The Significance of Method Parameters and Local Variables as Beacons for Comprehension: An Empirical Study

It is widely accepted that meaningful variable names are important for comprehension. We claim that not all variables are equally significant to the process of code comprehension. We conducted a controlled experiment in which 9 professional developers try to understand 6 methods from production util classes, either with the original variable names or with names replaced by meaningless single letters. In our experiments, we have masked all the methods names. One group of subjects was presented with 3 different versions of the code in the following order: all variable names replaced by single letters, only parameters names or local variables replaced by single letters, and all the original variable names. We then compared the time it took them to resolved the method functionality to a different group who received the method with the original variable names revealed from the beginning. During the experiments we communicated with the subjects and used the thinking aloud technique in order to get a better understanding of each developer’s thinking process. This helped us determine, among other things, which of the variables were more beneficial to their understanding of the methods. Results show that parameter names are more significant for comprehension then local variables, as 80% of the subjects indicated that parameters where more beneficial then locals regardless of the order the variable names were revealed. Replicating previous work on identifier naming, we can also reconfirms that names have a large impact on the comprehension of code. In half of the methods there was a statistically significant difference between the time it took each group to reach an understanding of the method, were it took less time for the subjects who were presented with the method and its original variable names. Surprisingly, the other three of the six methods we used turned out to have problematic names. In fact, they demonstrated that misleading names are worse than meaningless names like consecutive letters of the alphabet, and can lead to errors in comprehension. But these names were not meant to be misleading. This reflects the subjective nature of naming, where a name that one developer thinks is meaningful can be misleading for another developer. Considering the way these methods were chosen, and the fact that they were extracted from real production code, suggests that it is a not uncommon phenomenon. We believe that this highlights the need for additional research on how variable names are interpreted and how better names can be chosen.