Browse Prior Art Database

Interpretive Retrieval Information Subsystem

IP.com Disclosure Number: IPCOM000079645D
Original Publication Date: 1973-Aug-01
Included in the Prior Art Database: 2005-Feb-26
Document File: 2 page(s) / 14K

Publishing Venue

IBM

Related People

Dennison, RE: AUTHOR [+3]

Abstract

The Interpretive Retrieval Information Subsystem (IRIS) incorporates a technique for searching a large data base, using words contained in the record for identification of the desired record instead of the actual location of the record. To retrieve data, it is necessary only to specify the general area of the data base and some combination of words known to exist in those records. Informational searches may also be done, by specifying the desired type of data in words and letting IRIS find out if any such data actually exists.

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

Page 1 of 2

Interpretive Retrieval Information Subsystem

The Interpretive Retrieval Information Subsystem (IRIS) incorporates a technique for searching a large data base, using words contained in the record for identification of the desired record instead of the actual location of the record. To retrieve data, it is necessary only to specify the general area of the data base and some combination of words known to exist in those records. Informational searches may also be done, by specifying the desired type of data in words and letting IRIS find out if any such data actually exists.

Data is stored in a relatively random fashion keeping data organized as to general area of content only. Records are assigned an arbitrary key or Record Number. These numbers are managed by a data base access method. The records are stored randomly on a disk file, and the record numbers are indexed on a drum file for those files requiring fast access. Slower files use disk indexes.

All information is cross-referenced by word and record location. The cross- reference operation breaks a record into words and records the location in a Bit String, which is a positional representation of the record location. These Bit Strings are kept to a reasonable length by the data record groupings. A group of closely related records is called a File. A group of loosely related files is called a Library. Libraries are grouped into Library Sets by application.

The words for a particular Library are recorded in a Word List data set. Each word has associated with it a unique number and indicators to enable the Bit String in the Bit String data set, for that word for a particular file in the Library, to be fetched. Bit String data has unused positions of reasonable length eliminated to conserve space in the Bit File. The combination of a Word File and Bit File is called a Glossary.

The Bit Strings are used for more than just keeping track of which record in a file that a given word is located. They are used to perform the logical operations indicated by the operators between the words in the search argument.

The three Boolean operators AND, OR and NOT are used to combine words in the search argument. For example, if records containing both words A and B are to be located, the search argument is: A and B. IRIS takes the word A, finds it in the word list and fetches the Bit String for that word for the selected file. Then the Bit String for word B is similarly fetched. The two Bit Strings are then ANDed to produce, as a result, a Bit String which represents all the rec...