Modelling Variability in Self-Adaptive systems: Towards a Research Agenda

Abstract

The combination of generative programming and com-ponent engineering applied to software product line en-gineering (SPLE) has focused thus far mostly on static systems (as previous editions of AOPLE indicate), with variability that is bound once. Meanwhile, an emergent paradigm in software engineering deals with self-adaptive and dynamic systems. While there is a well-known and agreed SPLE process for static systems, there has been less focus on dynamically adaptive systems. As such it appears imperative to include it in an extended research agenda. In the present paper we observe limitations related to domain engineering in SPLE and identify what fundamen-tal concepts, such as context and binding time, must be re-tought in order to achieve SPLE for dynamically adaptive systems. The main contribution of this paper is a set of research questions, aimed at defining a common research agenda for addressing these limitations. 1