Towards Consistent Evolution of Feature Models

摘要

This paper explores the possibility of consistent evolution of feature models (FMs), which should resolve the requested changes and maintain the consistency of FMs. According to the definition of FMs, we first analyze the primitive elements of FMs and suggest a set of atomic operations on FMs. Then we analyze and apply the semantics of change to FMs to support consistency maintenance during FMs evolution. The resolution of a requested change to an FM requires obtaining and executing a sequence of additional changes derived from the requested change for keeping the consistency of the FM. Our approach limits the consistency maintenance of an FM in a local range affected only by the requested change instead of the whole FM, which reduces the effort and improves the efficiency for the evolution and maintenance of FMs.