What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 130 Stack Exchange communities.

Sometimes users don't bother marking answers as correct, even though it might have a valid answer.

So, how about a way for experienced users to mark solutions?

If the questioner has not visited a question in X days, a control appears for users over a specific reputation to mark what they view as correct, and if enough users unanimously vote for an answer then it is accepted.

This action would notify the OP, who would then have the option to come and override this.

There's probably all sorts of badges and rep that could be worked out around that, but it'd be nice to simply get a clearer picture of actual solved/answered/unanswered questions.

share|improve this question
1  
I think this is a great idea. I guess that it could work like the vote to close a question. –  Barry Jun 24 '10 at 12:18
1  
I just write "###Proposed solution" in my answer, and folks up-vote it... –  Shog9 Jun 24 '10 at 14:07
    
Maybe a better option would be for uses to not get their points for question upvotes until they select an answer. Or something like they only get half of the points until they select an answer. –  Matthew Whited Jun 24 '10 at 14:09
    
@Matthew I'd disagree with that as well. That punishes people for not getting a correct answer, which is a pretty mean thing to do. Plus, we already have an accept-rate for that, and reverse motivation in the form of the +2 author reputation for accepting an answer. The fact is, the problem users tend to be the ones who don't care about reputation, meaning the only people who are really affected are those who may have legitimate reason for not accepting answers. –  Grace Note Jun 24 '10 at 14:14
    
@Grace, that is very true in every walk of life... and I didn't realize the +2 was in place for selecting an answer (I havn't asked any questions in a while) so thanks for the update. –  Matthew Whited Jun 24 '10 at 14:21
    
@Jared Harley It has been over an hour, and I received no notification, thanks for trying –  devinb Jun 24 '10 at 14:26

3 Answers 3

The community of all users, high and low reputation alike, already has a method for marking questions as proposed solutions in the form of upvotes. Pretty much by definition, a useful answer is one that can solve the problem. Yes, this does have problems with joke answers that get upvotes, like a certain ASCII ghost, but past that the general process is that the community upvotes indicate what the rest of the community thinks is a viable solution.

Ultimately, the "acceptance" marker indicates that a solution was successful for the author. The community does not have a voice in this, and never should. It doesn't matter how many other people were able to adapt a particular solution, if it doesn't work for the author then it should not be marked. The accept checkmark even indicates in the tooltip that it is what the author found to work. The lack of an acceptance mark may simply mean that the author is still deliberating on which works best.

This is why you have badges like Populist, where another answer gets far more votes than the accepted answer. This is the indication of what may be a more generally usable solution, without speaking (perhaps falsely) in the name of the original author.

There is also no way of knowing whether the author is still around or not. Nor should the author have to override the decision of the community as to what answer the author found most useful.

Finally, with regards to reputation limits, that is not very useful to implement. Reputation of a user has no impact on their knowledge of whether an answer is good or not. A low reputation user who knows the situation and may have encountered the problem in the past should not be less qualified than a high reputation user who has never seen the stuff before. That user should not be more qualified, either, because outside of the author's action we do not specifically value any one user's opinion greater than any other. So votes remain a very equalizing measure that's basically available to everyone.

share|improve this answer
    
Not enough space in the comment, so I answered in meta.stackexchange.com/questions/54832/… –  Peter Boughton Jun 24 '10 at 12:39

Similar features have been suggested before:

I'm going to paraphrase my previous answers.

Votes mean 'useful in general'

Accepted means 'useful to me'

Those two are completely and fundamentally different. By upvoting, the community is highlighting the answer they feel is the most generally useful. By accepting, the user is highlighting the answer which actually helped them the most.

Accepting is something which can only be done by the OP, because they are the only person who can definitively say what solved their issue. No amount of high-rep users can equal the weight of the one person who actually has the problem.

Also, it should be noted that upvotes in general already select which answer the community users "feels are correct." So having an additional checkbox would be rather redundant.

share|improve this answer
    
See my reply to Grace - I'm not interested in going against whether the OP says a question is accepted or not; I'm looking for a way to resolve questions that will never be accepted because the OP has left. (Hence the ability to override if the OP returns). Again, I'd be happy to mutate the question here into a different proposal if you've got other ideas how to identify questions still in need of attention vs abandoned questions. –  Peter Boughton Jun 24 '10 at 12:44
    
@Peter: As I mentioned in my response, the only person who has the ability to determine whether or not the question helped the OP is the OP themselves. No matter how long they have been gone for, they are still the only person who can tell us whether or not an answer was correct. If they have left and are never coming back, then we will have to go by the highest upvoted answer. But, no matter how well upvoted it is, it does not mean that it is correct. –  devinb Jun 24 '10 at 12:47
    
Sure, and that's why I wasn't proposing auto-accept based on upvotes - there are plenty of questions where inexperienced users will think they're upvoting a correct answer. That's why I was looking for a way to limit the action to "expert" users, but of course it's hard to identify these accurately. –  Peter Boughton Jun 24 '10 at 12:51
1  
@Peter An expert has no say on whether a question helped another user or not. Only that user has that say. –  Grace Note Jun 24 '10 at 12:54

Grace:

In reverse order...

  • I don't always agree with the way SO rep works, but that's how SO is, so we have to go with it.

  • There is a way of knowing whether the author is still around or not - all profiles have a "last seen" entry.

  • Eh? Badges aren't linked to questions, and a user having a badge doesn't solve this problem.

  • And as I said, if the author actively disagrees, they can override the decision - this is targetted at catching ex-users and drive-by users, rather than active users who just don't like any of the answers (I've got plenty of unaccepted questions because I haven't received a solution!)

  • As for votes being the controlling factor, that's partly true, but doesn't help with trying to find people that still need help - it's frustrating having to go through questions with X answers finding some that have 10+ upvotes, whilst others have just 1 or 2.

This FR is concerned with finding out which questions are truly unsolved, vs those that are clearly solved but simply not marked because the question was asked by someone that no longer cares. If there's another way to do that, feel free to make the suggestion and I'll modify the question.

share|improve this answer
2  
If you sort the answers by "votes" you'll have what the community thinks is the best answer right below the question - what more do you need? –  ChrisF Jun 24 '10 at 12:42
    
Regarding the 'last seen' entry, are you proposing that if the author hasn't been seen for months that people should vote differently? My point was, Stack Overflow operates on asynchronous communication. When the author actually gets around to a response is entirely independent from the community action, so speaking in her name falsely can result in false attributions for a long time, if not forever. –  Grace Note Jun 24 '10 at 12:45
    
Regarding badges, I'm not saying that badges solve the problem. What I'm saying is that the goal of badges is to promote positive behavior, and this identifies that the community finding one answer more useful than the accepted answer is positive behavior. And this doesn't mean that the author is wrong, it just means that the community has a different opinion. Which is perfectly acceptable. –  Grace Note Jun 24 '10 at 12:46
    
I want a way to identify questions that need to be looked at (because they haven't been answered to the satisfaction of the OP), and to ignore questions that have got "community accepted" answers, but the OP has lost interest. –  Peter Boughton Jun 24 '10 at 12:46
    
Regarding active disagreement, I think the fact that the author has not picked an accepted answer is their form of active disagreement. I should not have to re-de-accept answers because I didn't agree with it in the first place. If the community wants to try and change my mind, let me change my mind on my own decision, don't try to change my mind for me. –  Grace Note Jun 24 '10 at 12:48
    
Voting is sorted out by ChrisF's comment. The problem with this feature request is that while, yes, it can help identify questions that are unsolved versus unmarked, but it also opens up the system for people to change opinions that they have no right to change. An acceptance vote is really the same as a standard upvote or downvote - a measure of the user's opinion. No other user has a right to change that without the user's consent. –  Grace Note Jun 24 '10 at 12:50
    
Drive-by questions, if it has a good answer and it is properly upvoted, then when somebody needs to find that answer, they'll find it regardless of whether there is an accepted answer or not. Remember that ultimately, the goal of finding answers is to find information you might be looking for. I don't need someone else to confirm that it worked for them in order to find it with a basic search. –  Grace Note Jun 24 '10 at 12:52
    
Grace, this isn't about changing opinions. I want a way to filter questions that are awaiting a solution. I want to exclude questions that are probably solved but unaccepted. This is not about judgement/etc, I just want to be productive in my question answering. –  Peter Boughton Jun 24 '10 at 12:55
    
Peter, an acceptance mark is an expression of opinion. So it doesn't matter what your goal is, if you are changing that mark, you are changing that user's opinion. It doesn't matter if they can override it, the fact is that they should not have to. –  Grace Note Jun 24 '10 at 12:56
1  
Peter - searching for questions with hasaccepted:0 will give you all questions with no accepted answer. Unfortunately there's no way in the search to find questions with less than a particular number of answers. You could search the data dump though. –  ChrisF Jun 24 '10 at 12:58
    
This also doesn't take into account that some questions with accepted answers are still unsolved. I've stated it in the past, but my very first answer on Stack Overflow was to a question that already had an accepted answer, because the author accepted it because "he didn't have a better choice". I provided that better choice, and he seemed very happy with it. With that said, the metric based on accepted answers is only semi-accurate at best. –  Grace Note Jun 24 '10 at 13:00

You must log in to answer this question.

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