Experimentation in software engineering: an introduction
Experimentation in software engineering: an introduction
| Hi-index | 0.00 |
A considerable number of software projects still exceed time and budget or completely fail, because the qualitative situations of these projects are not visible to the management. The problem can be resolved by monitoring the quality of project results and by steering a project at certain major points (so-called Quality Gates). At each Quality Gate the project results are checked against predefined criteria being derived from carefully chosen metrics. Many software companies use Quality Gates but unfortunately a theoretical reflection on the definition of criteria for Quality Gates is missing. This paper shows, when and how these criteria can be identified and improved over time. Our results obtained from students' software projects show, that the application of a systematic top-down approach (such as GQM) delivers better criteria and that roughly a considerable number of the criteria could be improved after experiences have been captured and reused systematically.