A Pragmatic Evaluation of the Enhanced MC/DC Coverage Criterion for the Testing of Safety-Critical Software

• Main Authors: Jun-Ru Chang, 張君儒
• Other Authors: Chin-Yu Huang
• Format: Others
• Language: en_US
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/41966202395160442641

碩士 === 國立清華大學 === 資訊工程學系 === 94 === With the size of test suites growing, the cost time of software regression testing becomes longer. In practice, finding an effective method to reduce the size of test suites is important. Modified condition/decision coverage (MC/DC) criterion is an effectual verification measurement and can reduce the total number of test cases effectively. To implement MC/DC, our method which focuses on the condition-level uses the data structure of n-cube and Gray Code to satisfy the MC/DC criterion. We also use the property of n-cube graph and the masked input of gates to reduce the searching time. In this thesis, we propose four steps to differentiate the necessary test cases and redundancy test cases. Besides, we also demonstrate a practical regression testing tool and we use JAVA to implement this tool. It can be used to all kinds of testing coverage measurement and can determine which test cases should be re-run. Finally, we will use a simple program to walkthrough the working process of this tool and give a fair comparison with the NASA’s method.