Supporting the adaptation of software quality models - an empirical investigation

Measuring and evaluating software quality are fundamental challenges for software engineers. Assessing software quality is hard to accomplish in practice not only because existing quality models are not easily applicable, but also because adjusting them to the needs of one's own organization and projects requires intensive effort. In this paper, we present a study on applying a flexible but rigorous adaptation process for quality models. The goal-oriented adaptation process is based on the existence of a quality meta-model that provides a structure for adapted models. In order to obtain first empirical insights, we compare adaptations guided by the tool-supported adaptation process with (ad-hoc) adaptations using a tool that can be used to create and edit quality models. In the study, we investigated the formal consistency, appropriateness, and efficiency of exemplary quality model adaptations. One important study result is that the quality models obtained applying thetool-supported process are considerably more consistently and appropriately adapted than the ones obtained by following an ad-hoc approach. Further, we could observe that model adaptation is significantly more efficient when applying the adaptation process.