Grammar-based test generation for software product line feature models

Product lines are often employed for the facilitation of software re-use, rapid application development and increase in productivity. Despite the numerous advantages of software product lines, the task of testing them is a cumbersome process due to the fact that the number of applications that need to be tested is exponential to the number of features represented in the product line. In this paper, we attempt to reduce the number of required tests for testing a software product line while at the same time preserving an acceptable fault coverage. For this purpose, we introduce eight coverage criteria based on the transformation of software product line feature models into formal context-free grammars. The theoretical foundation for the proposed coverage criteria is based on the development of equivalence partitions on the software product line configuration space and the use of boundary value analysis for test suite generation. We have performed experiments on several SPLOT feature models, the results of which show that the test suite generation strategies based on the proposed coverage criteria are effective in significantly reducing the number of required tests and at the same time maintaining a high fault coverage ratio.