Investigação de operadores essenciais de mutação para programas orientados a aspectos

Abstract

Context: The literature on software testing reports on the application of the Mutation Analysis criterion or mutation testing as a promising approach for revealing faults in aspect-oriented (AO) programs. However, it is widely known that this criterion is highly costly due to the large number of generated mutants and the effort required to identify equivalent mutants. We highlight that little existing research on mutation testing for AO programs focuses on cost reduction strategies. Objective: this work aims at investigating the cost reduction of mutation testing for AO programs. In particular, we intend to reduce the cost of mutation testing by identifying a reduced set of mutation operators that are capable of keeping the effectiveness in guaranteeing the quality of the designed test sets. Method: to achieve the goals, we applied an approach called Sufficient Procedure. Such approach yields sufficient (sets of) mutation operators. Test sets that are adequate with respect to mutants produced by sufficient operators are able to reveal the majority of faults simulated by a whole set of mutants. Results: by applying the Sufficient Procedure, we obtained substantial cost reductions for three groups of AO programs. The cost reduction in the experiments range from 52% to 62%. The final mutation scores yielded by the test sets that are adequate to mutants produced by the sufficient operators range from 92% to 94%. Conclusion: with the achieved results, we conclude that it is possible to reduce the cost of mutation testing applied to AO programs without significant losses with respect to the capacity of revealing prespecified fault types. The Sufficient Procedure has shown to be able to support cost reduction and to maintain the effectiveness of the criterion.