Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

SYLLOG: A KNOWLEDGE BASED DATA MANAGEMENT SYSTEM

IP.com Disclosure Number: IPCOM000128181D
Original Publication Date: 1981-Dec-31
Included in the Prior Art Database: 2005-Sep-15
Document File: 19 page(s) / 50K

Publishing Venue

Software Patent Institute

Related People

Adrian Walker: AUTHOR [+3]

Abstract

An experimental data base system, called SYLLOG, is described. The system, which has been prototyped in the language SETL, provides a screen-oriented English-like language for use by non-programmers in setting up and using a data base. To set up a new data base, some standardized English sentences are typed in, and are combined into syllogisms which indicate how the data will be interpreted. Then, once the data have been loaded, the knowledge in the syllogisms is used for retrievals. The knowledge is used for retrievals by. a backchaining algorithm which operates on the syllogisms alone. A tree resulting from the backchaining controls an iterative algorithm which searches the data base. It is shown that the combined backchain-iteration algorithm is correct for schemas in which no syllogism calls itself, and that under this restriction, the query language is at least as powerful as the relational algebra. An extension is described to handle recursive syllogisms, such as' those which yield the transitive closure of a relation.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 9% of the total text.

Page 1 of 19

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

SYLLOG: A KNOWLEDGE BASED DATA MANAGEMENT SYSTEM

Adrian Walker

JUNE 19 81.

Report No. 034 (c) Adrian Walker 1981 All rights reserved

Th.i.s work. teas supported by the National Science Foundation under Grant No. NSF'.-MCS- 80-043+9.

ABSTRACT

An experimental data base system, called SYLLOG, is described. The system, which has been prototyped in the language SETL, provides a screen-oriented English-like language for use by non-programmers in setting up and using a data base.

To set up a new data base, some standardized English sentences are typed in, and are combined into syllogisms which indicate how the data will be interpreted. Then, once the data have been loaded, the knowledge in the syllogisms is used for retrievals.

The knowledge is used for retrievals by. a backchaining algorithm which operates on the syllogisms alone. A tree resulting from the backchaining controls an iterative algorithm which searches the data base. It is shown that the combined backchain-iteration algorithm is correct for schemas in which no syllogism calls itself, and that under this restriction, the query language is at least as powerful as the relational algebra. An extension is described to handle recursive syllogisms, such as' those which yield the transitive closure of a relation.

1. INTRODUCTION

The relational model for data bases [4l effectively frees a user from the details of: physical access to data, and it provides an uncluttered framework in which such topics as normalization and the meaning of updates can be discussed [5,71. Yet, for people who are not programmers or mathematicians, relational data bases can be difficult to use, even when provided with a high level query language such as Query-by-Example [121. One view of this difficulty is that, while the user knows many common sense rules about the real world situation which a data base describes, the data base system does not (it only has the raw data) , so there is plenty of room for misunderstandings.

In work on computer systems which can represent the knowledge of a human expert, [91 the emphasis is on capturing everyday rules of thumb about a particular subject (e.g. medical diagnosis), and then using the rules to make deductions. Very often, such rules are expressed in a form

If conjunction of premises then conclusion

and are called production rules. Such rules are chained together to make deductions.

New York University Page 1 Dec 31, 1981

Page 2 of 19

SYLLOG: A KNOWLEDGE BASED DATA MANAGEMENT SYSTEM

Clearly, both the relational model of data and the production rule model share some features with the first order predicate calculus [31. A relational data base can be viewed as a set of explicitly listed predicates (a model), and a set of production rules call be thought of as rules of inference for making deductions. However, logic, as a formalism for everyday computer use, is beset by the problem that its notation is di...