Lay summary
Software product line requirements engineering, that is the discipline of dealing with requirements when using a software product line, has gained significant attention in this decade. However, modeling the variability of a product line and deriving product requirements specifications from the product line requirements are still major challenges, in particular when requirements are represented by graphic models.The SPREBA project aims at enabling and improving variability modeling and product derivation in software product line models which use graphical modeling languages. The project has the following main goals:- Overcome the weaknesses of existing graphic variability models for software product lines,- Solve the configuration problems which occur in today’s variability configuration models (e.g. Feature-Oriented Domain Analysis),- Enable semi-automatic product derivation,- Automatically synthesize the graphic requirements models for a product derived from the product line.Our principal ideas are to model variants and variation points in graphic requirements models with aspects on the one hand and to specify the product configuration options in sophisticated configuration tables on the other hand. Variants frequently cross-cut the commonalities of a system, thus making aspects a good choice for modeling them. Furthermore, aspect weaving techniques can be used for generating product requirements models when products are derived from the product line specification. Sophisticated configuration tables enable semi-automatic product derivation. For example, the effects and consequences of a binding decision can be calculated and checked automatically, thus allowing the exploration of tentative decisions and answering what-if questions.We will investigate these techniques both for loosely coupled modeling languages such as UML and for integrated modeling languages, for example the ADORA language. The focus will lie on integrated modeling languages, because they provide better capabilities for decision evaluation and product model synthesis.