Investigating styles in variability modeling: Hierarchical vs. constrained styles

Context: A common way to represent product lines is with variability modeling. Yet, there aredifferent ways to extract and organize relevant characteristics of variability. Comprehensibilityof these models and the ease of creating models are important for the efficiency of anyvariability management approach.Objective: The goal of this paper is to investigate the comprehensibility of two common stylesto organize variability into models – hierarchical and constrained – where the dependenciesbetween choices are specified either through the hierarchy of the model or as cross-cuttingconstraints, respectively.Method: We conducted a controlled experiment with a sample of 90 participants who werestudents with prior training in modeling. Each participant was provided with two variabilitymodels specified in Common Variability Language (CVL) and was asked to answer questionsrequiring interpretation of provided models. The models included 9 to 20 nodes and 8 to 19edges and used the main variability elements. After answering the questions, the participantswere asked to create a model based on a textual description.Results: The results indicate that the hierarchical modeling style was easier to comprehendfrom a subjective point of view, but there was also a significant interaction effect with the degreeof dependency in the models, that influenced objective comprehension. With respect to modelcreation, we found that the use of a constrained modeling style resulted in higher correctness ofvariability models.Conclusions: Prior exposure to modeling style and the degree of dependency among elementsin the model determine what modeling style a participant chose when creating the model from natural language descriptions. Participants tended to choose a hierarchical style for modelingsituations with high dependency and a constrained style for situations with low dependency.Furthermore, the degree of dependency also influences the comprehension of the variabilitymodel.