Design recovery of legacy database applications based on possibilistic reasoning

Industrial database applications often evolve over three or more generations of developers, cover several hundred thousand lines of code and maintain a vast amount of data. A rapidly growing number of companies face the problem that they have to adapt or modernise such existing legacy database applications (LDA) in order to keep up with emerging requirements. The documentation of such LDAs is often obsolete as they have been developed over several generations of programmers. This paper presents an application of possibilistic reasoning to infer the semantic information that is necessary to recover the conceptual design of an LDA. A dedicated graphical language, called generic fuzzy reasoning nets, is introduced to specify and customise the applied reverse engineering process. The actual reasoning process is performed by a nonmonotonic inference engine based on fuzzy Petri nets which supports lazy execution of expensive analysis operations