Measuring the efficacy of code clone information: an empirical study

Abstract

Much recent research effort has been devoted to designing efficient code clone detection algorithms and tools. However, there has been little human-based empirical study of how the output of those tools is used by developers when performing maintenance tasks. In this study 43 computer science graduate students completed a bug localization task in which a clone was present while researchers observed their activities. The goal of the study was to understand how those developers use clone information to perform this task. The results of this study showed that participants who used the clone information correctly, i.e. they first identified a defect then used it to look for clones of the defect, were more effective than participants who used the information incorrectly. The results also showed that participants who had industrial experience were more effective than those without industrial experience.