An evaluation of Boolean expression testing techniques

Increase in the size and complexity of the software developed has made software testing a challenging exercise. A number of testing techniques are available but they differ in terms of statement coverage, condition coverage and particularly in fault detection capabilities. The size of the test suite also differs from one technique to other. Fault that has propagated into the system inadvertently, especially into the branch statements, have severe effects as they affect the logic of the program. In this paper, an experimental evaluation of the popular branch-testing techniques (Elmendorf's method, Boolean Operator (BOR), Modified Condition/Decision Coverage (MCDC), and Reinforced Criteria/Decision Coverage (RCDC)) is presented. These techniques are evaluated on the basis of types of faults they identify, size of the test suite and their effectiveness in fault detection. For experiments, various branch statements used and referred in literature are selected. Test cases and mutants were prepared for these branch statements. Mutants were prepared by seeding single operator and operand faults into the statements. The results indicate that for a subset of fault types BOR is effective. A variant of MCDC and RCDC demonstrate better performance on the full class of faults and are only slightly worse than Elmendorf's (CEG) method test suite.