A Taxonomy System to Identify Human Error Causes for Software Defects

Abstract

The identification of root causes is critical for software organizations to prevent defects and improve software quality. Human error is thought to be a leading cause of software defects inducements. However, no integrated classification system exists targeting to identify human errors of software engineers. The present study aims to develop a valid taxonomy and integrated into corresponding data collecting system for China Aviation Software Industry (CASI). The taxonomy is developed by a structural analysis based on cognition models. It is evaluated and evolved based on the feedback from the potential users by interview. Then it is put into trial application and assessed in 9 companies. The results show that the taxonomy system of software developers (TSSD) appears to be a reliable and useful system for studying human error in CASI. It has potential application value in defect causal analysis (DCA) for the generalized software industry.