An Empirical Evaluation and Analysis of the Fault-Detection Capability of MUMCUT for General Boolean Expressions * +
Tsong Yueh Chen, Kwan Yong Sim, Chang‐ai Sun, Tsun-Him Tse · Swinburne figshare (Swinburne University of Technology) · 2004
Boolean expressions are extensively used in software specifications. It is important to generate a small-sized test set for Boolean expressions without sacrificing the fault-detection capability. MUMCUT is an efficient test case generation strategy for Boolean expressions in Irreducible Disjointed Normal Form (IDNF). In the real world, however, Boolean expressions written by a software designer or programmer are not normally in IDNF. In this paper, we apply MUMCUT to generate test cases for general Boolean expressions and develop a mutation-based empirical evaluation on the effectiveness of this application. The experimental data show that MUMCUT can still detect single seeded faults in up to 98.20% of general Boolean expressions. We also analyze patterns where test cases generated by MUMCUT cannot detect the seeded faults.