FAMA: Automated analyses of software product lines. A CSP-based reasoning framework
David Benavides and Pablo Trinidad ( University of Seville )
- 15:00 19th June 2008 ( week 9, Trinity Term 2008 )Common Room
In recent years, software product lines have been introduced as a new
technical paradigm to build software products focusing on the
development of a set of distinctive products sharing a common part
rather than building different products one by one from scratch. There
is an important piece of this paradigm that serves as a model to
represent the set of products of the software product line. To this end,
one of the most widely used models is the so�called feature model that
was proposed back in 1990 and has been a topic of research throughout
recent years. Feature models have been mostly used as graphical
notations to communicate between different stakeholders. The automated
analysis of feature models was stated as a research challenge in the
original proposal, however only in recent years some publications have
paid attention to this task.
In this presentation, we introduce FAMA (FeAture Model Analyser), a new framework to automate the analysis of software product lines in general and feature models in particular. Its main advantages lie in its formal semantics which avoid misinterpretation, its abstraction that allows to extend the framework with other models than feature models, the capability of analysing extended feature models where feature attributes are included, and finally the support of multiple solvers, such us CSP, SAT and BDD-based solvers, in the implementation of feature model analysis. Putting all this together we can set the basis to develop automated tool support.
technical paradigm to build software products focusing on the
development of a set of distinctive products sharing a common part
rather than building different products one by one from scratch. There
is an important piece of this paradigm that serves as a model to
represent the set of products of the software product line. To this end,
one of the most widely used models is the so�called feature model that
was proposed back in 1990 and has been a topic of research throughout
recent years. Feature models have been mostly used as graphical
notations to communicate between different stakeholders. The automated
analysis of feature models was stated as a research challenge in the
original proposal, however only in recent years some publications have
paid attention to this task.
In this presentation, we introduce FAMA (FeAture Model Analyser), a new framework to automate the analysis of software product lines in general and feature models in particular. Its main advantages lie in its formal semantics which avoid misinterpretation, its abstraction that allows to extend the framework with other models than feature models, the capability of analysing extended feature models where feature attributes are included, and finally the support of multiple solvers, such us CSP, SAT and BDD-based solvers, in the implementation of feature model analysis. Putting all this together we can set the basis to develop automated tool support.