A human factor analysis for software reliability improvement based on a quality engineering approach in design-review process

Human factors affecting software reliability in the development process should be clarified and controlled to improve software productivity and quality. Several references have proposed for hypothetical models to describe the causal relationships among the human factors and the software faults occurring in development activities. However, there axe few studies which can verify a human factor model through software development experiments. In this paper, we conduct an experiment to clarify human factors and their interactions affecting software reliability by assuming a model of human factors which consist of inhabitors and inducers. In this experiment, we focus on the software design-review process which is more effective than the other processes in the elimination and prevention of software faults. For an analysis of experimental results, a quality engineering approach based on a signal-to-noise ratio is introduced to clarify the relationships among human factors and software reliability measured by the number of seeded faults detected by review activities, and the effectiveness of significant human factors judged by the design of experiment is evaluated. As a result, applying the orthogonal array L-18(2(1) x 3(7)) to the human factor experiment, we obtain the optimal levels for the selected inhabitors and inducers. Moreover, we conduct an additional experiment to approve the experimental results by using the signal-to-noise ratio.