Enhancing Software Visualization with Information Retrieval

I have enhanced Metric Attitude. It is a visualization tool based on static analysis that provides a mental picture by viewing an object-oriented software system by means of polymetric views. In particular, we have integrated an Information Retrieval engine and named this new version of visualization tool as Metric Attitude++. It allows the user to formulate a textual query and to show on the visual representation of the subject software the elements that are more similar to that query. This could be useful in all those cases in which a user needs to identify (or to localize) features implemented in the source code. Several filters are also available to hide possibly irrelevant details and to ease the browsing and then the comprehension of a software system. Finally, we have applied Metric Attitude++ on a number of object-oriented software systems. In this paper, we report preliminary results of a quantitative study on a widely studied open-source software, namely JEdit. On the basis of our results it seems that Metric Attitude++ can be effectively applied to different kinds of source code comprehension tasks and to concept location in source code, in particular.