8

This site encourages users to try to answer old unanswered questions. I've made several attempts to wade into the 35K+ unanswered questions, but it is like finding a needle in a haystack. From a review of old meta posts, various aspects of this problem have been hashed and rehashed for years.

The vast majority of these questions are, realistically, ones that are simply not good candidates to ever get answered. Their presence makes it nearly impossible to find the few questions that are. The bloat includes questions that will eventually be removed automatically via rules, but are obvious to a human will not benefit from the wait. Many received some upvotes and/or some initial views, so they have languished without triggering the rules. For many, the comments provided the answer so activity just stopped. Many of the questions are related to obscure hardware or user-specific issues that have long since been overtaken by events and likely of interest to few other people. There are a number of other similar characteristics that are obvious to anyone who has looked for old questions that can be put to bed.

As it stands, very few of the old questions that might get answered will get answered simply because it is too hard to find them. If the site is serious about getting old questions answered, there needs to be a way to separate the wheat from the chaff, to enrich the fishing ground so good candidates are more accessible. One approach is to start with the pool of all old, unanswered questions and peel off layers of poor candidates. The other is to search the pool for good candidates and promote them to an "answer me" pot.

One argument against doing this is that questions in the "poor candidate" pool will never be answered. The truth is that they are unlikely to be answered anyway, and not doing this means that the good candidates won't get answered either.

10
  • 2
    Old questions with low score and a low view count will automatically be deleted, in case you didn't know. I can't give you any number though.
    – slhck
    Oct 4, 2014 at 9:59
  • 1
    Right. The unanswered question pool reflects this policy. I'm not talking about changing the policy, I'm referring to a method to more easily find good candidates in a big sea of questions, most of which will never get answered.
    – fixer1234
    Oct 4, 2014 at 14:31
  • Well, I'd start with the unanswered list, which is sorted by votes. If there's anything that is abandoned, already answered, or outdated, we need people to flag it so we can delete it. This can't be automated…
    – slhck
    Oct 4, 2014 at 14:33
  • Been there, done that. That's what prompted the post. Virtually all of the questions are there because they are within the policy for not being deleted yet and I'm not talking about deleting any or changing the policy. I'm talking about creating a pool of questions within the unanswered questions that have been screened by more rigorous guidelines so that a much higher percentage in that pool are good candidates. This can be done by automated means to exclude the least likely questions, via readers tagging questions for exclusion, or via readers tagging for inclusion.
    – fixer1234
    Oct 4, 2014 at 14:56
  • 1
    Just focus on your preferred tags and wade through them that way? I often browse unanswered excel questions, I remember which ones have no registered author and which ones aren't answerable, so my list stays limited. Oct 4, 2014 at 23:42
  • I came across a number of posts like that. It seems to have come up every few years. In the two years since that post, the count has gone from 13,000 to 35,000. It would be useful even to have user controllable filters to get the pool of questions to wade through to a more manageable size.
    – fixer1234
    Oct 6, 2014 at 0:26
  • Is a significant portion of these questions effectively duplicates of good, answered questions? At least on SO that seems to be a factor -- improving duplicate handling one way or another would address that, but I'm not sure if it fits here.
    – tripleee
    Oct 10, 2014 at 7:04
  • I will say that it does happen - I just answered this question after tag-diving, which sat unanswered for 5 years
    – Robotnik
    Oct 13, 2014 at 23:10
  • 1
    And you saw the comment that he wasn't using it anymore?
    – fixer1234
    Oct 13, 2014 at 23:19
  • 2
    @fixer1234 - yeah? That shouldn't matter, the answer might still be useful to other people.
    – Robotnik
    Oct 14, 2014 at 3:33

3 Answers 3

11

If you go through the list of unanswered questions and you see a question that you would consider a "bad question" because you think that it can't be answered, downvote it.

If a question can't be answered, it's not helpful. By downvoting it, you make it clear to other users that it's not worth looking at.

Likewise, if you see a question that you would consider good and answerable, upvote it, to make it clear that it's a good question and worth looking at.

Beyond that, I really don't see what else should be done about the situation at hand. It's not like we're going to form a committee to go through all unanswered questions and promote or demote them into pools. A committee like that exists, it's the community and voting is the process that takes care of this problem.

2
  • 3
    Voting up or down one question at a time would make no difference. Nobody is going to read through thousands of old, bad questions, either to find one that is answerable or to vote on each one. Questions come in so fast that if one doesn't get answered in the first few days and doesn't have a bounty it becomes lost, and then votes don't make much difference. It is just assumed that if it was answerable, it would have been answered when it was fresh. I'm not suggesting committees. I'll think about potential solutions and put them in an answer.
    – fixer1234
    Oct 6, 2014 at 10:40
  • 1
    @fixer1234 not necessarily, sometimes an old but good question is difficult and can bring a non-user in just to answer it Oct 10, 2014 at 9:08
3

In response to Oliver Salzburg's answer, I said I would think about possible solutions and post them. I want to close the loop.

I dug into the unanswered questions, did some analysis, and came up with several pages of absolutely brilliant insights and suggestions. Then I learned about using the advanced search features. In conclusion, nevermind.

It turns out that the advanced search features are good enough to accomplish the task. To identify "likely" good questions, I used this set of criteria: views:50 score:2 answers:0 duplicate:0 closed:0. This found about 2,560 candidate questions (changing views to 30 added another 200). This is a manageable number to work through. Virtually all of these are "quality" questions. Surprisingly, the old questions in this group aren't substantially different from the newer ones in terms of being obsolete.

The majority are still not answerable. There are a host of reasons for this. Common ones: the question is already answered in the comments, or the question was about a system hiccup. A wild guess--maybe 10% of the questions in this group are answerable by somebody. So out of this slice of the unanswered questions there are probably 250-300 that could be answered. Adding tags to the search criteria would reduce the review task to a fast list.

3
  • If the question is answered in the comments, or the OP has edited their question with the solution, you should first leave a comment, asking them to put it as an answer, and (if after a couple of days they don't) - Add it as an answer yourself with the 'Community Wiki' checkbox ticked (and edit it out of the question).
    – Robotnik
    Oct 15, 2014 at 6:29
  • I used the term "answered" loosely. Mostly, they are questions, hints, or collections of suggestions that pointed the OP in the right direction and they subsequently implied that they had solved the problem. There isn't enough there to convert it to an answer. Often, the OP is long gone.
    – fixer1234
    Oct 15, 2014 at 7:04
  • 2
    I would still recommend turning the comments into at least a CW answer in that case. If someone comes along needing an answer to that question, they at least have it all collected in a single place rather than in a jumble of comments. If someone familiar with the subject eventually comes along, they can take that CW answer and expand on it turning it into a proper answer, either by editing or by posting a new answer. IMO that's a win for everyone involved. Remember that comments are second-class citizens on Stack Exchange and can be deleted for almost any reason.
    – user
    Oct 15, 2014 at 14:14
0

A short, & possibly useless point...

So many times, the 'correct' answer is posted, yet the OP never checks the box to accredit it.

Can an admin 'push' that answer in any way, so it will show as 'answered'?

3
  • 2
    Only the OP can decide whether that is the answer that satisfies them, which is the meaning of the checkmark (not that it is the right answer). A different answer might be the right one for someone else. If the OP leaves a comment like, "Thanks, that's the exact answer I was looking for and it solved my problem", and then doesn't checkmark it and disappears from the site, I could see an argument for a mechanism to award the credit. If the answer is just in the comments, somebody would need to convert it to an answer. Then I suppose the same accreditation logic could apply.
    – fixer1234
    Oct 17, 2014 at 18:43
  • I appreciate your comment, though perhaps my line of reasoning wasn't clear [even to myself] - picking up points for a good answer wasn't my motivation; clearing of old 'useless' Q/A & promoting of 'good' Q/A was my intention.
    – Tetsujin
    Oct 17, 2014 at 18:48
  • I agree that getting an answer that clearly solved the problem checkmarked provides valuable information for others. If I see an old question with answers but no acceptance, I wonder if perhaps the answers weren't good and the question could be put to bed with a better one. If I see an accepted answer, it is clear that the existing answer actually solved the problem.
    – fixer1234
    Oct 17, 2014 at 19:01

You must log in to answer this question.

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