Improving Software Quality through Understanding and Early Estimations

Abstract: Software quality is an important factor of a large and complex software system. Hence, there is a need for measuring and controlling quality. Software quality has, however, many different f actors and involves implicit expectations of the behaviour of the software. This thesis investigates two aspects of improving the software quality, in particular the reliability factor. The first part of this thesis investigates how software reliability can be assessed and planned for during the planning of test. It also describes how the reliability can be estimated during development. By estimating or measuring the reliability throughout the development, it can be used to control the software development process and resources can be directed to where they are best needed. The software quality is improved by better use of the resources. The second part investigates software quality improvement through understanding of the software process. The quality of the product is partly a result of the quality of the process and by studying the effects on the software quality when process improvements are introduced the software quality can be improved. In this part, defect injection and removal are also studied. If the introduction and removal of defects can be understood, software quality and in particular the software reliability can be improved. The results presented in this thesis provide support for controlling and improving the software quality. The software reliability can be assessed in the planning of test and it can then be controlled throughout the development. The thesis also provides an understanding of the defect injection and removal, which helps improving the software quality by decreasing the number of injected defects and a more effective defect removal.