Rule management and inferencing in relational databases

To handle facts and rules in a database system, it is desirable to have a uniform representation and manipulation scheme for both. In such a case, rules and data are entered, deleted, updated, and queried in a transparent way. In order to support recursive queries and transitive closure queries, a homogeneous approach is very attractive, since it allows rules to be treated as data in a conventional relational database. To be able to infer information from these rules in addition to the stored data in the database, it is necessary for the rules to be triggered and executed as and when required. The authors discuss these issues and the requirements and features of such a system. They also provide a description of a prototype system developed in a Prolog environment and the core manipulating capabilities and inferencing strategy provided for its proper operation