Browse Prior Art Database

Universal Database Manager

IP.com Disclosure Number: IPCOM000047281D
Original Publication Date: 1983-Oct-01
Included in the Prior Art Database: 2005-Feb-07
Document File: 3 page(s) / 31K

Publishing Venue

IBM

Related People

Giles, DA: AUTHOR [+4]

Abstract

This article describes a mechanism supporting a field-oriented model of data, which is capable of supporting any of the known architectures for database management. This mechanism is based on the triple data model, which has been described elsewhere, and is referred to as the "universal triple machine". It is capable of supporting several different data models at the same time, and also the structures needed for data dictionaries and generalized knowledge representation. A database is usually organised according to some "data model" which places a built-in structure on the data stored, e.g., a tabular structure. Any data model provides a method of "factoring" the stored data, which simplifies some aspects of data handling.

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 49% of the total text.

Page 1 of 3

Universal Database Manager

This article describes a mechanism supporting a field-oriented model of data, which is capable of supporting any of the known architectures for database management. This mechanism is based on the triple data model, which has been described elsewhere, and is referred to as the "universal triple machine". It is capable of supporting several different data models at the same time, and also the structures needed for data dictionaries and generalized knowledge representation. A database is usually organised according to some "data model" which places a built-in structure on the data stored, e.g., a tabular structure. Any data model provides a method of "factoring" the stored data, which simplifies some aspects of data handling. It follows that an application which attempts to interpret the stored data must "understand" the built-in structure in order to produce meaningful results. The data model may therefore be regarded as additional semantics of the application. At the same time the data model limits the information which may be stored, since the built-in structure prohibits storage of some facts. A data model which does not suffer this disadvantage must have no built-in structure, and is said to be free of semantics. This means that an application need have no prior knowledge, since all information is contained in the database. The triple data model has been proposed elsewhere and shown to meet this criterion. A triple is an elementary ordered grouping of three fields, which is used to record an elementary fact, such as: "SMITH" "HAS SALARY" "10,000"

It is therefore analogous to an elementary sentence. Collections of triples are capable of representing arbitrarily complex information structures. In particular, some triples may represent dictionary information which is used in interpreting other triples. The universal triple machine provides a means of recording a proper name (referring to some real world object) and associating it with a unique identifier generated by the database manager. This unique identifier is known outside the database system and may be used in place of the proper name, if desired. It may also be used where no proper name exists. Identifiers may have fixed or varying length and are used in a triple in place of proper names, which may not be unique. Six basic operations are defined on the triple machine as follows: 1. INSERT-NAME (name) --> id

2. DELETE-NAME (id)

3. FIND-NAME (id) --> name

4. INSERT-TRIPLE (id,id,id)

5. DELETE-TRIPLE (id,id,id)

6. FIND-TRIPLE (id,id,id) --> id,id,id The last of these operations provides an information retrieval facility. It returns all triples which match the arguments supplied. Dummy arguments may be used which match against any identifier. Hence, this operation is capable of listing the entire database. A useful extension to the FIND-TRIPLE operation permits a name to be substituted for an id in any argument, and thus provides symbolic searching. An...