Publications

Abstract : Visualization of program execution is a powerful means to comprehend program behavior. Sequence diagrams have long been used for representing the control flow of object-oriented programs. Since the sequence diagrams generated from run time can become large and unmanageable, several compaction measures are proposed to represent the diagram in a concise way. While the compact diagrams provide a summary of the execution, means to explore them in a systematic way are necessary. Hence, user-driven exploration methods have been proposed. One of them is the use of hyperlinks to unveil the depth of the sequence diagram on-demand. The compacted sequence diagram provides more flexibility to interactively explore the compacted regions in a hierarchical fashion. The user decides which regions to expand and up to what level of depth they must be expanded through the support of hyperlinks in order to comprehend the program. However, some of the crucial logic embedded inside the methods remain hidden. In this work, we show how an activity diagram can be constructed from the hyperlinked sequence diagram to portray the control flow logic at the sub-method level at different levels of granularity. Although the idea of an activity diagram for the entire execution can attract criticism owing to its large size, our interest is from the standpoint of portraying sliced executions on-demand. We contend that sliced activity diagrams can provide valuable insights into the execution by restricting the focus on the essentials. Finally, we show the utility through a variety of experiments.