Streamlining Feature-Oriented Designs

Software development for embedded systems gains momentum but faces many challenges. Especially the constraints of deeply embedded systems, i.e., extreme resource and performance constraints, seem to prohibit the successful application of modern and approved programming and modularization techniques. In this paper we indicate that this objection is not necessarily justified. We propose to use refinement chain optimization to tailor and streamline feature-oriented designs to satisfy the resource constraints of (deeply) embedded systems. By means of a quantitative analysis of a case study we show that our proposal leads to a performance and footprint improvement significant for (deeply) embedded systems.