Key points are not available for this paper at this time.
While software visualization has been widely used for navigation, its use for understanding has been much more limited. Software visualizations are expensive to develop, require large amounts of information that is often difficult to collect, and even then, are good only at addressing the specific task for which they were developed. Our work on software visualization in the Desert environment is aimed at making software visualization a viable approach to understanding. We do this by providing ready access to a variety of information about the system at hand, a range of high-quality, high-density visualizations, and a simple interface that lets the programmer rapidly create new software visualizations for understanding problems as they arise. In this paper we describe the approach we have taken.
Steven P. Reiss (Wed,) studied this question.