In the spirit of Improving on 19,000 unanswered questions? and with our number of questions now over 93,000 I am wondering whether we would like to set ourselves a challenge of having 80% of our questions answered (i.e. <20% unanswered) when our question count ticks over 100,000?

Today we have 20,096 19,070 out of 90,859 100,000 (22.1% 19.1%) of our questions unanswered.

Unlike the Main site, discussion is encouraged here on Meta so "answers" to this "question" could include ideas about how questions may be moved from our unanswered list.

For the definition of what constitutes a question counted on the unanswered tab, this Meta SE FAQ should be consulted:

  • With 80,930 questions answered when we hit 100,000 questions asked, we well and truly succeeded in our challenge, and so I think it is time to set status-completed. Thank you to all who helped our site meet this challenge. – PolyGeo Jul 10 at 9:18
up vote 21 down vote accepted

Vote Early, Vote Often

  • Upvoting answers, when warranted, quickly moves questions off the Unanswered list if they have a net 0 vote.
  • Downvoting answers, when warranted, helps to push low quality questions towards the Roomba which can delete them from the site (and thus off the Unanswered list), if the community voting indicates that they have little value.

This encourages writers of good content to continue contributing, prevents visitors from seeing low quality content, and generally leads to a site with predominantly clear questions with clear answers, which is what we all look for when we have a problem.

A search expression that can be used to find questions with only answers that lack net upvotes (and thus appear as Unanswered) is answers:1 closed:0 isanswered:0 hasaccepted:0 (from Advanced search expression to find questions with answer(s) when no answer has net upvotes?). However, if you use this then please review the answer (and the question it is being applied to) before upvoting it. If it does not warrant an upvote then it may be more useful to the site to apply a downvote (if warranted) to it instead.

  • The problem is questions never marked as answered despite perfectly acceptable answers being given – Stevetech Dec 1 '17 at 18:25
  • 9
    @Stevetech I always think of the Accept checkmark being nothing more than indicating the answer which helped the asker the most (may be a lot but may be just a little) and gets rewarded by the asker. To me the votes say much more than the Accept checkmark about whether an answer is likely to be correct. – PolyGeo Dec 1 '17 at 23:38
  • 3
    @Stevetech To expand a little on PolyGeo's comment, the problem isn't questions that are never marked as answered, as once a Q&A has an upvoted answer with a positive score it will no longer show in the unanswered tab. – Midavalo Dec 2 '17 at 16:07
  • You miss the point. If answers don’t get marked as accepted the person answering has no idea if the op has even looked at it. This discourages posting answers to other questions, hence lots of unanswered questions – Stevetech Dec 2 '17 at 16:16
  • 1
    @PolyGeo It may be worth pasting a sample search string to find questions with unaccepted un-upvoted answers to filter from those with 0 answers (to aid in the upvoting). – Midavalo Dec 2 '17 at 16:17
  • 4
    @Stevetech I understand what you're saying, but I disagree - I could answer a question and get 7 upvotes from the community but no accept from the asker. To me this signifies the community accepts my answer and makes it worth answering other questions. Even if I get one upvote that means somebody likes my answer. All the accept mark indicates is that the answer worked for the asker, nothing more. I put more weight on the community vote. – Midavalo Dec 2 '17 at 16:19
  • Yes but you have 23000 rep. Someone with a smaller score might not be so blase – Stevetech Dec 2 '17 at 16:20
  • 5
    @Stevetech an accept = 15, upvotes = 10 each. A couple of upvotes quickly pass the accept reputation – Midavalo Dec 2 '17 at 16:23
  • You obviously miss the point so let’s leave it there – Stevetech Dec 2 '17 at 16:24
  • 3
    My point is, upvoting 0-score answers in the "unanswered" list still gives reputation to the answerer, as well as removing them from the unanswered list – Midavalo Dec 2 '17 at 16:24
  • 1
    @Midavalo I do not know such a search expression so I am asking at Meta SE: meta.stackexchange.com/questions/303962/… – PolyGeo Dec 2 '17 at 22:33
  • 1
    @Midavalo I got an answer within 5 mins (on a Sunday). – PolyGeo Dec 2 '17 at 22:43
  • 2
    My experience has been that the unanswered questions are either really obscure or really specific to a single user, and therefore, not of general use to other users. I have tried answering what looked to me like interesting and difficult questions and got zero upvotes for my efforts, which tends to put you of trying. It isn't all about rep, but you at least want some recognition for a research effort. – John Powell aka Barça Dec 7 '17 at 21:54
  • 2
    @JohnPowellakaBarça It may seem harsh but if an unanswered question is "really obscure or really specific to a single user, and therefore, not of general use to other users" then assessing it to be not useful and downvoting it rather than expending effort that you could apply to a more useful question is worth considering. – PolyGeo Dec 7 '17 at 22:51
  • 1
    @RobertoRibeiro The full criteria for post deletion are at meta.stackexchange.com/questions/5221 but in this instance I am defining low quality questions as those with 0 or negative votes. Sometimes they are only kept on the site due to upvoted answers. If those are poor answers then downvoting them pushes the Q&A towards deletion. Any question with a net upvote is safe from system deletion. – PolyGeo Jan 19 at 0:19

I think we need to tackle this on three fronts.

  1. A question without an answer which has a positive score: One aim could be to look for your specific tag you like to answer (for me its qgis and pyqgis) and search for questions in the unanswered section which have answers and if they are good to upvote them.
  2. Questions solved by a comment: In those cases you can clearly see that the OP could solve her/his problem with the help of a comment I advocate the solution, that you post the comment as answer.
  3. Questions which are good but don't have an answer: simply answer them if you can.

Option 1 and 2 are for me the easiest to persuade but I don't know how many cases are fall into those categories

  • I think encouraging users to even upvote (be it question or answer) is going to be difficult as there doesn't seem to be a large number of frequent voters. – Joseph Nov 30 '17 at 14:00
  • @Joseph I tend to agree with you when I have a look at the front page and see the upvote distribution on questions but this is just a gut feeling – LaughU Nov 30 '17 at 14:39
  • 2
    @Joseph It usually only takes one vote to give an answer a positive score, so there's not a huge need to encourage a lot of voters - if you find one you can upvote – Midavalo Nov 30 '17 at 15:02
  • 6
    @LaughU Regarding #2 I wouldn't worry about "stealing" reputation. You can either encourage the commenter to answer, or you answer it yourself. The commenter doesn't lose any reputation if you don't make it CW, but if you make it CW you then miss out on the chance of some reputation after taking the time to actually turn the comment into an answer which the commenter didn't bother doing to start with. – Midavalo Nov 30 '17 at 15:05
  • 4
    I agree with @Midavalo that "guilting" the "answers from comments (#2)" to become Community Wiki is likely to be counter-productive because you are removing an incentive which is available to everyone to try and improve our comment/answer conversion ratio. Besides it is rarely just a copy/paste because the comments need to be assessed to see which have the content to make them suitable as answers before posting them as such. – PolyGeo Dec 1 '17 at 1:10
  • 1
    I am also against advising making an answer CW. It distorts what CW is for. Comments are second-class citizens; they should go with time. I’d instead upvote easily a valid/useful answer converted from comments even if it is a copy/paste. – Andre Silva Dec 1 '17 at 1:44
  • 1
    @Midavalo,@PolyGeo the arguments against CW answers are making sense to me. My approach was more from the perspective of the commenter rather than from the person who posts this comment as an answer. – LaughU Dec 1 '17 at 7:35
  • 1
    Don't forget that you are free to revise your answer at any time. It may be that one part of it currently prevents upvotes or risks downvotes. – PolyGeo Dec 1 '17 at 20:14
  • 2
    I passed a few comments into answers, but usually when it was obvious the solution was listed. I usually say the answer comes from the comments to leave credits. My aim is to try and help to improve the rate of answered questions. My work was sometimes rewarded by some rep points, but I don't feel guilty to help out passing answers from comment to a real answer. After all, the person commenting could have took time to answer. That doesn't mean one can't try to improve it when copy n' pasting :) PS : dunno how to do a communty wiki yet... – gisnside Dec 1 '17 at 20:17
  • 3
    @gisnside exactly - we have some users providing excellent content in temporary comments and I think you are doing the community a favour by publishing that content in an answer. If the original commenter (or anyone else) later uses the same content to write a better answer then that is the SE model at work. – PolyGeo Dec 1 '17 at 21:56
  • 1
    @gisnside I also give credit to the user who wrote the comment (using their name and linking to the comment if appropriate) – Midavalo Dec 2 '17 at 16:09
  • 1
    @LaughU The commenter is still free to post an answer, even if you have already done so based on their comment. In fact they are free to post an answer instead of commenting to start with, but for whatever reason they chose not to. I would not worry about how the commenter feels if you post their comment as an answer, as they have already had (and still have) that option. – Midavalo Dec 2 '17 at 16:10

Is there something to be said about questions that are older than say 5 years that is a comment on their relevancy?

If someone posts a question and it isn't answered for even 1 year, or even several years, to me it means that question is either not relevant, belongs on another site/forum, or doesn't have an answer.

In any event, either the asker has moved on, solved their own problem and not posted the answer, or didn't require an answer in order to complete their work.

I come across questions that are so obscure or require a workaround so great that they don't get answered. Should those stay on this Q&A site as open questions?

  • 2
    If I think such questions may have any value then I edit or close them to try to make them more likely to be answered. I encounter much resistance to such editing and closing from some within our community. If I think questions have no value I downvote them to try and push them into the path of the Roomba (automated deletion). – PolyGeo Mar 1 at 21:53
  • @PolyGeo well thanks for accepting my answer! Do you think there shoudl be any action taken to close questions older than a certain date in order to clean things up? – DPSSpatial Mar 2 at 16:29
  • I understand your point and agree with it to some extent, but do you think, for example, the most scored unanswered questions should be removed? I hypothesize eventually, some of these get to be answered? Perhaps adding some examples of questions to your answer would help making a point. I don't think interesting and valid questions which no one knows how to answer show be removed or closed. – Andre Silva Mar 2 at 17:02
  • 3
    @DPSSpatial to me the value of a question should always be about its quality rather than its age so reviewing old questions is a volunteered activity that I think we should value. If questions are clear and appear to still be useful I upvote to help keep them, if they are not useful I downvote them to help the Roomba delete them, if they appear useful but could be clearer I try to improve them by editing or closing. I accepted your answer because I think it makes the point that having unanswered questions that hang around unimproved for years is something that we should try to avoid. – PolyGeo Mar 2 at 20:10

GIS.SE is used by some software organizations/collectives as a sort of help-desk (I've nothing against this, it's great...), but I note there is a tendency for an answer to be given by someone on the firm, and that answer then not up voted. This is normal behaviour in a help-desk environment ~ an enquiry comes in, one person picks it up, and responds and no other help-desker gets involved.

For example this appears to be the case with CartoDB such as:

and similarly for MapInfo:

I was wondering whether there was any way to encourage some sort of cross checking (and upvoting) in such situations?

  1. One thing that clears the path to have more answers, and consequently, more options to upvote for is to carefully edit the entire post (even if this 'complete' edit depends on getting info from OP through commenting). Unfortunately, it is not uncommon in our community to observe posts only partially edited (under the argument that any edit is a valid edit, which it is, but it costs time of visibility from new answers in the active tab).

  2. Another powerful action is taking the time to curate canonical posts (Q/A) and link/reference duplicates.

I have seen good outcomes after trying to take care of above issues considering specific tags:

Case 1:

One example was the experience 'moderating' the tag (Are kriging questions getting enough attention?). A considerable time has passed since I tried to clean the tag but the unanswered rate has hold to <30% since then (it was previously >40%).

Case 2:

The other example is , which I hypothesize that after I helped curating it more closely, it increased activity from other responders as well. For example, parameterizing the SEDE query pointed in PolyGeo's answer here for the 108 most populous tags, resulted in this tag having the least unanswered rate (6.12%) after .


Of course, those tags I have cited have a small number of questions with them, but I think the idea can be extrapolated to larger tags as well.

  • I commend the idea of trying to perform large edits to improve questions and to try and curate canonical Q&As. However, with improvement to posts relying entirely on volunteered efforts I think the value of multiple nudges of questions towards a more answerable state by modest edits, and/or towards potential answerers/improvers by better tagging should not be underestimated/undervalued. I think it is usually better to do something with an available slice of time rather than holding off on doing anything while trying to find the time to do everything to a post. – PolyGeo Dec 16 '17 at 2:31
  • I just tried running that SEDE query again and even though lidar has 506 (to meet the minQuestionCount) it does not seem to be showing up. From gis.stackexchange.com/tags/lidar/topusers I think it should have still appeared in 2nd place with 7.1%. I can't explain why it is missing but I'm no SEDE/SQL expert. – PolyGeo Apr 22 at 1:42
  • @PolyGeo, it is because it is not a top50 tag in number of questions. That query has two variables: i) minimum number of questions; ii) maximum number of tags (the ones with most question count). When I ran that query I set it with 108 top tags. – Andre Silva Apr 22 at 1:59
  • 1
    That gave me the nudge I needed to use fork query for the first time and run it on the 200 top tags. It placed lidar third after field-calculator and cursor. – PolyGeo Apr 22 at 2:08

You must log in to answer this question.

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