The real value of an SO answer is probably best approximated by how many people it helped. In my experience, SO is incredibly useful when I'm learning a new language or framework, and I want to quickly blast through hurdles caused by non-intuitive syntax or functionality. In those situations I'm on there almost constantly. For popular languages like Objective-C for iOS, there are thousands of people who are in that same situation.
Surely there must be a way to capture the value-added for all those thousands of people in the reputation system? Many people get to these answers from google, so it sounds like a simple matter of pinging them to (sign-in) and upvote the answer that helped them out.
If, on the other hand, as the author says SO is no longer actually values reputation, and is screwing with their algorithms accordingly, that is a serious problem. Not sure what can be done about that.
It's a negative dynamic that starts from the popular, "low-hanging fruit" kinds of questions like "how do you subtract 2 dates in framework X". Such a question would gain many points both to the OP and the answerers. The best answer to a question like that is short, unambiguous and concise. Consequently, you have high-reputation users skilled in answering popular questions. Then the popular questions are mostly over and people start asking questions about a particular problem hidden in their piece of code, and high-rep users sneer at it because they see it as "not researched".
> The only major broken part (IMO) of SO is that the SE network is becoming so disjointed and your rep doesn't transfer between sites.
By design, I would argue. And a good decision, too.
To use a facetious example just to illustrate my point: without separate reputations, one could build up a huge reputation on http://apple.stackexchange.com/ and then use that influence over at http://askubuntu.com/ to provide a mediocre answer.
This mediocre answer could then get upvoted to the top by those who do not know a better way purely because of this user's reputation. Meanwhile the perfect answer provided by a newcomer to the SE network may remain hidden in second place.
N.B. I think the "bonus" reputation points that can be transferred are a way of addressing this imbalance while pushing people to use other SE sites at the same time.
> A lot of time has been spent by developers trying to get points. There is a cost to that.
Agreed, however one answer will help many developers, which should more than offset the initial effort
> The site has enabled software vendors to slack off on their documentation.
Documentation has not deteriorated, in my opinion. In fact, a point could be made that it is now much easier for small vendors to have outstanding developer support through stack overflow. A lot of open source project use it as the main support line.
> It would be interesting to know what the global number for reputation points is. There may be a correlation to money saved.
It is probably correlated to the time spent on the site, but the money saved is certainly not related, as users don't gain any reputation by reading answers or simply browsing the site.
One person asks a single naive question once: "How do I undo the most recent local commit with GIT", gets 200,000 reputation.
Another person answers 100 complex, delicate questions useful to high-performance computation underlying widely-available cloud services; gets an average score of 5, has 5000 reputation.
Another person occasionally asks and answers a few questions, reaching 2000 reputation, but does a lot of editing work, triages new posts, fixes up tag pages etc.
Almost all my reputation is from a question [1] that I just happened to see and answer before anyone could close it as "bad quality" or "duplicate".
Had I been a minute or two later I suspect it would have been closed as poor quality.
I only spotted it by pure chance because I had just wondered the same thing before having a "duh" moment when I realised it was a really simple solution to what sounds complicated (but isn't).
So I was able to interpret the "real" question and answer it.
But, as you say, never a bit of feedback from the question asker. No acceptance, just abandonment.
It just slowly gathers karma, around 10 per month for ~10 years.
> One person asks a single naive question once: "How do I undo the most recent local commit with GIT", gets 200,000 reputation.
I imagine the person that asked the question about git was the first to ask. There's a benefit to being the first, or asking unique questions. The value of the first time that was asked and answered usefully on SO is vastly greater than the 100th time, and the value of having it at one time compared to a year later is also great. Sometimes what matters is how it's asked, so it looks like other people's questions.
> Another person answers 100 complex, delicate questions useful to high-performance computation underlying widely-available cloud services; gets an average score of 5, has 5000 reputation.
I would guess those complex and delicate questions didn't help a lot of people, or they would have upvoted, right? Whether that's because people didn't see them or it's so esoteric as to not really help many people, the result is the same.
> Who has been more useful to the site?
The person who asked a question that got a useful answer first, or that answered a question usefully first, is greater.
Stack Overflow would be a pretty shit site if the first time someone answered that git question was in 2021 and not a decade or more ago, so optimizing for people identifying, asking and answering unique questions that match the questions other people have and are easily identified when people search for that problem makes sense IMO.
In the mechanical turk metaphor, you're willing to pay a lot for someone to take out your overflowing smelly trash because that's what you need, but you're not willing to pay as much for someone to do the same thing a day later if it's not even half full and doesn't smell. You reward people more for doing the things you need.
> I would say that reputation numbers don't necessarily measure engagement per se in S.O. very well.
I've found that if I answer a general JavaScript question, even if its an absurdly simple one, I garner more reputation in a few minutes than if I I give a detailed answer to 2-3 in depth canvas questions.
Here's why: SO points are not based on complexity or thoroughness, but based on value - value that is perceived by the community. A quick answer to an simple Javascript question may seem rudimentary to an experienced programmer, but if it has hundreds of points, the community has determined that the answer is helpful in a big way. Chances are there are many, many people that have experienced the same problem, which makes an answer to that problem - no matter how simple - incredibly valuable to the community as a whole.
Edit: This is probably obvious for SO users, but it's worth mentioning for anybody that is confused or wondering why the point system works the way it does.
B) Someone is gifting free interesting content to the site; even if it's imperfect does the reaction have to be so harsh? This is, personally, exactly why I never post on stack overflow. You're acting like you're doing them a favor by letting them post there, when they're doing you a favor by giving you interesting content in the first place.
In addition, the reputation model is actually incentivising the wrong thing. Answering some simple question that many people see will give you an answer that gets upvoted to the double or even triple digits. Answering something hard and well might only get a handful of votes if it's a niche topic.
(disclosure: 11K rep, mostly on tools like make, autofoo and cmake.)
To me, the #1 problem with Stack Overflow is the number of plain bad answers that are given and often accepted. CSS question? Here’s someone’s un-researched and invalid first guess, that doen’t work in practice. +10 karma and accepted.
Hard question? Here’s some loser’s gut feeling of “can’t be done” in exchange for +2 karma, and then a knowledgeable follow-up a day later from someone who has actually been there, done that, and figured it out.
So often it’s apparent people are just throwing best-guess answers out there for karma, and this is hugely unhelpful. There should be a large penalty for stating incorrect guesses as fact.
It’s to the point where I instantly mistrust any answer from someone with over 10k reputation as I learned there is a good chance they are just shooting from the hip for karma.
everybody is arguing about why the user wants to do it that why
I fucking hate this, and it's been happening forever. It's those fucking assholes farming points for rep, who don't really want to help OP but still want their answer accepted.
You see it on SO too, people answering basic questions 2 seconds after they're postedd with long pastes from documentation that are sorta related to the original question.
The reputation might be a sort inverse filter: thousands of people participate in the programming jokes questions whereas even perfect answers to questions like "Under conditions X, Y, Z how can my foo do bar?" are unlikely to get much love.
It's those hard questions I love StackOverflow for (and there's no rush of people sniping them).
The SO community culture rewards precise, concise and direct answers to simple and popular questions like "how do you subtract 2 dates in Python". In the first years, those were the low-hanging fruit and a source of big reputation points for some people. Unpopular questions regarding an obscure bug in the user's code or perhaps a misunderstanding of the OP aren't seen as "pretty" - they require answerers to understand the context of the particular problem and will not gain them many reputation points, even if they provide a useful answer to the OP.
My SO experience lately has been the accepted answer is 12 years old and the correct, modern solution is down at the bottom with 1 point and I can't upvote it because they want me to jump through some hoops to get "reputation" before I can vote
At this stage SO is a reference work. The kind of things it's useful for are things like "what's the syntax for [algo] in [language]". There's specific answers people are looking for, and they stay that way.
It also means everything(tm) is there. It's been ages since I've needed to ask anything at all there. Someone has had the problem before, and gotten the answer.
Individual reputation might not matter much in this context. If the answer to your question has 100 upvotes, why does it matter who they fell to? If someone answers a question, does it matter how many other upvotes they got? Does it make them more authoritative?
You'll find loads of simple answers with loads of upvotes, and deep dives with just a few.
My own account is just as odd. The most upvoted answer by far is a one-liner that says "untick this box". That time I found an actual error in a Swift lib, I spent ages detailing it, and got barely any votes.
No, the issue is that, in SO karma, relatively common knowledge seems to be valued orders of magnitude more than insight on problems few people would be able to solve. While some of those questions are probably just obscure trivia, it also includes advanced knowledge of (say) compiler optimization. Cases where most people on the site wouldn't be familiar enough with the subject matter to tell an insightful answer from ignorant ranting.
It can also be a measurement of how much time you're willing to sink into talking about programming, rather than actually doing it. (HN has the same problem.)
I responded to a question on SO once where I elaborated on the accepted answer - my solution was more functional as it provided assignment operator override.
Some hot shot user with an absurd amount of badges told me, in a condescending tone, that my solution did not answer the question and linked me to some TOS article or something about staying on topic. This was the first time I had ever answered a question on SO.
My takeaway from the situation is that SO is full of accounts that farm badges/rep. To what end, I do not know - perhaps they reference it on their resume or portfolios.
I called the guy out, seems he has since deleted the comment.
https://stackoverflow.com/questions/3413470/what-is-stdmove-...
reply