Modeling variability in template-based code generators for product line engineering

Abstract

Generating software from abstract models is a prime activity in model-driven engineering. Adaptable and extendable code generators are important to address changing technologies as well as user needs. However, they are less established, as variability is often designed as configuration options of monolithic systems. Thus, code generation is often tied to a fixed set of features, hardly reusable in different contexts, and without means for configuration of variants. In this paper, we present an approach for developing product lines of template-based code generators. This approach applies concepts from feature-oriented programming to make variability explicit and manageable. Moreover, it relies on explicit variability regions (VR) in a code generator's templates, refinements of VRs, and the aggregation of templates and refinements into reusable layers. A concrete product is defined by selecting one or multiple layers. If necessary, additional layers required due to VR refinements are automatically selected.