On a New Detecting Technique for Conjunctive Literal Insertion Fault in Boolean Expressions

Abstract

Fault-based testing aims to generate test cases that can detect anticipated faults. A recent study has unified the definition of conjunctive literal insertion fault (LIF[.]) based on the abstract syntax tree (AST) representation of a Boolean expression. LIF[.] is considered harder to detect as compared with other faults in Boolean expressions. For irredundant disjunctive normal form (IDNF) Boolean expressions, Multiple Unique True Points (MUTP) strategy guarantees LIF[.] detection. For general form (GF) Boolean expressions, Modified Condition Decision Coverage (MCDC) was used and its fault detecting ability was reported as 70.9%. In this paper we propose a test case generation strategy to detect LIF[.]. Unlike the MUTP strategy, our strategy can be applied to general form Boolean expressions. Empirical results on two renowned software systems' Boolean specifications indicate that, on average, the test suites satisfying our strategy detect approximately 99.0% of LIF[.] for GF expressions and 100.0% of LIF[.] for IDNF expressions. Moreover, the size of the required test suite is approximately half of the MCDC test suite.