I'll have to jump out and point out that this is exactly the type of study that I feel the academia nowadays more or less are forced to perform, but at the same

I'll have to jump out and point out that this is exactly the type of study that I feel the academia nowadays more or less are forced to perform, but at the same time have questionable setup and limited impact. Overall the study is performed a very limited time frame, with a last portion of the data analysis performed through emailing back and forth. The research scope is limited to 300+ organization, which in fact has 30000+ engineer in total. And the tool was developed without much of normal industry standard understanding, aka, introduce a code analysis tool requiring users explicitly invoking it (I think anyone who is at a typical senior level in large Corp knows that without management pushing, such change is never going to work). The mentors of the author might be full time researcher? I am not really nitpicking or condescending. I am dismayed by the waste of time and effort on both sides. I'll read the paper later. But I suspect it would be basically some well known facts in the dunstry...