Development of the Maintainability Index for SPLs Feature Models Using Fuzzy Logic
Davi Cedraz S. de Oliveira, Carla Ilane Moreira Bezerra · 2019
The variability of the common features in an Software Product Line (SPL) can be managed by an feature model, an artifact that consist of a tree-shaped diagram, that describe the features identified in the products and the possible relationships between them. Guarantee the quality of the feature model may be essential to ensure that errors do not propagate across all products. The process of evaluating the quality of a product or artifact can be done using measures, which may reflect the characteristics, sub-characteristics or attributes of quality. However, the isolated values of each measure do not allow access to a whole quality of the feature model, since most of the measures cover several specific aspects that are not correlated. In this context, this paper proposes the aggregation of measures in order to evaluate the maintainability of the feature model in SPL. We aim to investigate how to aggregate these measures and access the respective sub-characteristics by means of a single aggregate value that has the same available information as a set of measures. For this, we have used the theory of Fuzzy Logic as a technique for aggregation of these measures. The new aggregate measure represents the maintainability index of a feature models (MIFM) was obtained. Moreover, to evaluate the MIFM, we applied it to a set of models. It was verified that the aggregate measure obtained allows to measure if a feature models has a high or low maintainability index, supporting the domain engineer in the evaluation of the maintenance of the feature model in a faster and more precise way.