Affiliation:
1. Massachusetts Institute of Technology, Cambridge, Mass.
Abstract
This paper examines the concept and implications of virtual information in data base systems. Virtual information is any fact which does not physically exist in the data base, but is nonetheless accessible through combinations of algorithms and other data. Physically recorded information is only one of a number of ways to obtain information from a data-base system. Viewing an information system as a collection of functions shows that pure data and pure algorithm from the endpoints of a spectrum of ways function values can be realized, with the middle range being various types of virtual information. Several classes of virtual information are identified, and their usefulness is examined to show the appropriateness of the concept in a data-base system. Finally, the model is evaluated in light of the implications of virtual information for inference and automatic restructuring within a data base.
Publisher
Association for Computing Machinery (ACM)
Subject
Information Systems,Software