A semantic approach in designing relational data bases

Abstract

This paper proposes a semantic approach to data base design in the relational approach. The analysis on entity types, associations and attributes presented may be applied to data base design in other approaches too. Entity types, associations, and attributes in an operational environment are recognized and their semantic properties are studied. The attributes are categorized into simple and repeating, and the associations are categorized into atomic and molecular. Based on the analysis, relation schemas are established for the data base. These relation schemas are mostly in fourth normal form. Further normalization and optimization guarantees that optimal fourth normal form relation schemas are to be obtained. In this paper, the concept of semantically irrelevant functional dependency is also proposed. If this concept is adopted, then we may have a more optimal data model, which is semantically sound, than the optimal fourth normal form data model.