A hybrid approach to feature-oriented programming in XVCL

Abstract

Feature-Oriented Programming (FOP) is a programming paradigm for developing programs by composing features. It is especially useful for software product line development, as each product line member implements some combinations of features. FOP attempts to modularize features and to enable their flexible composition into programs. Recent studies have shown that it is not practical to modularize and then compose features that have fine-grained impact on base programs. In this paper, we present a hybrid approach to feature modularization/composition problem. We modularize only separable features that can be well contained in dedicated files. We handle inseparable features by annotating base programs using preprocessing-like directives. We show how the hybrid approach can be achieved in XVCL, a generative technique to manage variabilities in software product lines. © 2010 Springer-Verlag Berlin Heidelberg.