Abstract
In large-scale software development, more than a half of the total development effort in our company is expended on the testing phase. Therefore, information for management of the fault-detection and correction activities in the testing phase, and for assurance of the quality / reliability at product delivery is very important in order to develop a highly reliable software product efficiently and economically. However, in the actual testing phase, an expert person with special knowledge and experience is required in order to derive the information mentioned above. This paper discusses the methodologies, e.g., derivation of quality and reliability assessment information, based on fault data with high accuracy to software product during testing without needing an expert person. Further, we discuss the effect and usefulness acquired by using these methodologies in the actual projects.
