Several authors have found evidence of large dependence clusters in the source code of a diverse range of systems, domains, and programming languages. This raises the question of how we might efficiently locate the fragments of code that give rise to
