Introducing TVL, a Text-based Feature Modelling

Feature models are a common way to represent variability in software product line engineering. For this purpose, most authors use a graphical notation based on FODA. The main drawback of those approaches is their lack of scalability: they generally do not fit real-size problems. Indeed, their graphical syntax does not account for attributes or complex constraints and becomes a burden for large feature models. In this paper, we present TVL, a text-based feature modelling notation that is both light and comprehensive, meaning that it covers most constructs of existing languages, including cardinality-based decomposition and feature attributes. The main objective of TVL is to provide engineers with a human-readable language supporting large-scale models through modularisation mechanisms. Furthermore, TVL can serve as an extensible storage format for feature modelling tools. We illustrate the various concepts of the language with short code fragments.

• Type: Conference
• Year: 2010
• Where: VAMOS
• Authors: Quentin Boucher, Andreas Classen, Paul Faber, Patrick Heymans
• Keywords: Feature Modelling, Feature Modelling Notation, Feature Models, Software Engineering, VAMOS 2010