Software Product Lines (SPLs), are families of software systems sharing common assets and exhibiting variabilities specific to each product member of the family. Commonalities and variabilities are often represented as features organised in a feature model. due to combinatorial explosion of the number of products induced by possible features combinations, exhaustive testing of SPLs is intractable. Therefore, sampling and prioritisation techniques have been proposed to generate fault-finding, sorted lists of products based on coverage criteria or weights assigned to features. Solely based on the feature model, these technique do not take into account behavioural usage of such products as a source of pri-oritisation. In this paper we assess the feasibility of integrating usage models into the testing process to derive statistical testing approaches for SPLs. Usage models are given as a Markov chains enabling the selection of probable/rare behaviours that can then be analysed against Featured Transition Systems (FTSs), acting as design models of the SPLs, to determine which features and products are realizing these behaviours. Statistical prioritisation can achieve a significant reduction of the state space, and modelling efforts can be rewarded by easing tool integration. In particular we used MaTeLo, a statistical test cases generation suite developed at ALL4TEC. Our experience report is based on the evaluation of our feasibil-X. Devroey, G. Perrouin (FNRS Postdoctoral Researcher), M. Cordy (FNRS Research Fellow), 2 Xavier Devroey et al. ity criteria on two different systems: Claroline, a configurable course management system, and Sferion TM , dealing with an embedded helicopter landing function.
Devroey, Xavier ; Perrouin, Gilles ; Cordy, Maxime ; Samih, Hamza ; Legay, Axel ; et. al. Statistical Prioritization for Software Product Line Testing: an Experience Report. In: Software & Systems Modeling,