Visualizing Multiple Program Executions to Assist Behavior Verification

Visualization techniques have been widely used in representing software artifacts. They play a central role in conveying program information to software developers. While numerous tools have been developed to visualize information such as static software architectures, dynamic program behaviors, and debugging processes, little attention has been paid to visualizing correlations and variations among program representations. This paper investigates the visualization of cross-references across multiple program executions based upon different testing inputs so that meaningful and viewable properties can be presented to the viewpoint from different perspectives. Visualizing such a comparison can help feature location and program behavior verification. It also helps programmers better understand and test their software which can have a significant impact on improving its reliability.