Partial Automation of Sensitivity Analysis by Mutant Schemata Approach
Department of Computer Science, Sultan Qaboos University, Oman
Abstract: According to Voas et al. testability is defined as the ease with in which faults may manifest themselves as failures when the software undergoes the testing process . They also went further by introducing an approach for measuring sensitivity in terms of estimates from Propagation, Infection, and Execution (PIE) analyses of software and calculating the testability of a program through sensitivity estimate. Their testability calculations ‘by hand’ to determine the stability of the PIE analysis technique had drawbacks such as more time-consuming, high cost and less precision in the overall results . Particularly the infection analysis part is one of the most expensive, sophisticated and time-consuming component of the PIE analysis technique. In order to solve this problem an investigation has been carried out by the author for establishing the feasibility of automating (or partially automating) the PIE analysis technique by means of a fast, and computationally less expensive Mutant Schemata Generation (MSG) approach . An MSG-Infection tool has been developed to automate PIE analyses partially. This paper explains the use of MSG-Infection tool in automating the sensitivity analysis of C-programs and presents the results demonstrating the performance improvements achieved due to the MSG-Approach.
Keywords: Testability analysis, PIE technique, mutation testing, mutant schemata approach, Mothra mutation system, PiSCES, MSG-Infection tool.
Received June 20, 2004; accepted February 2, 2005