Many research and development efforts are underway to make database technology suitable for CAD applications. This thesis discusses some of the database requirements for CAD applications which are fundamentally different from those for business applications. The differences in database requirements between business and CAD applications make it difficult to adapt commercially available database management systems to CAD applications.
In recent years, many data models for organizing the data have been proposed to capture more of the semantics of the data for business applications, but these data models are not adequate for design databases and are difficult to apply. Although the relational model is most attractive for design databases among data models developed earlier, ordering and member duplication which are important for CAD applications are not directly supported by the data model.
In this thesis, we extend the relational model to support tuple ordering and tuple duplication, and present an extended relational algebra which is a collection of operators on extended relations as the basis for a high-level data language. Our data model, called the Extended Relational Model, which includes the classical relational model as a special case can achieve a high degree of data independence and is based on Childs`` extended set theory.
We also discuss the concepts of extended data dependencies for the extended relational model and the conditions under which an extended relation can be decomposed into its projections without loss of information. Three types of extended data dependencies about extended relations are studied and related to the decomposition process of extended relations. The extended data dependencies provide a means of incorporating data semantics into the extended relational model.