Visualization of feature-traceability in variant-rich systems

Abstract

Purpose. Feature-traceability is defined as the ability to trace features and the different aspects of it in code. Variant-rich systems are systems that have many variants, such as software product lines. Feature-traceability in variant-rich systems is hard, as a result, the purpose of this thesis is to suggest views that will improve feature-traceability and other feature related problems in variant-rich systems. Methods. The purpose was achieved by performing a literature survey, then, by following a design science methodology, views for improved feature-traceability was conceived. These conceived views were successively evaluated and improved upon further. Result. The result of this thesis is an analysis of the current state of featuretraceability in research, concept views that present ideas of what could be implemented in a feature-traceability tool. Additionally, an open-source plugin for the integrated development environment Eclipse called Feature Dashboard. Feature Dashboard enables the user to, not only, trace features to their implementation, but to get information such as which features are implemented in the same file, observe where in the folder structure a feature is located, metrics about features, feature metrics for folders/files, and inspect which features are shared between projects. Conclusion. This thesis shows that Feature Dashboard can be used to improve feature-traceability, maintainability, understandability, and other feature management related tasks for a project. As Feature Dashboard is open source, it offers a platform that enables future researchers to, either implement new views or to compare the result of Feature Dashboard to other views. Furthermore, anyone can suggest improvements and extend it with additional functionality, thus benefiting everyone.