The development of a partial design recovery environment for legacy systems

Computer Assisted Program Understanding systems take input primarily in the form of source code and produce output representing system concepts in some useful form. This paper discusses the development of a program design recovery environment based on structural and behavioral recognition of programming plans. Our research investigates program and plan representation methods and the issues related to the detection of code fragments using pattern matching techniques. In particular, we consider the integration of diverse tools allowing selection from a choice of strategies. Our investigation focuses on techniques to allow partial design recovery when complete recognition is not feasible. Finally, we discuss the underlying process paradigm, called "Goal-Question-Analysis-Action".