Interactive 3-D Visualization of Feature-traces

The maintenance or reengineering of an object-oriented system includes its reverse engineering. In other words its internal structure and behavior needs to be understood. Many researchers have proposed different techniques to support the reverse engineering effort. The two predominant approaches to reverse engineering are: • static analysis of source code • dynamic analysis of behavior of a system at execution time. Both techniques have strengths and short comings. The static analysis of an object-oriented software system face difficulties such as polymorphism and it may be difficult to correlate parts with system functionality. Dynamic analysis approaches generally do not achieve full coverage of a software system. Moreover, due to the volume of data captured during dynamic analysis of a system, it is difficult to infer high-level views. In this thesis we address this issue of software maintenance and reengineering and propose a novel visualization technique which combines static analysis of source code with dynamic information extracted by exercising features of a system. We refer to these as dynamic feature-traces. This technique supports the software engineer in understanding the behavior of software systems by visualizing it in terms of its internal structure. Using our visualizations we focus on modeling features of a system, how to detect and locate them in the source code, hotspots of behavior and feature interaction as a means to understand how different features behave in a software system.