A Combined Global-Analytical Quality Framework for Data Models

In the domain of data model quality two independent approaches can be identified. The first one proposes a global view mainly based on quality models and frameworks, focusing on high level quality characteristics such as minimality, maintainability and evolvability and on metrics for measuring them. A second research track has concentrated for decades on the analysis of specific problems, ranging from unnormalized structures to unsatisfiability. The latter proposes means for formalizing, detecting and correcting particular defect patterns. Both of these approaches address data model quality issues, but in independent ways. In this paper, we present an attempt to address database schema quality through both approaches in a common framework. We summarize the main concepts and reasoning basis of a project devoted to database schema quality. We propose an operational framework that combines the contribution of both global and analytical views of quality. Our global view focuses on defects categories to evaluate schema quality and error side effect. Our analytical view translates into detection and correction methods of these defects. The final purpose of this work is to propose a precise, intuitive and easy to use quality management methodology for database schema.