A software product line domain requirement model construction method based on model difference and model composition

Software product line is one of the most effective strategies for reuse, and it can improve the development efficiency and quality of software. It is well known that variability modeling, which captures the commonality and variability of software product families, is one of the most important techniques of software product lines. Current software product line variability modeling method, which relies on the domain analyst's analysis ability heavily, are all based on the domain analysis of product families. When software product families are large, it is difficult to establish the variability model. In this paper, a new domain requirement modeling method is presented. In order to achieve separation of concern, feature model is used to facilitate modeling of the different aspect of software product line families. Hence, the domain analysts can thereby concentrate on the modeling of different aspects of the software products. In the end, the domain requirement models are constructed by model differencing and merging algorithm. In this paper, the method for comparing different feature models based on signature and structure is presented. The feature model composition method with constraint is also presented. Experimental are conducted to verify our method and the results showing the efficiency of our domain requirement modelling method.