Title:
Static and Dynamic Execute After relationships as a replacement of
dependence-based program slices

Abstract:
We present our recent research on the so-called Execute After
dependencies, which are - in essence - control flow-based dependencies
that do not take into account data-flow information. These dependencies
can be used to compute "conservative program slices" which approximate
slices computed by System Dependence Graph-based algorithms. Due to this
simplification the approach is really scalable compared to the SDG-based
slicers. At the same time, the precision declines only very slightly, by
some 4% on average, while much larger difference would be expected. We
experimented with dependencies on different levels from statement,
through procedure to class level, and with static and dynamic variants
as well. Some extensions to specific architectures, like database
systems and applications in impact analysis and testing are sketched as
well.