Optimizing problem space representations through domain multi-modeling

This work states that there is a need for an optimized problem space representation for heterogeneous domains. We identify two modeling paradigms widely used in practice: Domain-Specific Modeling (DSM) and Feature-Oriented Domain Modeling (FODM). Each modeling paradigm favors different domain characteristics. Especially the fact that software often is embedded either in a system or in a process and, therefore, is strongly influenced by its environment enforces the demand for a combined representation. We propose a concept for a multi-modeling approach based on existing technology. Multi-modeling means the combination of the two main modeling paradigms to represent a heterogeneous domain. The major benefit of the approach is the reduction of representation complexity by optimizing the representation of single subdomains. This will be shown on one representative case study from the automotive domain. Another advantage is the improved stakeholder communication because of familiar notations. A discussion of limitations shows potential for future work.