An empirical study on change recommendation
Manishankar Mondal, Chanchal Kumar Roy, Kevin A. Schneider · Computer Science and Software Engineering · 2015
Recommending changes to programmers by exploiting their repetition tendencies during system evolution has been investigated by a number of studies. In our research we perform a change type (additions, deletions, and modifications) based analysis of the efficiency of change recommendation. We also investigate the programmer sensitivity of the repeated changes (i.e., the extent the same changes are repeated by the same programmers) of different change types. The existing studies did not perform such investigations. However, these investigations can be important for efficient ranking (i.e., prioritizing) and filtering of recommendations. According to our investigation on thousands of commits of five diverse subject systems we observe that modifications have a very low tendency (around 1.3%) of being repeated. We should primarily focus on recommending additions, and deletions. More importantly, overall 71% of the repeated changes are programmer sensitive. We believe that a change recommendation system that prioritizes recommendations considering programmer sensitivity can help programmers reuse previous changes in a time-efficient manner.