An empirical investigation of software reliability indicators

University of Technology, Sydney. Faculty of Engineering and Information Technology.This thesis investigates how individuals and organisations, without technical skills, might determine the reliability of open source software given the increased use of such software. Software reliability is normally indicated by the growth and subsequent decline of defects in the software. A notable observation is that reliability growth models require a definitive stabilisation phase during which testing can reveal the growth and decline of defects as the indication of the increase in reliability of software. However, there is not necessarily a definitive stabilisation phase in the open source software development. More importantly, the presence or absence of the stabilisation phase is an attribute of a software development method and is not restricted to open source software. When software is developed without a definitive stabilisation phase, reliability growth models are not applicable because the conditions for their validity have not been achieved. Consequently, this thesis looks for alternative information based on tests to aid decision-making about software acquisition. Data was collected by conducting semi-structured interviews from 29 participants who were currently engaged in software development. The information of tests; coverage, sufficiency and rigours of tests concerns the testing that has been performed on the software product and gives expectations on how well the software product has been tested