Browse Prior Art Database

Revision Line Retrieval Scheme

IP.com Disclosure Number: IPCOM000077633D
Original Publication Date: 1972-Aug-01
Included in the Prior Art Database: 2005-Feb-25
Document File: 2 page(s) / 14K

Publishing Venue

IBM

Related People

Greek, JC: AUTHOR [+2]

Abstract

During the typing of dictated manuscripts it may happen that difficult words, such as personal names, are misspelt by the typist. If such misspelt words occur frequently in the text, it is tedious for them to be tracked down and corrected individually. The revision line retrieval scheme facilitates this work, assuming that an input/output typewriter is used in which the written text together with the control codes is entered into a bulk store, for example, a magnetic record medium.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 2

Revision Line Retrieval Scheme

During the typing of dictated manuscripts it may happen that difficult words, such as personal names, are misspelt by the typist. If such misspelt words occur frequently in the text, it is tedious for them to be tracked down and corrected individually. The revision line retrieval scheme facilitates this work, assuming that an input/output typewriter is used in which the written text together with the control codes is entered into a bulk store, for example, a magnetic record medium.

For revision, the text is processed section-by-section. Longer text portions (e.g., m characters) are transferred from the bulk store into an m character dynamic shift register acting as a buffer.

In addition to the data codes, control codes, such as dummy and space
(x) codes, delete codes, carriage return (CR), operation (OP), match flags (M), memory position mark (MP), etc., are stored, if necessary.

The unload buffer contains dummy codes, the MP mark, and the OP flag. The latter defines the position of the next character and the operating point in the register for output and revision operations.

For revision, the buffered text is sequentially searched for words identical to the misspelt word stored in a query register. Words can be compared either in their totality or to a particular recognition depth, with words or word parts found to be identical being M-flagged. Upon completion of the compare operation in the buffer, each M-flagged text line is printed for revision, ensuring that the editor need check and correct only those lines which contain at least one misspelt word.

It is assumed that the correct text portion to be recorded is: ---data codes and comparison codes in--- ---Coded information and control codes--- ---a single data
code input to a decode---.

However, due to errors in keying, the data and control codes actually stored are as follows: (CR)---(x)data(x)dodea(x)and (x)comparison(x)dodea(x)in(x)-- (CR)---(x)doded(x)information(x)and(x)control(x)dodes(x)-- (CRO--- (x)a(x)single(x)data(x)dode(x)input(x)to(x)a(x)decode(x) ---(CR).

It is furthermore assumed that during an initial print-out typical errors are noted. The relevant text portion is transferred from the bulk memory to the buffer register. As a typically misspelt word "dodea" (in lieu of "codes") is found, a search mode is initiated by depressing a search key. Characters "d", "o", "d", "e", "a", representative of the misspelt word, are sequentially entered into a query register. The compare operation begins by comparing first the keyed "d" with each first character of each text word so that in the first operational step text characters su...