An Observational Theory of Integration Testing for Component-Based Software Development

Integration testing plays a crucial role in componentbased software development. In complementary to the existing works on the selection of test cases and measurement of test adequacy in integration testing, this paper focuses on questions about how to observe the behaviours of a large and complicated system during dynamic testing. We first analyse the structure of whitebox integration testing and propose a family of integration testing methods. We then discuss and formalise the requirements of proper uses of test drivers and component stubs in incremental integration. Finally, we propose a set of axioms for integration testing of concurrent systems.