Automated Evolutionary Test Data Generation with Domain Reduction for Aspect-Oriented Programs

Mark Harman,Tao Xie

2007-01-01

Abstract:Aspect-Oriented Programming is an emerging technique that helps improve separation of concerns in software systems. It has received a great deal of recent interest. However, algorithms and empirical results for testing of aspect-oriented programs are lagging some way behind this upsurge in interest. To date, there are few published approaches to automated test data generation for aspectoriented programs and, therefore, there remain few empirical results on testing aspect-oriented programs. In this paper, we present a new approach to automated test data generation for aspect-oriented programs and its supporting system, providing empirical evidence to validate its applicability. The new approach uses domain reduction to further improve the performance of test data generation, providing empirical evidence that domain reduction can significantly reduce the computational expense of test data generation for aspectoriented programs. The paper also presents the results of a study into effort reduction when focusing on testing behavior in aspects instead of all behavior in the whole program.