Using auxiliary artifacts during code inspection activity: findings from an exploratory study

Code inspection is an important activity to identify defects in the source code and improve software quality. However, even when using techniques such as checklists, inspectors consider implicit decision-making knowledge. In this paper, we perform an exploratory study with groups of inspectors (Group1 and Group2) with two objectives: 1) to present findings on how (and if) auxiliary artifacts interfere in decision making during the code inspection activity, and 2) to show whether there is any influence on the number of defects identified by inspectors when using or not auxiliary artifacts. Both groups used the computational code inspection support of the CRISTA tool, but only Group1 used auxiliary artifacts (requirements, UML diagrams, software metrics). We identified 10 findings. All of them are related to the inspectors' decision making and the influence of using artifacts on defects identification. The findings suggested that when inspectors use auxiliary artifacts, their effectiveness in identifying defects is improved. Besides, their decision making is more homogeneous than that of inspectors who do not use auxiliary artifacts. However, more investigations are necessary to make the results more generalizable. As future work, different strategies for code inspection techniques can be defined based on the findings.