What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 131 Stack Exchange communities.

Recently, I have answered a few questions that were upvoted by factors more than the accepted answer, but not accepted. This one, at the time of my writing, has an accepted answer that the asker said was not what he wanted. I don't know why. In the future, when people view this, people might only look at the accepted answer and move on. Would it be possible to have a "community accepted" feature? Maybe only the people at the top can vote to accept it or something. How much rep, if any, should go to that person?

share|improve this question
1  
@George, final note on your situation in particular. In this case, the community accepted answer would be WRONG. The poster said explicitly that your answer was NOT what they wanted. I'll grant that the answer they marked as correct was just as 'unwanted', but this is a case where neither should be accepted. –  devinb Jul 8 '09 at 21:55
    
no notes are final. i dont see where my answer was not wanted. i see that he didnt want br, noting mentioned about wbr or its alternatives. –  George IV Jul 9 '09 at 1:02
44  
Oh the irony of this question not having an accepted answer. –  Dominic Rodger Mar 1 '10 at 13:17
1  
@GeorgeIV What's br and wbr? –  Konrad Viltersten Feb 13 '13 at 8:45
    
@Dominic Rodger, not that much as no answer fulfilled the needs of the asker. Nevertheless, the community clearly showed the most reasonable option. Sometimes, it can be more complex : I upvoted the best answer, but still would think that sometimes it could be fine to designate an answer for top users. SOF, at least on the Android tag is growing with unanswered questions, and it can give a bad perception of it. I think it would be worth addressing that. But ok, after all, that's not that bad the way it is, and we surely avoid tons of conflicts. :) At least, that's what my votes means. –  Snicolas Dec 21 '13 at 0:19

11 Answers 11

The number of up-votes shows the level of acceptance by the community.

If the "accepted answer" has no up-votes, but yours has 12, clearly yours is the community-accepted answer :)

share|improve this answer
    
Yes, I had that thought, but what about the case where the asker forgets to accept and the discrepancy is not so huge. –  George IV Jul 9 '09 at 1:05
24  
If the discrepancy isn't that big, then it's not really fair to call one the "community-accepted answer" :) –  Jonathan Sampson Jul 9 '09 at 2:52
    
Couldn't that also be because maybe the question is about a very specific / low level issue in which only a few have interest? –  TweeZz Jun 3 '11 at 9:44
    
I agree; I generally regard a lot of upvotes as a means of determining the approach generally preferred. Having an additional flag seems redundant. –  Lynn Crumbling Aug 3 '11 at 17:32
    
What about when the accepted answer is just factually wrong? This results in a very misleading display to non-registered users of the site, who always get the accepted answer displayed first, and visually distinguished (green check mark next to it). Not ideal. –  Konrad Rudolph Jun 13 '12 at 9:31
    
@KonradRudolph - then you flag it with reasons and it can get unaccepted. –  bPratik Aug 29 '12 at 0:06
    
@bPratik Flagging is generally not to be used for factually wrong answers; flagging is for non-answers or objectionable content. As a moderator on one of the Stack Exchange sites, this would probably be an invalid flag as far as I’m concerned. –  Konrad Rudolph Aug 29 '12 at 6:52
    
@KonradRudolph - even if the answer is utterly blatantly wrong? I think I have flagged with proper reason before and it was actioned... –  bPratik Aug 29 '12 at 9:15
    
@bPratik -- You should also upvote any articulate critiques in the deficient answer's comments. Leave a clearly critical comment of your own, if you're able to add something new. –  zero2cx Sep 21 '12 at 8:30
    
That's working best in an active tag like Java, C# and JS. When you work with a smaller field, you're lucky if you get any upvotes at all. When 10000 people read you, can cast a vote and you get low, it says something. If 3 people read you and only two can cast a vote, the low score means jack. :) –  Konrad Viltersten Feb 13 '13 at 8:50
    
@Jonathan totally off-topic - I sent you a couple E-Mails last year, but never received a reply. Not sure whether they ever made it to you? –  Pëkka Feb 14 at 17:06
    
@Pëkka Reach out to me on twitter, please. –  Jonathan Sampson Feb 14 at 19:37

The "accepted answer" feature was never intended to mark which answer is best or even if the answer is correct. It is, simply stated, the answer that the original author found most useful in solving their problem.

The people's-choice favorite answer is selected through the voting process. The "accepted answer" is all the original author. If you want to let the users also select the "accepted answer", the purpose of having a selected answer becomes redundant and useless. That decision has to come from the author, or you're just conflating the two features.

share|improve this answer
    
+1 Well explained...:) –  Sachin Shanbhag Feb 23 '11 at 5:19
    
That's a reasonable explanation for the intention and state of things. But it doesn't completely address OPs point: users that neither know or don't bother to fulfill the past tense of "found most useful". (But of course there's a whole lot of related meta topics.. :) –  mario Feb 23 '11 at 6:54
2  
@mario: The author's feature request was to let the readers accept, in lieu of a forgetful/uninformed question author. My post says "no, that is not a good idea." –  Robert Cartaino Feb 23 '11 at 15:45
    
@mario: And how could anyone except the OP know what was most useful to them? –  Hendrik Vogt Feb 23 '11 at 17:41
4  
@Hendrik: "thanks, you solved my problem!" –  mario Feb 23 '11 at 18:00

The site is at it's heart a 'question & answer' site. Although many people have taken to posting answers which appeal to the community at large, there are cases where they do so at the expense of the actual question asked by the poster. An example (NOT REAL) situation:

Question: I know this isn't best practice, but restraints force me to do XYZ, how can I implement it (or work with it) to minimize the risks/down-sides?

Answer: It is NEVER acceptable to do XYZ.

The answer will then get upvoted hundreds of times, even though the answerer did not answer the question. Someone else may attempt to answer the question correctly, but because the community is so set against doing XYZ, they won't upvote that answer. Then the poster marks it as correct.

In this case, the correct answer (the answer the poster was looking for) will appear directly beneath the question, and the 'best practices' or 'community accepted' answer will appear right below it.

There have definitely been times where I think the OP picked the wrong answer, (especially when one of my own beautiful and well-crafted answers are in the mix) but since I am not the asker then I do not know exactly what he/she is looking for. Only they do. You may have answered the question to the best of your ability, but only the OP can select the answer that helped them the most.

share|improve this answer

I would be in favor of a notification -- if there isn't one already -- when an answer receives more votes than the currently accepted answer. This would allow the OP to easily revisit the question and change the accepted answer if a better one came along.

share|improve this answer
    
I like that idea +1. That helps them out and future readers of the question who may not read anything past the accepted answer. –  George IV Jul 9 '09 at 1:06
    
Maybe with some kind of threshold, so that going one or two above isn't a big deal. It's only when there's a substantial margin that it appears significant, and even then it might not help much (the asker might simply have gone away). –  Donal Fellows Nov 10 '10 at 13:40

We have a badge called Populist which addresses the fact that the community finds your post much better than the accepted answer. Implementing something along the lines of "Community Accepted" is incredibly redundant since the community is voting on it with their upvotes.

share|improve this answer
    
Populist is fairly difficult to achieve. I advocate a silver version. Maybe 10/5? But that helps me more than people who go to that question in the future. I could not care less how many badges I have. –  George IV Jul 9 '09 at 1:08

@Jonathan idea may hold true for popular tags. But take or some other less popular tag. There are very few people who contribute, and most of the questions are done by new users who won't upvote nor accept an answer. Eventually those who participate in this non-popular tags will get sick of spending time helping others when they known their work will end up being unrecognized. For instance, I gave 72 answers in and got only 33 upvotes (I encourage people to review my contributions).

The fact is that we can't trust new users to accept answers when their problem is solved. Sometimes the OP even adds a comment saying "Thanks, that solved my problem!" but it never accepts the answer and also never comes back. This is really frustrating and it's a real problem.

We need a mechanism that allows the community to accept an answer.

My suggestion is that users who own a silver tag-specific badge (that is being used on the question) should be allowed to cast a vote to Accept an answer. Pretty much the same way we do it to close/reopen a question. I already stated this idea before and later added that it might be better if we deny/prevent users that answered the question to also vote for the accepted answer (to avoid a potential conflict of interest).

share|improve this answer
1  
How does anyone get a silver tag-specific badge on a less popular tag? –  Ian Ringrose Apr 19 '11 at 13:49
    
Working as a camel. For non-popular tags we could use a bronze tag instead. –  karlphillip Apr 19 '11 at 13:52
4  
+1 for can't trust new users. It's a crapshoot when answering a question where the OP has a rep of 1. Are they even going to read my answer? Will they acknowledge that it worked (or didn't)? Will they accept it? Makes me not want to waste my time unless the user has a minimum rep AND a minimum acceptance rate. And if others think like I do it makes it harder for new users to get answers (the bad seeds spoil it for the bunch and all that). –  ggutenberg May 1 '11 at 20:45
    
+1 for point out the real issue with less-popular groups. Sadly, people running the site are members in very popular groups and I suspect they don't quite understand the problem (they know of it but not really feel it). –  Konrad Viltersten Feb 13 '13 at 9:08
    
@karlphillip +1 Agree. That's quite a problem. Happens in .htaccess and mod_rewrite tags also, that I know. But most probably there wont be enough people to accept the answer either. An average of 6 or 7 visitors, that usually don't have a clue about the subject, are there just waiting for their questions to be answered and don't know or don't care about accepting or upvoting, hardly will vote for another answer to be accepted. –  Felipe Alameda A Mar 19 '13 at 11:02

Wow, this is quite a blast from the past (so to speak). I actually suggested this many months ago, back when StackOverflow was still in private beta:

Every part of Stack Overflow is run and moderated by the community except one: the selection of the accepted answer. As long as you have one person in charge of selected answers you will always have to deal with their particular biases, which decrease the value of Stack Overflow as an objective reference to programming questions.

It might be a good idea that in addition to the answer selected by the asker that there be a community selected answer. This would be completely orthogonal to the up/down votes for the questions, and would only allow one selected answer per person per question. An approach like this would greatly serve to make Stack Overflow a more equitable and fair system.

Here is how it could work:

Expose the "select answer" link to everyone (or perhaps just registered users) and allow them to select at most one answer per question. The answer with the highest number (5 or more) of selections becomes the community-selected answer. To reduce potential for abuse, there should be no reputation or badge associated with this answer selection.

(this is in response to criticism that Stack Overflow is merely the "blind leading the blind": http://blogging-harmful.blogspot.com/2008/08/stack-overflow-blind-leading-blind.html)

And the official response was:

I did. Votes and sorting are the de-facto community answer.

community mode essentially achieves this; see FAQ. Posts automatically get moved into community wiki mode when certain criteria are met. Also, votes are de-facto choosing the community accepted answer; it's the one with the most votes!

It's certainly an interesting suggestion, but unless something has changed in the nature of StackOverflow in the intervening months (as well it could have) I think that the official response would be much the same today.

share|improve this answer

I think the idea is to have it stick to the top, right below a person's accepted answer. Not everyone views posts the same way, and if the community thinks that an answer is best, and it's on page two for me, I might not see it. However, if the community can not only upvote, but say "this is the right solution", then the "right" answer will stick to the top for everyone.

share|improve this answer

Why can't we just have a 'community answer' in addition to the 'normal answer'?

Just because I upvote something, doesn't mean it always answers the question, it may contain very useful information which the actual answer has missed out.

share|improve this answer

The issue you're addressing is basically irresponsible askers or users that don't use the site features correctly. It is almost the same issue as users who don't come back and accept an answer after asking a question.

Unfortunately a type of cruft. I think that for users that un-signup as it were (their name appears gray), something should assume ownership of that question and be able to move that checkmark around.

share|improve this answer

How about not giving "community accepted" status to any answer, but instead adding "accepted by community" indicator to the question. So it would not bring any extra reputation to the answer with most votes, and it would not be a property of that answer or shown in the answer at all.

It would be applied to questions which have an unaccepted answer with some minimum amount of votes, which is also more that votes on accepted answer (if there is an accepted answer).

It would be useful when looking at this question list (note: user specific listing), at least.

share|improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .