FRAMEWORK FOR THE DESIGN AND EVALUATION OF REVERSE ENGINEERING METHODS FOR RELATIONAL DATABASES

It is often difficult to obtain a good conceptual understanding of a l egacy database, especially when there is a lack of documentation. Data base reverse engineering attempts to provide solutions for this proble m. It is the part of system maintenance work that produces a sufficien t understanding of a legacy database and its application domain to all ow appropriate changes to be made. However, research on database rever se engineering has largely ignored design and evaluation issues of the ir methods (i.e., foundations and processes). This research proposes a framework for the design and evaluation of reverse engineering method s of relational databases. This framework consists of eight criteria: 1) the situation chosen as the basis for reverse engineering, 2) the c onceptual model chosen to represent the reverse engineering results, 3 ) the prerequisites of the database to be reverse engineered, 4) the t horoughness of domain semantics acquisition, 5) rules and heuristics e mployed by the reverse engineering process, 6) performance efficiency of the reverse engineering process, 7) completeness and robustness and 8) validation. These criteria are important to be considered in desig ning reverse engineering methods, so that they can perform reverse eng ineering for a broad range of legacy databases at a high level of auto mation and provide a conceptual schema that is semantically rich and c orrect.