Automated Software Conformance Testing

English) Software has found its way into our everyday lives and as a consequence the impact of software failures can range from just being inconvenient to being life-threatening in the worst case. Consequently, software quality is a major concern in the development of modern software systems. Testing is considered as a practical technique for finding bugs in computer programs. An important aspect to be considered is conformance testing, which is used to validate the correctness of an implementation with respect to its model. In such a setting a conformance relation formalizes the notion of correctness. One commonly used conformance relation is the input-output conformance (ioco) relation. This thesis aims to provide theories and techniques that facilitate the application of ioco testing in an industrial context. Originally, the input-output conformance relation was formulated over labeled transition systems and so far some of the underlying assumptions have only been stated on an informal basis. This thesis reflects the input-output conformance relation to reactive processes of the Unifying Theories of Programming framework. Thereby, all assumptions of ioco are formalized. Furthermore, this thesis shows how some of these assumptions can be relaxed. Knowing the pros and cons of a particular testing technique leads to the next step in conformance testing: the selection of an appropriate set of test cases. One promising test case selection technique in the context of ioco testing is the use of formal test objectives, i.e. test purposes. However, nowadays such test purposes are usually developed by hand which makes the quality of the resulting test suite dependent on the skills of the tester. This thesis investigates techniques that support test engineers in complementing manually generated test purposes and by automatically deriving test purposes. Once a set of test cases has been generated and executed a lot of time is spent on the analysis of the test execution results. This thesis describes an approach that intends to assist test engineers in grouping test cases with respect to detected failures. By analyzing only one test case per group the time and effort needed for the test result analysis can be significantly reduced. To sum up, this thesis makes input-output conformance testing more amenable to practical implementation.