Answer-set programming (ASP) is a well-acknowledged paradigm for declarative problem solving, yet comparably little effort has been spent on the investigation of methods to support the development of answer-set programs. In particular, systematic testing of programs, constituting an integral part of conventional software development, has not been discussed for ASP thus far. In this paper, we fill this gap and develop notions enabling the structural testing of answer-set programs, i.e., we address testing based on test cases that are chosen with respect to the internal structure of a given answer-set program. More specifically, we introduce different notions of coverage that measure to what extent a collection of test inputs covers certain important structural components of the program. In particular, we introduce metrics corresponding to path and branch coverage from conventional testing. We also discuss complexity aspects of the considered notions and give strategies how test inputs that yield increasing (up to total) coverage can be automatically generated.
展开▼
