On-the-fly extraction of hierarchical object graphs

Abstract

Reverse engineering techniques are usually applied to extract concrete architecture models. However, these techniques usually extract models that just reveal static architectures, such as class diagrams. On the other hand, the extraction of dynamic architecture models is particularly useful for an initial understanding on how a system works or to evaluate the impact of possible maintenance tasks. This paper describes an approach to extract hierarchical object graphs (OGs) from running systems. The proposed graphs have the following distinguishing features: (a) they support the summarization of objects in domains, (b) they support the complete spectrum of relations and entities that are common in object-oriented systems, (c) they support multithreading systems, and (d) they include a language to alert about expected (or unexpected) relations between the extracted objects. We also describe the design and implementation of a tool for visualizing the proposed OGs. Finally, we provide two case studies. The first study shows how our approach can contribute to understand the running architecture of two systems (myAppointments and JHotDraw). The second study illustrates how OGs can help to locate defective software components in the JHotDraw system.