Visual exploration of large-scale evolving software

The comprehensive understanding of today’s software systems is a daunting activity, because of the sheer size and complexity that such systems exhibit. Moreover, software systems evolve, which dramatically increases the amount of data one needs to analyze in order to gain insights into such systems. Indeed, software complexity is recognized as one of the major challenges to the development and main-tenance of industrial-size software projects. Our vision is a 3D visualization approach which helps software engineers build knowledge about their systems. We settled on an intuitive metaphor, which depicts software sys-tems as cities. To validate the ideas emerging from our re-search, we implemented a tool called CodeCity. We devised a set of visualization techniques to support tasks related to program comprehension, design quality assessment, and evolution analysis, and applied them on large open-source systems written in Java, C++, or Smalltalk. Our next re-search goals are enriching our metaphor with meaningful representations for relations and encoding higher-level in-formation. 1