To ensure software is performing as intended, it can be black-box or white-box tested. Category partition is a black-box, specification-based testing technique that begins by identifying the parameters, categories (characteristics of parameters), and choices (acceptable values for categories). These choices are then combined to form test frames on the basis of various criteria such as Base Choice and Each Choice. To ensure that the combinations of choices are feasible, constraints on choices are introduced. Combining choices, while accounting for constraints, to form an each choice adequate test set is feasible (eg, using constrained covering arrays from combinatorial testing). However, the Base Choice criterion has not been defined to specifically account for constraints on choices, resulting in adverse consequences. In this paper, we introduce two extensions to the Base Choice criterion, namely, Constrained Base Choice and Extended Constrained Base Choice to specifically account for (complex) constraints on choices. We use a number of academic and industrial case studies to compare different adequacy criteria, including the new ones, in terms of cost and effectiveness at finding faults. Results show the performance of the new criteria equivalent to a 3-way combination criterion with a much smaller cost.

Additional Metadata
Keywords Base choice, Category partition, Combination criteria, Combinatorial testing, Constraints, Single annotations
Persistent URL dx.doi.org/10.1002/smr.1868
Journal Journal of Software: Evolution and Process
Citation
Khalsa, S.K. (Sunint Kaur), & Labiche, Y. (2017). Extending Category Partition's Base Choice criterion to better support constraints. Journal of Software: Evolution and Process. doi:10.1002/smr.1868