Quality measurement and the utilisation of measurement results in a software development process
No Thumbnail Available
Doctoral thesis (monograph)
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
Report / Helsinki University of Technology, Networking Laboratory, 2/2003
AbstractThe research topic of the thesis is quality measurement and the utilisation of measurement results in a software development process of a large switching system. The first problem studied, the utilisation, means that measurement results are poorly analysed; data is collected, but not enough conclusions are drawn about how that experience data could help in the planning and execution of software development projects. So, the problem is how to learn more about the collected material. The second research problem is further development of measurements from the point of view of the utilisation. We think this is because the existing metrics have too much focused on later phases instead of using purposeful driving metrics for controlling the process towards high final quality in early phases. The main objectives of this study are to examine the coupling of existing quality metrics to design practices, to generate new quality indicators, and to develop the utilisation of measurement results further. Basically, the study aims at finding out relationships between post-release measurement results and the driving attributes. It also aims at making proposals for methods to improve and introduce new measures; and to create a utilisation model for better use of measurement results. Research methods involve both empirical research and constructive design of models for measurement processes and utilisation. We apply statistical methods to sixteen data sets gathered from real software modules; and discuss correlation in the light of nineteen case studies. Eleven stated hypotheses are tested by their evidence based on the statistical analyses of the material collected during the 1990's from industrial projects. Generation of process and utilisation models is based on constructive modelling, on experience of good practices, and on adoption of elements from modern measurement theory as well as from goal- and attribute-oriented approaches. First, the results contribute to enhancement of empirical knowledge regarding implementation, problems, obstacles as well as lessons learnt on utilisation of measurements in an industrial environment. As a second result, the study provides analysed test results for each of the eleven hypotheses. A new result worth mentioning is a conclusion that 70-80 % of individual modules, which have been faultless in Function Test, are also faultless in System Test and during their first 6 months after delivery. Furthermore, the thesis presents an improved measurement framework including a renewed measurement process. The process is composed of four parts for planning, definition, performing - and especially - for utilising measurements. The thesis has also resulted in a novel utilisation model that can be used to improve the use of measurement results. Finally, we present a supplementary result that requires further research and validation. In particular, we have constructed a six-level measurement utilisation capability model.
telecommunication switching, software quality, measurements, metrics, capability, maturity
- Additional errata file available.