speed input on and off. After a choice is made, the program performs the function requested or echoes the character selected, and then returns to regular touchtone typing mode. If the pound sign is pressed, the program cancels the 1-menu and returns to touchtone typing mode immediately.

Choice 9 on the 1-menu toggles high-speed input on and off. In high-speed input mode, the maximum length of time between successive button-presses is reduced from 500 milliseconds to 275 milliseconds. (Note that this time limit applies only to multiple button-presses intended to signify a single character or function; the program places no restrictions on the amount of time between characters.) This mode . is designed for confident experts who are willing to punch the buttons more rapidly in return for a faster system response.

Button 0 triggers a help message that explains touchtone typing to a novice user.

Although button 0 and button 1 have special functions when pressed only once, when pressed four times they produce the numerals '0' and '1' respectively, in the same manner as the other numbered touchtone buttons.

The star button is used to back up and erase characters previously input. One press of the star button is analogous to the backspace key on the typewriter, backing up and erasing the previous character (the last character on the program's internal input buffer) ; two presses of the star button erases all characters entered so far (clearing the input buffer and starting over) .

The pound sign is used to terminate words and lines. One press of the pound sign is analogous to the space bar on the typewriter, inserting a space between words; two presses of the pound sign is like the carriage return key: it ends the line, which usually signals the end of the user's input.

An example of instructions audibly conveyed to the user are: "Your touchtone keyboard can be used like a typewriter, with three letters on each key.

Pressing a button once types the first letter, pressing it twice quickly types the second letter, and pressing it three times quickly types the third letter. Pressing four times types the 18 number of the button itself, zero through nine. To use characters not shown on a button, press button 1, and you will be able to choose the character you want.

Here are the three most important functions you should know. A pound sign enter a space between words; a double pound sign ends your search phrase; and a star backs you up."

Automatic Page Composition For Making Faxed Documents More Readable

When, in response to a search request, an article of text is transmitted by FAX to a user of the system, it is preferable that the faxed text be formatted in a manner characteristic of a text published in a newspaper or magazine. Such a format would typically have the title of the article at the top of the page and centered on the page, the name of the author (if any) just below the^"title and also centered. The text of the article would be below the author's name, possibly in two columns. Such a format produces a page that is easier and more pleasurable to read. Data bases, however, typically transmit their text in a format that is quite formal and stilted. For example, each line might start flush left, the first being, "/Tl: Economic Forecast", the second line being "AU=John Smith", the third "JN=The U.S. Economist, March 20, 1987" , with the text following.

Previous work by others has resulted in computerized systems whereby a text file, usually stored in ASCII, is processed by a computer for typesetting. In the present invention, a text file stored in ASCII or other computer code, is processed by a computer so that, without any type being set, it can be converted to a FAX format such as G3.

In the current invention, after a data base has been searched, the results are downloaded into the computer in the form of a standard text file. This text file is then sequentially submitted to a composition program which identifies each element of the down loaded record. The elements include the text of the article and identifying elements such as title, author, bibliographic source and date of publication. The composition program then assigns appropriate printing specifications for each element. Such specifications include typeface, typesize, lead-in, number of columns, right justification if desired, left justification if desired, framing if desired, size of margins, flush left setting if desired, flush right if desired, centering if desired. The specifications are those specified in advance by an operator of the system (i.e, someone in charge of the system, as opposed to someone who makes a search request, i.e., a user of the system.) The computer than processes each element so that each element is sent in ASCII to a FAX board such that the board converts the ASCII message to electronic signals that are transmitted to a FAX machine over a standard telephone line or other telecommunication linkage for printing according to the desired specifications.

Operator Eavesdropping

Normally, the system requires a touchtone phone for input of search topics and mailing addresses, using the Touchtone Typing method. For users with pulse/rotary-rdial telephones, and for touchtone users who do not want to bother with typing on their telephone keypad, the system can be modified to provide another method of entering alphabetic information - reading it aloud with an operator listening.

In this alternate method, when it is necessary to enter alphabetic information, the system tells the user to speak the words into the telephone, spelling aloud any difficult or unusual words. Then the system makes a one-way connection between the user and a keyboard operator, so that the operator can hear the user, but not vice versa. The operator listens to the user's words and types them on a standard computer keyboard, which directs them to the system as the user's input.

Once the user's input has been completely entered by the operator, the system spells it out aloud with digitized voice, and prompts the user to say aloud any changes he/she wants to make. The operator edits the entry, and stays connected until the user is satisfied.

If there are serious communication difficulties - the user's accent, a bad long distance connection, mumbling - the operator may press a button on his/her terminal and make the connection two-way, allowing misunderstandings to be resolved in a more human way.

Well-trained operators can do more than just type entries. An operator with a knowledge of Boolean searching and a familiarity with the system can- improve the user's search topic. With the addition of "message buttons" (buttons on the operator's key board that read canned voice prompts to the user, such as "Your search is too broad; can you be more specific?" and "Do you want to search on synonym X also?"), the operator can perform an online reference interview quickly and efficiently.

Sample session

The following illustrates a search on a system, named Three Magazines for purposes of illustration, that combines three full text bibliographic data bases into a single seamless service. The user's touchtone entries are underlined, and explanatory comments are in parentheses.

Welcome to Three Magazines, a FAX information service.

If you have a FAX number, enter it now as if you were dialing it, starting with your three digit area code.

If you don't have a FAX, press the pound sign.

2152548354

Your FAX number is:

2-1-5 2-5-4 8-3-5-4

If this is -correct, press 1. To change it press 2.

Enter your choice now, or press the star to repeat the choices,

Please use your touch tone keypad to enter the term or phrase to be searched.

For instructions, press zero.

Your touchtone keyboard can be used as a typewriter, with three letters on each key. Pressing a button once types the first letter, pressing it twice quickly types the second letter, and pressing it three times quickly types the third letter. Pressing four times types the number of the button itself, zero through nine. To use characters not shown on a button, press button 1, and you will be able to choose the character you want. Here are the three most important functions you should know. A pound sign enters a space between the words; a double pound sign ends your search phrase; and a star backs you up. Continue entering now.

(User uses keypad to type: ) KUWAIT AND IRAQ

Your topic is:

K-U-W-A-I-T space A-N-D space I-R-A-Q

To start searching, press 1.

To re-enter your topic, press 2.

Enter your choice now, or press the star to repeat the choices. We're searching eight hundred thousand articles for you.

It'll take about a minute.

Please wait...

Thank you for waiting.

We've found some articles for you to consider.

Our computer voice will read them to you now.

Title # 1, out of 2. (Article taken from a trade magazine.)

Iraq masses '100,000 troops near Kuwait border, (news item)

Do you want this article?

Press 1 for Yes, 2 for No, or 3 for Maybe.

To repeat the title, press the star.

We're sending the summary.

(Summary of article is being downloaded and faxed to user) Moving on to the next title.

Title #2, out of 2. (Article taken from a popular magazine.)

Kuwait and Iraq at odds over OPEC quotas. (column)

Do you want this article?

Press 1 for yes, 2 for No, or 3 for Maybe.

To repeat the title, press the star.

We're sending the summary.

That was the last title of the group.

To hear more titles, press 1.

To leave Three Magazines, press 2.

Enter your choice now, or press the star to repeat the choices, Thank you for using Three Magazines. Please call again.

Delayed Fax

In a preferred embodiment of the invention, if a user does not own or have a FAX machine readily at hand, the system offers a facility whereby a FAX transmission can be held up (for example up to 48 ^"hours) until the user is ready for it.

Delayed FAX is activated by entering a specified number, (for example, 999) at a specified time, (for example, when the system asks for a FAX number) . The system then assigns the user a serial number which applies only to the searching that will be done as a result of the present call and conveys this serial number to the user (for example, by reading it over the phone). The user may then search the system exactly as if a FAX number had been entered, but all data that would normally be faxed immediately is stored in the system's computer until called for later.

After- finishing the search on the system , the user has the time which transmission can be held up (e.g., 48 hours) to locate a FAX machine and call a number (e.g., a toll-free number), which reaches the system. The system requests first the user's serial number, then the telephone number of the FAX machine that will receive the saved data. The system then calls the specified fax number and finally delivers the saved data.

After the user has called for his saved FAX data, or after the time transmission can be held up (e.g.,48 hours) has elapsed, the saved data and its accompanying serial number is erased from the system computer and may no longer be retrieved.

Multidisciplinarv Searching Using Curve Sampling

A data base can be considered to be a collection of

"records" with each record divided into "fields of search". For example, a record may contain an entire magazine article, with the text of the article being one field, the title of the article of being another field,. the author of the article being another field and the name and issue date of the magazine being yet another field. Instead of magazine articles, other records might concern themselves with a newspaper article, a journal article, an abstract of an article, book or patent, or directory information; the foregoing are illustrative, other types of records and fields are also stored in data bases.

The system described herein can use a unique method to gather the most current, most relevant records from a number of databases of varied subject content, presenting them to the user one at a time in order of relevance. At its most basic level, the method has three parts: searching, calculating, and sampling.

In the searching phase, the method tests the user's search phrase in each database, constructing a table of "hits" in each database (the number of records containing the search phrase in one or more fields specified by either the user, the system or both), as in Table 1. This step may be accomplished by searching each database in turn or by searching all databases simultaneously.

Table 1. Searching Phase

Database Name Hits

In the calculating phase, the method adds up the total number of hits for the entire search. [The Tables can, of course, be created as digitally stored information within the computer.] It then creates a new column in the table, and calculates a weighting factor for each database, the number of hits in that data base divided by the total number of hits that were found in the group of databases. This table is then sorted on this weighting factor column in decreasing order, and databases with no hits are deleted, as in Table 2. The table is

Table 2. Calculating Phase

Database Name

Database 5 Database 2 Database 1 Database 3

Totals

In the final, sampling phase, a "batch" of hits is sampled from the curve, the size of the batch being predefined (e.g., 10) . The batch size is the total number of records or specified fields therefrom that will be sent to the user - unless the user requests additional records. The method starts within the first database in the sorted curve (i.e., the one with the most hits), multiplies its weighting factor by the batch size, rounds off to the next integer, and retrieves from it the number of hits equal to the result. The method proceeds to the next database and repeats the process to find out how many hits it will contribute, and continues down the curve until enough hits have been chosen to fill the batch, as shown in Table 3. The chosen hits are then retrieved, usually in the form of titles, and offered to the user for selection.

Table 3. Sampling Phase Database 5 provides 5 hits Database 2 provides 3 hits Database 1 provides 2 hits

Total hits provided - 10

If the user is not satisfied and desires more hits, another batch may be extracted from the curve in much the same way.

Table 4. Sampling Phase, second batch

Database 5 provides 5 hits leaving 30 hits Database 2 provides 3 hits leaving 17 hits Database 1 provides 2 hits leaving 12 hits

Total hits provided - 10

The net effect of the method is to produce manageable batches of hits, with each database represented in roughly the same proportions that it was represented in the entire search set. The batch of hits is sorted in favor of databases with more hits, those which are more likely to be relevant than databases with few hits. Also, bibliographic databases generally offer the most current data first (last-in-first-out order); therefore, most current and most relevant hits appear first in the resulting batch.

Title-only Curve Searching

In bibliographic searching, the curve can be enhanced to produce hits that are even more relevant, by limiting the search set to articles that contain the user's search phrase in the field of search that is the title.

By default, most database host systems look for a user's search phrase in the title, descriptors (subject headings) , and full text of each article. Searching all these fields together can lead to irrelevant hits, however, especially if the subject being searched is only mentioned once in the text of an article. If a search phrase is found in the title of an article, however, it is likely that the article gives much more information on the desired subject. Searching by the title only, therefore, yields fewer but far more relevant hits.

On the basis of this reasoning, the system can conduct its initial scan as narrowly as possible — for example, searching only by title field in bibliographic databases. If it finds no hits in any database, or if the user requests more hits than this narrow-scope search provided, the system does further searching with a broader search — for example, searching by the descriptor field in bibliographic databases — thus sacrificing relevance for more hits.

The invention should not be considered to be limited to the specific embodiments described above. The arrangement of the components in Figure 1 is but one example of how a set of stand ard hardware items can be arranged to deliver the unique combina tion of audio access to electronic databses with FAX delivery of data. Add that touchtone telephone can, of course, be replaced by equivalent Dual Tone Multi Frequency (DTMF) hardware. Add that mailing the search results can be done by any mailing procedure, but preferably by an overnight delivery system, such as Western Union Mailgram, in which case the results would be transmitted in an appropriate format to the delivery system. All such variations should be deemed within the scope and spirit of the claims that follow.

Claims

WHAT IS CLAIMED IS:

1. A system for processing information entered as key hits at the keypad of a touchtone telephone or equivalent DTMF hardware said system comprising a programmed computer connected to the telephone or equivalent hardware, said computer capable, according to a code that equates the number of successive key hits of a particular key with a letter of the alphabet, of converting the number of successive key hits into digitally stored information representing that letter of the alphabet, said computer further capable of converting successively entered letters into digitally stored information representing words or phrases.

2. A system of Claim 1 wherein further said computer is capable of converting said digitally stored words or phrases into language .that can be recognized as a search request by a data base.

3. A system of Claim 2 wherein further said computer is capable of transmitting, via a modem board, the search request to the data base in the language of the data base, and said computer is capable of receiving, via said modem board, letters, words and phrases coded as digital data by the data base in response to the search request.

4. The system of Claim 3 wherein said computer is further capable of processing the digital data received from the data base into one or more of the following formats: a format that can be transmitted via a FAX board to a FAX machine for printing of the letters, words and phrases by said FAX machine, a format that can be transmitted over a voice board to a telephone receiver so that the letters words and phrases can be heard by said caller, a format that can be transmitted for printing the words, letters and phrases in a format appropriate for mailing.

5. A system for processing information entered by a voice at the mouthpiece of a telephone said system comprising a programmed computer connected to a telephone, said computer capable of converting words and phrases entered by the voice into digitally stored information representing said words and phrases, said computer capable of converting said digitally stored words or phrases into language that can be recognized as a search request by a data base, said computer capable of transmitting, via a modem board, the search request to the data base in the language of the data base, and said computer capable of receiving, via said modem board, letters, words and phrases coded as digital data by the data base in response to the search request.

6..A system of Claim 5 said computer further capable of processing the digital data received from the data base to one or more of the following formats: a format that can be transmitted via a FAX board to a FAX machine for printing of the letters, words and phrases by said FAX machine, a format that can be transmitted over a voice board to a telephone receiver so that the letters words and phrases can be heard by said caller, a format that can be transmitted for printing the words, letters and phrases in a format appropriate for mailing.

7. A system for converting a succession of key hits made at the key pad of a touchtone telephone or equivalent DTMF hardware into digitally stored information representing a letter of the alphabet, said system comprising means for detecting successive key hits at a particular key, means for measuring the times interval between each pair of successive hits, means for determining whether each time interval was less than a predetermined time limit, means for determining how many successive time intervals each within the time limit occurred, and means for converting a particular number of successive time intervals each within the time limit to digitally stored information representing a letter of the alphabet.

8. A system of Claim 7 which further comprises means for converting digitally stored words or phrases into language that can be recognized as a search request by a data base.

9. A system of Claim 7 which further comprises means for transmitting, in an audible manner via a voice board, said letter to a telephone receiver.

10. A system for reformatting a text file and transmitting it to a FAX machine for transmission which system comprises a computer or network of computers said computer or network comprising

(a) means for receiving a text file in computer code such as ASCII;

(c) means for assigning one or more printing specifications to each element, said specifications selected from the group, type face, type size, lead in, number of columns, left justification, right justification, framing, margin size, flush left, flush right, and center; (d) means for processing each element such that each element is sent via a FAX board over a connecting line to a FAX machine where it is printed according to its printing specifications.

11. A system for accessing records in a group of at least two data bases in response to a search request specified by a user, said system comprising

(2) means for specifying which field or fields in a record are to be searched;

(3) means for transmitting the search request to said data bases;

(8) means for transmitting to the user, from each of the R records selected from a data base, the contents of a specified field.

12. A method for accessing one or more data bases which method comprises the steps of

(3) transmitting over the first telephone a word or other indication that the digitized voice had correctly spoken the word or phrase, and

(4) transmitting, via a keyboard located in the proximity of the receiving means, the word or phrase to a data base base as part of a search request.