However, the better-quality stuff is often great, and upvotes are a great way to discover such stuff.
Given that only one third of my answers at stackoverflow have more than two upvotes, and one fifth has zero, by that logic my answers are utter shit, not worth reading.
According to the people that I've helped (I get about one email per week about various StackOverflow/LinuxQuestions stuff I've discussed, even though I haven't answered any questions or comments in a year), like
this or
this or
this, I do not believe so. I don't claim they are great; I only claim that their score does not match their usefulness.
My highest voted answer starts with
"Why does everyone insist on reinventing the wheel again and again?", and isn't the accepted answer
.
(It's about using POSIX.1
nftw() to walk directory trees, instead of opendir()/readdir()/closedir(), as the latter is hard to implement in a robust way, correctly handling e.g. files that are added, deleted, renamed, or moved during the walk. If you care more about BSD than POSIX compatibility, use
fts() instead; me, I like POSIX.)
It would be much better if instead of "reputation" it was called "popularity", because that is what it reflects: popularity, not quality or reputation.
I've only ever asked
one question at the StackExchange network, related to perfect binary search trees; otherwise I've just trawled for unpopular but interesting questions, and answering those. So, perhaps the set of questions and answers I've dealt with is particularly skewed. It is definitely true that the "atmosphere" differs a lot on different StackExchange sites, like math.stackexchange.com vs stackoverflow.com.