Now it depends on what you think about what Stack Overflow is supposed to be. I personally think Stack Overflow should be focused on providing information that not already exists, I don't think it should try to replicate or link to every manual and tutorial out there. This stuff already exists and is hopefully actively maintained by the language and library developers, so what is the point of duplicating it? I think Stack Overflow should not be about teaching people how to code but help them to solve hard problems.
What constitutes a hard problem of course heavily depends on one's experience and abilities so the line is kind of arbitrary but I would say that if it is in a beginners tutorial or the manual, then it is not a hard problem. I certainly unterstand the desire to get problems solved within minutes but I also think that is not really a good thing. No matter how good and experienced a developer you are, you will come across problems that take hours and days to solve again and again. Reading ten pages of a manual should really not be considered an annoyance but a valuable skill.
Very often you will already have a rough idea of how you want to solve your problem and just look for some very specific information, but if you take the time to read about it you will often find that your initial solution is actually not that great and there is a different preferred way to do it. So if there is one advice I would give to beginners than it definitely is »Read the fucking manual!« and probably read it twice.
Personally, I had decent experiences in most cases. I never got over the fact that it refuses to become the de-facto point of reference for stacks, i.e. questions like "what are the best web libraries for $language", "what is the best practice for deploying $framework" are brutally rejected. And it's a shame, because those would have helped immensely -- it's very easy to learn a language, but it's really hard to figure out what you should use from the surrounding ecosystem.
Problem is that most people don't want to maintain these lists. The handful of exceptions stand out precisely because they have a few dedicated maintainers who've dutifully kept them relevant for years... Most of them, though, died from neglect and were eventually deleted out of mercy.
StackOverflow is a good resource precisely because it's curated and maintained, including rejecting content that doesn't work on the site. Just like Wikipedia doesn't allow you to have your own page listing your opinions.
That's why you need voting, same as traditional SO.
> and go out of date incredibly quickly
I can see that being a problem -- but it's a problem even for regular SO questions, since languages evolve (the best way to do X in python 2.3 is not the same as in python 2.4, let alone 3.x).
You could simply timestamp votes and progressively devalue old ones, like tajen suggested.
> If you search for such a thing, you'll find blog posts [...] useless because they are out of date
Absolutely. Blog posts and wikis are terrible at keeping up, they are fundamentally static and require a lot of work. For example Fullstack Python [1] is nice but it's probably an immense effort. It would be nice to have some semiautomated way to highlight stacks that can self-maintain and stay relevant.
In my opinion, it's just another unfortunate side effect of anonymity on the internet. It's terrible, and I wish people were nicer to each other.
The result? The only people that comment are trolls with fake account, people who have too strong opinions or don't care about their next job interview or people with "correct" opinions.
Meanwhile I rarely see anyone here or at the Ars Technica comment sections suggest we nuke one country or the other (and if anybody do we collectively moderate them beyond ground level.)