Towards a generic approach for meta-model- and domain- independent model variability

Variability originates from product line engineering and is an important part of today’s software development. However, existing approaches mostly concentrate only on the variability in software product lines, and are usually not universal enough to consider variability in other development activities (e.g., modelling and hardware). Additionally, the complexity of variability in software is generally hard to capture and to handle. We propose a generic model-based solution which can generally handle variability on Ecore-based meta-models. The approach includes a formal description for variability, a way to express the configuration of variants, a compact DSL to describe the semantics of model variability and model-to-model transformations, and an engine which transforms input models into models with injected variability. This work provides a complete and domain-independent solution for variability handling. The applicability of the proposed approach will be validated in two case studies, considering the two independent domains of mobile platforms and architecture knowledge reuse.