How to construct a suite of program slices

Abstract

Program slicing is a technique to extract the part of a program (the slice) that influences or is influenced by a set of variables at a given point. Computing minimal slices is undecidable in the general case. Obtaining the minimal slice of a given program is computationally prohibitive even for very small programs. Hence, no matter what program slicer we use, in general, we cannot be sure that our slices are minimal. This is probably the fundamental reason why no benchmark collection of minimal program slices exists, even though this would be of great interest. In this work, we present the first suite of quasi-minimal slices (i.e., we cannot prove that they are minimal, but we provide technological evidences, based on different techniques, that they probably are). We explain the process of constructing the suite, the methodology and tools that were used, and the obtained results. The suite comes with a collection of Erlang benchmarks together with different slicing criteria and the associated quasi-minimal slices. This suite can be used to evaluate and compare program slicers, but it is particularly useful to develop slicers, because it contains scripts that allow for automatically validating a slicer against the whole suite. Concretely, these scripts produce reports about the impact on recall and precision of any change done during the development of the slicer.