STIRDAT: A system tightly integrating relational databases and a theorem prover

Abstract

There has been a growing awareness of the need to share data and services among autonomous heterogeneous information sources. For obvious reasons, these sources are storing information in ways that meet their own needs and hence in different formats. However, for a combination of information from heterogeneous sources to be truly useful, a user needs to be able to use the shared information without having to learn the particular data format or the locations of the relevant data. This idea can be taken a step further as the importance of combining techniques from various areas in computer science becomes more obvious. There are a number of benefits that can be derived from combining, for example, some of the features of Artificial Intelligence systems, such as logical inference, with those of traditional database systems. This paper describes STIRDAT, a system tightly and transparently integrating data in a number of relational databases stored at different sites with a theorem prover. It provides a clear illustration of the benefits to be gained from a system combining techniques that have proven independently useful in different areas of Computer Science.