Database Design Revisited

Abstract

Information Systems design is affected by the simplicity of the relational model types, though these types do not correspond directly to entities of real applications. In this paper, we present another approach, where the Information System designers would be able to directly represent the real world in a database model, very close to the Entity-Relationship (ER) model. A query and manipulation language that can handle composite data types, close to ER diagrams, is the Conceptual Universal Database Language (CUDL). We demonstrate that a database modelled by ER diagrams can be directly expressed to the CUDL Abstraction Level (CAL), the data model corresponding to CUDL, by a set of rules for direct transformation of ER diagrams to CAL. This set consists of a basic set of five rules, which is extended with an extra rule that deals with specific situations appearing in practical applications. Consequently, the resulting more powerful and composite data can be directly maintained with the CUDL. In this way, the development process consisting in conceptual design (ER), transformation to Logical Relational Schema, usage of SQL for data manipulation/retrieval and the reverse steps to present the results in the conceptual level is simplified by conceptual design (ER), transformation to CAL and the usage of CUDL for direct manipulation/retrieval of real world (conceptual) structures.