An approach to analyzing the information content of existing databases

Abstract

Traditionally, application databases have been developed for efficiency of computer operations. The logical structure of the data is buried in the database schemas, application programs, and users' and programmers' minds. To manage information as a corporate resource the logical data structures must be made explicit. In this way organizational databases can be conceptually integrated. The magnitude and complexity of corporate databases makes a manual approach to this task infeasible. We propose a computer aided approach. The approach is based on extracting implicit functional dependencies from existing database schemas and forming an Entity-Relationship representation. Where the semantics of the schema are unclear, rules are applied and additional input sought from the Data Administrator. Because of its widespread use in production systems, we focus on CODASYL Network DBMSs in this paper. The approach also has applications in databases from one DBMS to another.