I have a small number of answers, and the most successful / least criticized ones were always on niche languages or topics. C/C++/Java? No chance. I corrected a wrong answer on a question in C++ once, which was immediately reverted with “nope, your answer is too different from the original, make a new answer”. Great, so now there’s an answer that is fundamentally wrong with 26 upvotes that apparently can’t be corrected because the correct version of the algorithm is too different from the original. I just stay away from answers in those languages now, they’re more effort than they are worth

> I corrected a wrong answer on a question in C++ once, which was immediately reverted with “nope, your answer is too different from the original, make a new answer”.

This is how Stack Overflow is designed to work. One single user is not supposed to be able to edit a top answer and change it into a different answer.

The Stack Overflow Way™ is to post a new answer that tells the correct answer and why the other answer is wrong. Then the community has two separate answers and they can vote on which one is more correct.

Confusing? Yes. Arcane? Yes. Did the system fail in this specific circumstance? Absolutely yes.

But they can't just let people edit highly voted answers into fundamentally different answers. That's why your edit was not approved.

That's apparently the intended design; SO bills itself as "questions and correct answers", but they've made it clear in the Meta site or in blog posts, I don't remember where, that the accepted answer and upvotes are not "this is correct" but "people found this helpful", and people can find "wrong" answers helpful, so they stay.

The intended outcome here is that you post your own answer, and leave a comment on the most upvoted one summarising your concern with it.

And that readers don't just pick the most upvoted one, (even though it's obviously the site's design that they should be able to do that).

So on your last point, I found something sneaky you can do to people:

The answers have a sort order, controllable by tabs below and to the right of the question. These go to the same page with a GET arg to change the order. The trick is that these are sticky - browsing to one saves it as your preference. So if you post a comment to a question and include that arg, it'll switch the whole site for whoever clicks it, without telling them.

One of the options is "newest first".

That kind of behaviour is what's most annoying about their site, their conventions sometimes overrule common sense or intelligent queries/dialogue.

Last I contributed anything on there I was being told by a single poster that my question was 'wrong', and descending into an argument - and started throwing SO convention at me. I just deleted the question. There's a sociopathic nature about those looking for ratings and authority that's totally at odds with simple sharing of knowledge and learning.

Same here, I’ve been a reader for 10 years or however long, but their rules were too difficult for me to follow in order to vote. I had to visit every day and do menial, non-useful things just to be able to vote or answer. This didn’t work for me because SO wasn’t something I typically visited every day. So I just couldn’t get into the habit of being enough of a user to browse and vote so it was a vicious cycle of apathy.

I appreciate SO and use it many times a year. I remember the world before SO (experts exchange was very bad), so I’m happy it exists. And it’s obviously getting along without my comments and moderation.

I always wished it would pair well with something like HN so the reputation effect from a daily activity would carry over to an infrequent activity.

I also really respected Joel Spoelsky so didn’t really have a constructive criticism as the site worked pretty well.

> their rules were too difficult for me to follow in order to vote. I had to visit every day and do menial, non-useful things just to be able to vote or answer

Has it changed a few years ago? When I started 3 years ago, I could answer right from the start and I got the ability to vote quickly enough, after a pair of vote on an answer, without having to do menial tasks every day. Reputation can also be gained with edits, is that what you are talking about?

Answering questions doesn't help unless those answers get upvotes. As a newbie, your answers are less likely to get upvotes, especially if you're slower than the old timers.

It's true. For my part, I started (and continued to) answer on a tag with not a lot of questions and answers, so most/some of the time I'd write the only answer. Though I had a professional interest into the tag at the time, I didn't choose it because it was easy to write the single answer on a question.

And that's an example of the system working at its best. You were incented to answer questions nobody else could or would answer. In my experience it doesn't work out that way often.

That’s funny, maybe it did change. I haven’t tried in many years as I gave up.

I definitely learned a lot of my attitudes about software engineering from "Joel on Software". It's weird to think that those works are "ancient" now.

I don't really have an answer to the problem, SO has some very smart people working for them who, I'm sure, think about this issue. For some reason technical discussions held by competent people on the internet seem to trend towards dumpster fire. I've also noticed this phenomenon on photography discussion groups.

I expect to burn a few karma points on this, and be ignored, but all evidence I have says that this is human nature at work, and that at scale it cannot be prevented.

I've seen moderators of small communities hold off the effect by setting a good example, using a deft touch, and not letting the community get too big.

My suspicion is that once the number of commenters gets much past Dunbar's number [1], the universal fall into "default jerk" happens. Once that's taken place, you're stuck in an Eternal September.

Communities in the real world don't scale well. I'm not sure why programmers expect them to on the Internet, with its well-known capacity for dehumanization.

1: https://en.m.wikipedia.org/wiki/Dunbar's_number

I think there's something to this. Not sure if it's Dunbar's Number specifically, but something. It seems like there's a pattern where a community is nice and small, with people joining at a trickle. Things change when a jerk joins and starts doing jerk stuff. This tends to drive away the nice and helpful people, attract more jerks, and pushes people still there to be more jerk-ish themselves. Sometimes deft management can prevent it with some well-timed bannings, but I think this gets increasingly difficult as the community gets larger, and may be effectively impossible in internet-scale communities.

You're describing the broken windows theory [0], it's not related to Dunbar's number.

[0] https://www.britannica.com/topic/broken-windows-theory

Yeah, I'm not sure it's Dunbar's number precisely - it was just what came to mind.

They could start by simply making it just a little more difficult to give negative feedback. For instance: Require that every down-vote is given a written reason for it.

And perhaps make it so that DOWNVOTES CAN BE DOWNVOTED (provided a written reason is given for that as well).

> SO has some very smart people working for them

I think they don't have enough (or any?) moderators working for them. They rely too much on making profits with the free help of trolls.

I understand it would be hard for them to change the power of the troll-users by now. Perhaps we simply need a new Stack Overflow, some competition.

You don't even have to have an account to post an answer, nor any points to post a question. You need only 15 points to upvote (2 upvotes on an answer, or three on a question).

Or no upvotes and 1 "accept" on an answer.

Experts Exchange, where the paywall could be bypassed if you scrolled down far enough.

At this point I just ask a question, get an answer, and if I don't like the commentary, I just delete it. I post there under my real name, which means I either have to address or accept dumb/nit-picking criticism, which is a no-win situation.

If the person who answers actually needs the points, I'll give that some consideration, but it seems like a lot of the time the solution is offered in the comments anyway. I don't know why that is.

Every once in a while I'll swoop in on a basic question and try to answer it before anyone else notices.

People deleting questions after they have been answered is one of the reasons I stopped contributing. It's so energy draining to take the time and explain something to somebody and do it in a way that others with the same problem can understand & apply it as well only for them to delete the question for some reason.

I prefer not to delete, especially to an official answer, but it’s hard to let the obnoxious comments stand at the same time.

You can now report comments for being rude/unfriendly/unkind/condescending.

You and the reply below have BOTH deleted questions from SO because of social overhead. That's a surprising outcome that's never occurred to me. I can imagine that many interesting questions have disappeared this way.

I have noticed answers in comments a lot too. It seems like some aspect of the gamification. One theory is that people don't want to put any straightforward answers in actual answers because they seem anything straightforward to be unworthy of an actual answer.

I have my own standards for what makes a proper answer, and if I'm too confused or rushed to meet those standards I'll just leave a comment instead.

I've been yelled at for doing that on Stack Exchange sites other than SO.

Agreed. Similar experience. Asked a question about architecture and dependency injection in .net. it was immediately removed as irrelavant because .net is better than Java.

Posted a week later with a friend's account and it stayed but my friend panicked at loss of points and inability to do their job without them and removed it themselves (tho gave me answers firtst).

I thought dependency injection in Java was mainly a way to emulate modules (as in Python or JS), so how is .NET different here since it doesn't have modules either?

It just turns into a "ACTUALLY!" contest, of who can pee higher. No matter how right you are, some guy will link a blog post detailing why in some obscure scenario that might not be valid, and proceed to downvote your obviously wrong answer.

That's strange, in all the time I've spent on SO sites, I can remember only one instance where that happened. Maybe it's community related and happens with certain topics more than with others?

Happens a lot in questions related to C. Most answers are followed by a discussion why the code provided in the answer is not portable. Oftentimes answers are more concerned with being technically correct than helpful.

In the particular case of C, it's often important to be pretty precise. If an answer is offering a non-portable solution, or worse still, relying on undefined behaviour, it should get called out. It's not like, say, Java, where you're generally unlikely to get bitten by dark corners of the language.

You're thinking about it all wrong. Don't get lost in the gamification. SO is intended to create a repository of useful answers to useful questions. If you provide a useful answer to a question which gets views, people will read your answer. Over time, many people will read your answer - and that is your "return on investment".

(Also, if many people read your answer and approve of it, then eventually you'll get some votes too.)

Having said that - there is some merit to sticking by the (nastily-enforced) standards. It does get people to improve their answers. Not saying that I approve of people's conduct on SO - it can be quite terrible.

On the tags with less question (<5 per days), it has never happened to me, I even managed to be among the top poster for one of them. And the few times I got corrections from more experienced posters, they were pretty cool about it. But then it's really tag-dependent.

This is the same for me, although I contributed for a while almost 10 years ago, and managed to earn just enough reputation points (that still increase a bit here and there) that I can get some ROI by asking questions and adding bounties. Two or three times they have saved me an immense amount of work trying to figure something out that otherwise might have taken weeks or I would have just given up on.

Another thing that is a problem is a senior guy like you contributes with much higher level, less specific advice than the kind of thing that gets points.

For instance "how do you concat strings" for various languages is highly upvoted. Senior advice about general code structure and maintainability doesn't really have a title.

I tried contributing regularly on Stack Overflow a while back, and it just wasn't fulfilling to me. These days I just post follow-up comments mostly on topics which generally have light coverage anyways.

Same with me. I do sometimes answer questions if they fit into a niche I have some particular knowledge about, but on the whole I've found it to be a mostly negative experience contributing on SO.

I feel the same way. I tried a couple times to contribute but found it too demoralizing.

likewise I also don't contribute questions either. I always say its a great place to find answers, terrible divisive place to ask questions.