Variability Management in Software Product Lines: An Investigation of Contemporary Industrial Challenges

Abstract

Abstract. Variability management is critical for achieving the large scale reuse promised by the software product line paradigm. It has been studied for almost 20 years. We assert that it is important to explore how well the body of knowledge of variability management solves the challenges faced by industrial practitioners, and what are the remaining and (or) emerging challenges. To gain such understanding of the challenges of variability management faced by practitioners, we have conducted an empirical study using focus group as data collection method. The results of the study highlight several technical challenges that are often faced by practitioners in their daily practices. Different from previous studies, the results also reveal and shed light on several non-technical challenges that were almost neglected by existing research.

Handling Variability in Software Architecture: Problems and Implications

“…Lately, there has been research on variability in product lines [4,8]. Moreover, as found in our previous work [19], software architects in practice think in terms of development activities that they perform when describing their needs for new methods or tools.…”

“…Therefore, our findings that target architecting areas where variability has a particular impact could thus help improve methods and tools. As argued in [8], tool support for variability in product lines to model variability has been proposed. On the other hand, complexity, knowledge harvest and management, design decisions and enforcement are not addressed by approaches.…”

“…So far, variability has primarily been studied in the software product line (SPL) domain [4,8]. In SPL, variability is understood as the ability of a software artifact to be configured, customized, extended, or changed for a specific context, in a preplanned manner [1].…”

“…Chen and Babar identified a broad range of technical and non-technical challenges related to variability in SPL [8].…”

“…Our work complements [8] by a) focusing on software architecting of systems beyond SPL, and b) investigating concrete tasks that architects perform rather than any task in the software life cycle.…”