Using Composition Trees to Model and Compare Software Process

Lian Wen, David Tuffley, Terry Rout

Abstract: Software processes described by natural languages are frequently ambiguous and it is usually difficult to compare the similarity and difference between one process defined in one standard and its counterpart defined in another standard. This paper proposes Composition Tree (CT) as a graphic language to model software process based on its purpose and expected outcomes. CT is a formal graphic notation originally designed for modeling component based software system. This paper demonstrates that CT can be a powerful notation to give a clear and unambiguous description of a software process as well. This paper also investigates an algorithm which can compare two CT-modeled processes and provide an intuitive view called a Comparison Composition Tree (CCT) to highlight the differences and similarities between the two processes.