Representing constraint business rules extracted from legacy systems

Abstract

Constraints are a class of business rules that most information systems implement. However, due to staff turnover and lackof documentation, precise knowledge of what constraints are enforced by a system is often not available. This can seriously hinder an organisation’s ability to understand the data stored in its systems, and to evolve the systems to implement new business policies. To help the situation, researchers have considered how to extract constraints out of legacy systems. While some powerful methods have been proposed for identifying constraints in application programs, little has been done so far to help users to comprehend the recovered constraints. To step up research in this direction, we study in this paper how the recovered constraints should be represented, so that they can be analysed, processed and then presented to the user in a comprehensible manner. We introduce a representation language that offers a balance between expressiveness, comprehensibility and reasoning power in handling the recovered constraints.