Tate, Jonathan (2003) Software process quality models: a comparative evaluation. Masters thesis, Durham University.
| PDF 8Mb | |
Archive (ZIP) (One CD) 2185Kb |
Abstract
Numerous software processes are implemented by software organisations in the production and maintenance of software products. Varying levels of success are observed in their execution, as processes vary in content and quality. A number of quality models for software processes have been published, each of which is intended to encompass the totality of quality factors and issues relevant to a specific notion of process quality. These quality models may be used to develop a new process, measure the quality of existing processes, or guide improvement of existing processes. It is therefore desirable that mechanisms exist to select the model of highest intrinsic quality and greatest relevance. In this thesis, mechanisms are proposed for the comparative evaluation of software process quality models. Case studies are performed in which existing software process quality models are applied to existing software processes. Case study results are used in empirical evaluation of models to augment theoretical evaluation results. Specific recommendations are made for selection of models against typical selection criteria. Assessment is performed of the assessment procedures against defined success criteria. Theoretical evaluation procedures are developed to measure process quality models against defined quality criteria. Measurements are performed of conformance of models to the requirements set for an ideal process quality model, and the relevance of model content to defined stakeholders in software processes. Comparison is also made of the scope and size of models. Empirical evaluation procedures are developed to assess model performance in the context of application to real software processes. These procedures assess the extent to which the results of process measurement using process quality models are observed to differ, and hence the importance of selecting one model in preference to others. Measurement is also performed of the extent of difference in the software processes evaluated in the case studies.
Item Type: | Thesis (Masters) |
---|---|
Award: | Master of Science |
Thesis Date: | 2003 |
Copyright: | Copyright of this thesis is held by the author |
Deposited On: | 09 Sep 2011 10:02 |