Supporting Consistency Checking between Features and Software Product Line Use Scenarios

Abstract

A key aspect for effective variability modeling of Software Product Lines (SPL) is to harmonize the need to achieve separation of concerns with the need to satisfy consistency of requirements and constraints. Techniques for variability modeling such as feature models used together with use scenarios help to achieve separation of stakeholders’ concerns but ensuring their joint consistency is largely unsupported. Therefore, inconsistent assumptions about system’s expected use scenarios and the way in which they vary according to the presence or absence of features reduce the models usefulness and possibly renders invalid SPL systems. In this paper we propose an approach to check consistency –the verification of semantic relationships among the models– between features and use scenarios that realize them. The novelty of this approach is that it is specially tailored for the SPL domain and considers complex composition situations where the customization of use scenarios for specific products depends on the presence or absence of sets of features. We illustrate our approach and supporting tools using variant constructs that specify how the inclusion of sets of variable features (that refer to uncommon requirements between products of a SPL) adapt use scenarios related to other features.