Problem Frames Construction from Feature Models

The Problem Frames (PF) approach is a well-known approach for describing, analyzing and structuring problems in requirements engineering. It defines certain patterns of problems to be problem frames which have solutions. The real world problems are solved by decomposing to sub-problems which could be matched against problem frames. However, whether existing problem frames is enough for covering all problems is undecided. In this paper, we propose to recognize problem frames in a specific application domain by using feature models. As feature models could link to solution space, our approach bridges the gap between problem frames and solution space. The newly constructed problem frames could be used to analyse problems. We illustrate their usage by presenting a problem frames based problem analysis which include problem descriptions and matches. Because our new problem frames are high level, the real world problems do not need to be decomposed before matched.