An ontological rule-based approach for analyzing dead and false optional features in feature models

Feature models are a common way to represent variability requirements of software product lines by expressing the set of feature combinations that software products can have. Assuring quality of feature models is thus of paramount importance for assuring quality in software product line engineering. However, feature models can have several types of defects that disminish benefits of software product line engineering.Two of such defects are dead features and false optional features. Several state-of-the-art techniques identify these defects, but only few of them tackle the problem of identifying their causes. Besides, the explanations they provide are cumbersome and hard to understand by humans. In this paper, we propose an ontological rule-based approach to: (a) identify dead and false optional features; (b)identify certain causes of these defects; and (c) explain these causes in natural language helping modelers to correct found defects. We represent our approach with a feature model taken from literature. A preliminary empirical evaluation of our approach over 31 FMs shows that our proposal is effective, accurate and scalable to 150 features.