Constellation visualization: Augmenting program dependence with dynamic information


This paper presents a scalable, statement-level visualization that shows related code in a way that supports human interpretation of clustering and context. The visualization is applicable to many software-engineering tasks through the utilization and visualization of problem-specific meta-data. The visualization models statement-level code relations from a system-dependence-graph model of the program being visualized. Dynamic, run-time information is used to augment the static program model to further enable visual cluster identification and interpretation. In addition, we performed a user study of our visualization on an example program domain. The results of the study show that our new visualization successfully revealed relevant context to the programmer participants.

2011 6th International Workshop on Visualizing Software for Understanding and Analysis (VISSOFT)