Pairwise testing has become a popular approach to software quality assurance because it often provides effective error detection at low cost. However, pairwise (2-way) coverage is not sufficient for assurance of mission-critical software. Combinatorial testing beyond pairwise is rarely used because good algorithms have not been available for complex combinations such as 3- way, 4-way, or more. In addition, significantly more tests are required for combinations beyond pairwise testing, and testers must determine expected results for each set of inputs. This article introduces new tools for automating the production of complete test cases covering up to 6-way combinations.
Crosstalk (Hill AFB): the Journal of Defense Software Engineering
combinatorial testing, formal methods, model checking, pairwise testing, software assurance, software testing