Design your analysis: a case study on implementation reusability of data-flow functions

Abstract

The development of efficient data flow analyses is a complicated task. As requirements change and special cases have to be considered, implementations may get hard to maintain, test and reuse. We propose to design these analyses regarding the principle of separation of concerns. Therefore, in this paper we present a reference design for data flow analyses in the context of the IFDS/IDE algorithm. We conducted a case study in order to inspect the level of reuse that can be achieved with our design and found it to be helpful for the efficient development of new analyses.