3 ms·
Bookmarked, and I'm excited for part two. Most of these points, I think, generalize to all code review workflows. Some teams, though, put a high priority on in
by wakamoleguy 9y ago
Bookmarked, and I'm excited for part two.
Most of these points, I think, generalize to all code review workflows. Some teams, though, put a high priority on individual ownership and accountability, which creates a few interesting dynamics. First, code reviews were mostly treated as nonblocking. With high trust, we assumed that most code was good by default. From that perspective, it was also discouraged to drop everything to do a code review. You were better off focusing on being productive until you found a natural breaking point. Because the code was often already shipping when it got reviewed, small style changes were typically moot and the focus was higher level.
There's definitely a trade-off with that approach in terms of catching bugs early, and it does depend on a high level of trust in developers. But I didn't feel like the posts assumptions quite covered that.
- watwut 9y agoI work in such environment, it is pretty cool. We have also testers who test things completely independently of developers , so that helps with bugs. Independently means that my biases over what could break and what is important dont affect them.
- KKKKkkkk1 9y agoI have never seen bugs caught in code review. I did see plenty of bugs go through code review just fine. I'm sure bugs might be caught occasionally, but I doubt code reviews are an effective tool for this.
- throwanem 9y agoI've seen it go both ways enough times to know better than thinking code reviews are a silver bullet, but still be glad to have them and still prefer having detailed review be part of the culture over having it not be.
- nine_k 9y agoI have seen and prevented bugs in others' code, and by others in my code. I've seen important misunderstandings unearthed by a few questions during code review, leading to certain design changes. If I never saw bugs during code review, it might be due to my peers producing flawless code, or me not looking hard enough. I see the latter as more probable.