Y. Wang and G. Ruhe (Canada)
Software Inspections, Requirements Documents, Empirical Studies, Decision Support, Rough Set Analysis.
Software inspection is an important means to verify and ensure high quality in software development projects. Many proposals have been made to improve Fagan's inspection method ("Fagan's method"). Various publicly published data have already clearly demonstrated the effectiveness and efficiency of software inspection. However, few empirical studies are currently available to guide the appropriate selection of inspectors based on their experience and skills. This paper presents results of an exploratory empirical study to address this open question. The empirical study reveals the cause-effect relationship among certain observable experience, skills of inspectors and the number of defects actually found in the requirement document. An integrated and interactive knowledge discovery approach, INTEGERGQM, was used as the underlying methodology to perform this empirical study. As a part of that, goal-oriented measurement and rough set analysis was applied. Fourteen subjects from a senior computer science course at the University of Calgary participated in the study. The result indicated that software design and testing experience had the strongest influence on the reading effectiveness of software requirement documents. On the other hand, coding experience has the least effect. These results provided further insight for the design of a subsequent study. Rough set based data analysis provided production rules to help decision makers for future projects to better choose the inspectors.
Important Links:
Go Back
