The Dominance Tree in Visualizing Software Dependencies

Abstract

Dominance analysis from graph theory allows one to locate subordinated software elements in a rooted dependency graph. It identifies the nesting structure for a dependency graph as a dominance tree, and, hence, adds information not immediately visible in large and complex graphs. Moreover, the subordination (or locality) can be leveraged for drawing dependency graphs. This paper envisions ways to leverage the dominance relation for structuring and presenting large dependency graphs. To explore the feasibility of these hinds of visualization, we measure dominance trees for large software systems written in different programming languages. These measurements give us the necessary information to design a usable visualization