Orientation de l'effort des tests unitaires dans les systèmes orientés objet : une approche basée sur les métriques logicielles
|Advisor:||Lamontagne, Luc D.; Badri, Mourad|
|Abstract:||Current software systems are large, complex and critical. The need for quality requires a lot of tests that consume a large amount of resources during the development and the maintenance of systems. Different techniques are used to reduce the costs of testing activities. Our work is in this context. It aims to guide the unit testing effort distribution on the riskiest software components using the source code attributes. We conducted several empirical analyses on different large object-oriented open source software systems. We identified and studied several metrics that characterize the unit testing effort according to different perspectives. We also studied their relationships with the software class metrics including quality indicators. The quality indicators are a synthetic metric that we introduced in our previous work. It captures control flow and different software attributes. We explored different approaches for unit testing effort orientation using source code attributes and machine learning algorithms. By grouping software metrics, we proposed an effort orientation approach based on software class risk analysis. In addition to the significant relationships between testing metrics and source code attributes, the results we obtained suggest the possibility of using source code metrics for unit testing effort orientation.|
|Document Type:||Thèse de doctorat|
|Open Access Date:||24 April 2018|
|Collection:||Thèses et mémoires|
All documents in CorpusUL are protected by Copyright Act of Canada.