Abstract

The purpose of this paper is to introduce a framework for assessing the expressive power of data models and to apply this framework to the relational model of data. From a designer's point of view, a data model such as the relational model should not only be formally defined and easy to understand, but should also provide a powerful set of constructs to model real-world phenomena. The expressive power of a data model, defined as the degree to which its constructs match with constructs encountered in reality, can be judged by two complementary principles: the interpretation principle and the representation principle. It is asserted that database designers attempt to minimize the number of ad hoc database constraints, and that a data model faithful to the two principles supports this design strategy. Subsequently, this constraint minimization strategy is used to assess the expressive power of a particular data model, i.e., the relational data model. The authors take the position that the expressive power of the relational model is not optimal, due to a lack of adherence to both the interpretation principle and the representation principle. The paper amplifies this position by means of a number of examples, all based on publications by Codd and Date.

Share

COinS