A conceptual approach to object-oriented data modeling

Abstract

Object-oriented data modeling is starting to replace the relational model for many recently emerging database applications. The complex nature of these databases precludes mapping of their data directly into a tabular relational structure. Current object-oriented data modeling lacks the standardization and mathematical soundness of the relational model. This thesis addresses this problem by proposing a conceptual data model called OPERA (Object Paradigm / Entity-Relationship Approach). OPERA incorporates the static features of the Entity-Relationship Model with the dynamic properties of object- orientation. In addition to OPERA, an object-oriented extension to the graphical query language GORDAS (Graph-Oriented Data Selection) is proposed. To demonstrate the effectiveness of the proposed model, we will model a United States combat systems support database, the EWIRDB (Electronic Warfare Integrated Reprogramming Database). We map the EWIRDB from its basic relational format to an object schema and then to an OPERA graph. Finally, this conceptual schema is mapped to a GORDAS schema graph and queries are performed on the database. OPERA is conceptually superior to the ER Model and its object-oriented variant, the Enhanced Entity-Relationship Model (EER) Model. We demonstrate this by representing methods as relationship types, which the ER and EER models are incapable of OPERA also aids in query formulation for visual query languages such as GORDAS by providing a query graph mapping template