The thing with the RTFM moments, is that, when I'm experimenting with a new technology, I often ask Google a question and it points me to a stack overflow page.

Yes, the question is obvious. And yes, it can be answered via 10-20 minutes of RTFM. But, the value of Stack Overflow is that it replaced 10-20 minutes of RTFM with a simple answer.

So, don't knock the RTFM questions on Stack Overflow. That's part of its value to the greater community.

(I must admit that I've answered very few Stack Overflow questions because every time I find an answer. I did take the time to write out a well-written answer for an obscure problem, and it got me plenty of points. It's nice that the community has a 1000x ROI, too.)

> But, the value of Stack Overflow is that it replaced 10-20 minutes of RTFM with a simple answer.

But at the same time you are learning with blinders on, i.e. you miss out all those things you would have also learned by reading the documentation.

This is not particularly bad when you're just starting out. But if you see RTFM-level questions from people which appear to already be in the business of building complex applications then you start to wonder if this kind of thing incentivizes help-vampirism.

Personally I prefer to answer niche questions or things that are not well-documented yet.

> you miss out all those things you would have also learned by reading the documentation.

That's assuming that the documentation is articulately and technically well written. Unfortunately that's more often the exception than the rule. That's rather the whole reason that StackOverflow even exists, I think.

It's also assuming the SO answer is less considerate of context than the documentation; this can be the case, but I find quite often good answers are quite comprehensive and explain things broadly.

thats exactly the point.

before: questions were always open. and over time people added those details.

now, after someone answer with a one liner and gets accepted, the question is "closed to prevent 'me too' comments"

it's going downhill for the same reason every internet community does: power crazy moderators who completely misses the point of the site they moderate.

The "me too" functionality is protection, not closing - registered users with a certain amount of reputation can still answer them. Similar to Wikipedia's system.

Every "me too" protected item I've seen has half a dozen "me too" answers deleted at the bottom. It's likely necessary.

I don't think it's just the moderators, there's also an incentive problem because a popular question/answer has no limits on its point-payout.

This leads to a kind of popularity gold-rush pattern. Hard important stuff languishes while easy stuff is has a glut of volunteers.

I'd be interested to see what happens if questions and answers only gave up to a fixed limit of Internet-points to the authors or contributors. Would it lead to a broader knowledge base?

I agree where things are badly documented. But look at the top tags, java and javascript are generally well-documented.

Nope, they are not particularly well documented.

I find JS is well documented on MDN, but doc writers can't possibly imagine all edge cases and interactions. A lot of these cases are "If you knew about that simple feature of instruction A and the simple feature of instruction B and how they work together then you can solve this problem".

Oh. I actually clean forgot that MDN has docs on JS, not just DOM :)

Yes. You're correct about that

Or that the docs exist in the first place.

Meh, it's just another resource. Sure you can make a case it encourages help-vampirism, but I would argue that help vampires are largely incapable of becoming competent anyway. Basically, anyone that will become a competent programmer has a hunger to learn and fit all the pieces together, and will be frustrated by endlessly googling and posting questions that take hours or days to receive a reply; no matter how much googling/documentation is involved, ultimately they will learn to synthesize solutions efficiently. Meanwhile there are many people who believe being able to program is just a question of accumulating enough facts and they never progress past asking whole-cloth questions. I'm sure some are capable of crossing the chasm, but if it doesn't happen early I don't think it ever happens.

For a help vampire, even SO is not rich enough to make them competent. For a competent developer, SO is not efficient enough for them to learn helplessness.

Generally if I'm going to SO instead of RingTFM, it's because I don't particularly need to learn. I just need to fix this one thing that's tangentially related to my work. And usually my deadline precludes me spending time learning it anyway, even if I'd like to.

You'll become a better programmer in the long run if you take the time to resist Google and read books/manuals. Unless you're on a deadline I guess.

> My top 3 answers are one-liners that solve "problems" any semi-competent developer shouldn't even need stackoverflow to solve.

Remember that SO's main traffic source isn't experts, it's novices searching for answers. I did some digging at one point, and the most popular questions on ServerFault are extremely basic, like 'how to make a symlink' basic (viewed 1 million times https://stackoverflow.com/questions/1951742/how-to-symlink-a...). If you're a developer or student just trying to figure out this one UNIX thing so you can get back to JavaScript or CS:Go or whatever, this is great.

I agree and appreciate that, but I don't know that it's always true. The problem is that a lot of documentation and man pages out there just aren't written well. They bury the lede and make you dig around to figure out how to do that simple thing you want to do. Googling for something and immediately finding that specific question on SO even if you could go look it up in pages of docs is a time saver.

Edit: Or perhaps the library has poor usability. I've been doing python programming for 6 years now and know the standard library like the back of my hand, but if I have to do anything reasonably complex with datetime you can bet I'll probably just look for someone on SO that has had the same problem and crib the answer.

Documentation should contain a complete reference off all options. StackOverflow rather describes... usecases.

It sounds fine to me, because the programmer can't think about all usecases, so they better be on a user-managed website.

I've answered a bunch of narrow questions like the ones you described, only to see them closed later because the community decided it doesn't want to include those questions. My attitude has always been, "Speak for yourself. I'm fine answering them." But, of course, they still get closed, so now I don't even bother.

There are people (like me) willing to kill some time solving useful problems that people in the real world are actually having, but SO's theory of broken windows is so extreme that they want to make me feel bad for doing that.

You can blame Jeff Atwood. He outlined his theory of the descent of sites into irrelevance in some blog posts, then put it into action. There's code on SO to prevent two questions having an identical title for example. Even though he's long gone from SO management, the legacy of his philosophy is still abundant.

It's weird that they're closing Documentation, because my experience is that SO doesn't actually want to be a Q&A site, it wants to be a crowdsourced documentation site disguised as a Q&A site for some reason. They'll close legitimate questions because they're not broadly applicable, which isn't really how Q&A works.