A software requirements specification(SRS)is a detailed description of a software system to be developed.This paper proposes and evaluates a lightweight review approach called value-oriented review(VOR)to detect defec...A software requirements specification(SRS)is a detailed description of a software system to be developed.This paper proposes and evaluates a lightweight review approach called value-oriented review(VOR)to detect defects in SRS.This approach comprises setting core values based on SRS and detecting the defects disturbing the core values.To evaluate the effectiveness of the proposed approach,we conducted a controlled experiment to investigate whether reviewers could identify and record the core values based on SRS and find defects disturbing the core values.Results of the evaluation with 56 software engineers showed that 91%of the reviewers identified appropriate core values and 82%of the reviewers detected defects based on the identified core values.Furthermore,the average number of defects detected using the proposed approach was slightly smaller than that detected using perspective-based reading(PBR);however,PBR requires defining review scenarios before attempting to detect any defects.The results also demonstrated that the proposed approach helped reviewers detect the omission defects,which are more difficult to detect from SRS than defects because of ambiguity or incorrect requirements.展开更多
基金This study was supported by the 2020 TSUBAME project of Tokyo Institute of Technology(Grant No.20D10597).
文摘A software requirements specification(SRS)is a detailed description of a software system to be developed.This paper proposes and evaluates a lightweight review approach called value-oriented review(VOR)to detect defects in SRS.This approach comprises setting core values based on SRS and detecting the defects disturbing the core values.To evaluate the effectiveness of the proposed approach,we conducted a controlled experiment to investigate whether reviewers could identify and record the core values based on SRS and find defects disturbing the core values.Results of the evaluation with 56 software engineers showed that 91%of the reviewers identified appropriate core values and 82%of the reviewers detected defects based on the identified core values.Furthermore,the average number of defects detected using the proposed approach was slightly smaller than that detected using perspective-based reading(PBR);however,PBR requires defining review scenarios before attempting to detect any defects.The results also demonstrated that the proposed approach helped reviewers detect the omission defects,which are more difficult to detect from SRS than defects because of ambiguity or incorrect requirements.
