[explained] the mind is a powerful place: how showing code comprehensibility metrics influences code understanding
Wyrich, M., Preikschat, A., Graziotin, D., & Wagner, S. (2020). The Mind Is a Powerful Place: How Showing Code Comprehensibility Metrics Influences Code Understanding. In Proceedings from 43rd International Conference of Software Engineering (ICSE 2021, pp. 512-523) DOI:10.1109/ICSE43902.2021.00055 [open access][open data].
Research has developed a plethora of metrics that either measure development or gather data from users after deployment. A trivial example would be to measure how many lines of code are written in one hour, as a way to measure developers’ productivity. Research, however, has oftentimes ignored the validation phase of metrics. This misalignment poses a danger to the digitalization industry: bad choices can be derived from metrics that do not measure what they are supposed to measure.
Following designs from psychology and cognitive science, we build an experiment that empirically demonstrates that a non-validated metric on source code understandability (a common way to assess software quality), when shown to software developers (as is usually the case with tools) will influence their subjective understanding of the code and its difficulty. The effect is large and goes in both directions (easy or hard).
The results have potential for high disruption in research as well as in the digitalization industry. Our paper will force a re-discussion of software metrics, what they are, their nature, and what they actually measure.
I kept this post short and sweet because Marvin Wyrich has recorded and edited a great video on the paper. I highly recommend watching it. Click on the thumbnail to watch it on YouTube1.
Consider watching his other videos as well as subscribing to his channel.
I do not use a commenting system anymore, but I would be glad to read your feedback. Feel free to contact me.