(A) form driven object-oriented reverse engineering methodology

Abstract

Most business organizations use business applications that support operations successfully, but existing applications, which have stable capabilities in present operations, are no longer suitable. Legacy applications are valuable assets; it is probably not an appropriate measure for most business organizations to give up the legacy system in the near future. However, the legacy system has been experiencing typical problems. First, the legacy applications utilize system-specific information, but they sometimes work without any documents. As a result, it is difficult to have further developments because of the lack of the business knowledge about legacy applications and the understanding of systems. Second, they are complex and expensive to maintain. In particular, their function can not be divided into more manageable components. Third, most legacy systems use technologies developed in the 1970s. They are host-based, and not easily integrated with current needs. There are several strategies to resolve the problems of the legacy applications. A desirable solution is to develop the applications which are easier to maintain. In this case, the migration strategy is crucial to recover the semantic of the legacy applications. To solve this problem, a reverse engineering technology is introduced. One of its major tasks is to understand the semantics of existing software; the other is to attract the recovered semantics to the abstractions of higher level. In this dissertation, we define the reverse engineering in terms of acquiring a higher conceptual abstraction from the physical systems. The object-oriented modeling technique is employed because objects can provide consistent services to both legacy and next generation applications. A main core of the object-oriented modeling is to identify objects. It may be easier to reengineer the current systems by using the objects. This dissertation proposes a form-driven object-oriented reverse engineering (FORE) methodology...