That said, it is still worth questioning their cost/benefit! The modern code review consensus has been promoted by tech companies with enough revenue to hide any costs. The problem is what I think of as the value of code reviews is hard to measure.
I'm pretty sure I've seen multiple papers where people test git and that it's very easy to turn up bugs in git? The limiting factor there seems to be the git developers wanting to spend time on bug fixes, not the difficulty of writing something that will point out a lot of bugs?
-
-
I've seen people make this kind of claim about software, but I've never not been able to find serious bugs, usually in < 1 hour of effort spent writing a fuzzer. I don't think this is surprising since I often find bugs when using software myself and I type slower than a fuzzer.
-
This is probably a better FTF conversation, but I think the only source code I've looked at that has protocols more complex than seen inside processor memory hierarchies are distributed databases. Everything else I've looked at or worked on is simple by comparison.
- 1 more reply
New conversation -
-
-
The paper I think of is about FS badness, where the central insight is that Unix doesn't provide particularly good APIs for data safety (fsyncgate). The testing confirms this, but you hardly need it at that point. (Also, git does a pretty good job at fixing reported errors, IME)
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.