|
|
|
|
|
|
by roguecoder
506 days ago
|
|
|
I was at one place where we tracked every bug introduced, and discovered more than 90% were in code written after 5pm. We dramatically cut our bug rate just by shutting down PRs outside of business hours. The problem is that when our performance declines, so does our ability to judge our performance. We can feel more productive while actually doing a much worse job. |
|
|
Intriguing. Did you find that remained true through DST periods, assuming DST observance? Meaning, did you find that it was literally the clock that determined when bugs would seep in, or did bugs also increase if you didn't counteract times changes for whatever human factor (circadian rhythm?) made 5 PM significant?
> The problem is that when our performance declines, so does our ability to judge our performance.
Sure, but what sees performance magically decline at 5 PM?
If it was the clock, did you try removing the clock from the equation? Did bugs show up the same if developers had no idea what time it was?
If it was some other human factor, did you see uniformity across all participants? Were the "night owls" who were just getting started at 5 PM just as likely to introduce bugs after 5 PM as those who had been working since 9 AM?