Software reliability is one of the most important characteristics of software quality. An approach to software reliability prediction based on software quality measurements has been introduced in previous research. Our approach is based on the concept of a reliability prediction system. The reliability prediction system pairs subsets of software engineering measures with associated models. Several reliability prediction systems can be constructed but the ones with the highest predictive power rely on software engineering measures that identify real software defects throughout the lifecycle of software development and are paired with structural models of the software application under study, allowing therefore the study of the effects of propagation. Extended finite state machines are an example of such a structural model. This paper describes a tool, called ARPS, which can support the development of such Extended Finite State Machine, the seeding of defects in the structural model and the evaluation of reliability.
展开▼
