An Extended Model of Software Configuration

TL;DR

This paper extends an existing model of software configuration to include feature toggles, aiming to unify terminology, improve research clarity, and facilitate meta-analysis across configuration techniques.

Contribution

The paper introduces MSCv2, an extended model that incorporates feature toggles and qualitative analysis, enhancing the original MSC for better research and practical application.

Findings

Different definitions of software configuration exist among researchers.

Repeated research questions on feature toggles and options due to unclear definitions.

MSCv2 helps clarify and unify research on software configuration.

Abstract

Feature toggles and configuration options are modern programmatic techniques to easily include or exclude functionality in a software product. The research contributions to these two techniques have most often been focused on either one of them separately. However, focusing on the similarities of these two techniques may enable a more fruitful combined family of research on software configuration, a term we use to encompass both techniques. Also, a common terminology may have enabled meta-analysis, a more practical application of the research on the two techniques, and prevented duplication of research effort. The goal of this research study is to aid researchers in conducting a family of research on software configuration by extending an existing model of software configuration that provides terminology for research studies. To achieve our goal, we started with Seigmund et al. Model of Software Configuration (MSC) which was developed based on interviews and publications on configuration options. We explicitly extend the MSC to include feature toggles and to add qualitative analysis of feature toggle-related resources. From our analysis, we proposed MSCv2 as an extended version of MSC and evaluated it through its application on five academic publications and the Chrome system. Our results indicate that multiple researchers studying the same system may provide different definitions of software configuration in their publications. Also, similar research questions may be answered on feature toggles and configuration options repeatedly because of a lack of a clear definition of software configuration. These observations indicate that having a model for defining software configuration may enable more clear and generalized research on the software configuration family of research. Practitioners benefit MSCv2 in their systems to better knowledge transfer to other practitioners and researchers.