Evaluating the Effectiveness of Random and Partition Testing by Delivered Reliability

The software engineering literature is full of test data selection and adequacy strategies. However, It is still a question whether these adequacy strategies are effective or not. So it is necessary to research how to evaluate the effectiveness of test strategy. The effectiveness of random and subdomain testing methodology is normally evaluated by failure detecting ability. However, detecting more failures does not guarantee that the software is more reliable because those failures detected maybe small and subtle ones that will seldom occur in reality. So in this paper, delivered reliability which presents the reliability of software after testing is introduced to evaluate their effectiveness. The better method delivers higher reliability after all test failures have been eliminated.