According to the authors, 90 percent of the cost of software is in maintenance, so they decide to make their contribution there. They use a visualization technique based on lines of pixels (for lines of code) and color (for types of change). They also carefully consider the number of dimensions, settling on two. They use a concurrent versions system (CVS) repository for data, figuring that it automatically keeps all changes, and has a diff command of its own. The authors report on two informal case studies, one of a Perl file and one of a C file. They hope to eventually show multi-file visualization.

From a visualization standpoint, this paper is limited, being only for CVS. However, the visualization is quite clear, and there are some attempts made for it to be used elsewhere. People interested in making software evolution easier to understand can benefit from reading and applying the ideas in this paper.