How to evaluate meta-models for software quality?
The use of appropriate software quality models is crucial for companies to achieve the product quality required to satisfy customer needs. Most current quality models provide little operationalization and lack adaptation guidelines, which limits their usefulness in practice. It has been proposed to use meta-models to specify an explicit structure in order to ensure that quality models conforming to it can be operationalized and adapted by requiring corresponding model elements and modeling constructs. To be applicable in practice, a meta-model needs to be general enough so that existing quality models can be transferred to the new structure provided by the meta-model while preserving the knowledge they contain. This paper presents an empirical approach for evaluating generality as well as its application to a selected meta-model and six industrial quality models. The results show that (1) the proposed meta-model is general enough to model most contents of the industrial quality models, (2) the generality of a meta-model contributes to its perceived ease of use and usefulness, and (3) the evaluation approach is applicable and reflects the perception of quality model experts well.
Applied Software Measurement. Joined International Conferences on Software Measurement, IWSM/MetriKon/Mensura 2010. Proceedings