Browse Prior Art Database

Abstract Coding of Textual Data

IP.com Disclosure Number: IPCOM000076042D
Original Publication Date: 1972-Jan-01
Included in the Prior Art Database: 2005-Feb-24
Document File: 3 page(s) / 60K

Publishing Venue

IBM

Related People

Lindsey, RD: AUTHOR [+2]

Abstract

Since searching each track of data recorded on a magnetic card for a particular track containing a desired record is time consuming, it is desirable to search in detail only those tracks where the desired record is likely to reside.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 3

Abstract Coding of Textual Data

Since searching each track of data recorded on a magnetic card for a particular track containing a desired record is time consuming, it is desirable to search in detail only those tracks where the desired record is likely to reside.

Illustrated is a flow chart of the searching out of prerecorded information on magnetic cards for a comparison to a query that is entered at the keyboard prior to the searching. The basic components of the system are a processing unit which contains a stored program of the machine definition, the magnetic card reader/recorder, and an IBM Magnetic Card "SELECTRIC" Typewriter (MC/ST) printer/keyboard. The information on the magnetic cards will be referred to as Index Records. Track one of each magnetic card is reserved for the abstract data of the Index Records that are contained on tracks 2-50 of the card. This abstract is automatically prepared by the machine the first time the card is searched and is recorded on track one. The abstract then is used on all succeeding searches to increase the searching rate. That is, during a search the abstract is searched to identify those tracks which must be searched in detail. Without abstracting, it would be necessary to read each Index Record and perform a detailed comparison with the query.

The following explanation of the abstracting is divided into two sections: (1) index record and query abstract preparation, and (2) use of abstracting to increase the searching rate. The flow diagram of (2) is shown. (1) All the possible printed characters that can be recorded on a magnetic card with the use of the MC/ST are divided into twelve approximately equiprobable groups. The division of the characters into twelve groups is made using published data of the probability of occurrence of characters in the English language, etc.

When a record is searched during the compiling of an abstract, the first character of each descriptor item is analyzed to determine which one of the 12 groups that it falls into. Each time a particular group is counted, one of twelve bits in two predefined registers is set (six in each register). Following the completion of the compiling of this information for an Index Record, the two six- bit characters are stored in a buffer area until two characters are stored for...