The Data Model Compiler: a tool for generating object-oriented database systems

Abstract

The Data Model Compiler project represents an effort to automatically produce object-oriented database systems. An analysis of data models of this genre leads to the conclusion that their significant differentiating characteristic is the set of fundamental, or built-in, relationships. This observation has lead to the development of a methodology for the generation of object-oriented data models which revolves around the specification of the basic relationships. Initially, the project focused on the development of a variety of object-oriented models in order to gain an understanding of the problem space. This paper concentrates upon that experience with a brief discussion of future plans. The particular application domains for which models have been constructed include business data processing, mechanical CAD, office information systems, and software performance modelling. The knowledge gained by developing these models served as the basis for the specification of a set of essential relationship characteristics. The plan for generating database software corresponding to a user-specified data model revolves around the definition of the fundamental relationships of the model in terms of the aforementioned characteristics.