Variability Modeling in Software Product Line: A Systematic Literature Review

摘要

Variability is the core concept characterizing software product line engineering. Over the past decades, variability modeling has been an emerging topic of extensive research that resulted in different units of variability (e.g., feature, decision, orthogonal, UML) with various variability modeling techniques. Hence, there is a need for a comprehensive study to shed light on the current status, diversity, and direction of the existing variability modeling techniques. The main objective of this research is to characterize the diversity of modeling variability and provide an overview of the status of existing literature. We conducted a systematic review with six formulated research questions and evaluated 74 studies published between the years 2004-2007. The results indicated that the majority of the studies proposed techniques for modeling variability in a separate model rather than modeling variability as an integral part of the development artifact, and the feature model was found as the most common unit of variability. Our study also identified more ambiguity in handling complexity issues as well as the need for a commonly accepted way of addressing variability model evolution. The strength of the evidence in support of the proposed approaches with illustrative examples and lack of robust tooling support that have confined the generalizability of the existing studies need further improvement with more robust empirical studies.