4

When navigating the site, you can see linked and related questions at the right hand side:

enter image description here

The number refers to the net score of the question, whereas the color refers to whether an answer has been accepted or not (green if yes, gray if not). Yet, for those questions without accepted answers, it does not tell you whether the question has received any answer at all. This means that if you are looking for answers or if you are looking to answer questions without answers, the current system makes you visit questions which might not have answers or that might have answers, respectively. This is highly inefficient.

I think this system can be improved, by conveying more information (in particular, to indicate whether the question has answers or not).

Is it possible to add this differentiation? One option is to use a third colour (e.g. gray for those with at least one answer, white for those without answer). Another option is to add a second number in the box, like 3/2 or 3-2 or 3(2) (for a question with 3 upvotes and 2 answers). This can also be useful in questions with accepted answers. There are of course other options (e.g. add an asterisk to those without answers, 3*).

Benefits

  • Those who have knowledge on the topic of a question (which is presumably why they are looking at the question in the first place) can get an immediate idea of related questions which have zero answers, increasing the likelihood they might attempt to provide answers. Yes, you can also voluntary search for favourite tags with unanswered questions, but since the essence of the activity in SE is spent looking at questions, the rate of response could rise quite a lot. The same could be said of questions with just one answer, versus one with say three answers.

  • Those who are looking for answers to an issue might be saved considerable time by entering related/linked questions with answers only, instead of wasting time by selecting (unknowingly) unanswered questions, which they are very likely not going to answer anyway (as they are looking for answers).

5
  • I'm pretty sure the Linked posts entries are created once (when you link from your question or a comment to a questions) and are never updated later on (for example when you edit a link out or remove a comment). To have it fetch the actual answer count for each page refresh or have an extra cache that needs to be invalidated might impose both complexity and/or slower page load performance. Is that an acceptable trade-off for your feature?
    – rene
    Commented Jul 24, 2017 at 9:18
  • @rene Your comment goes beyond my understanding of how SE works, but to me it seems as an engineering problem, of which I am sure SE have faced many already. So more than a trade-off, it might be a matter of whether the benefits of it outweight the costs of implementing it.
    – luchonacho
    Commented Jul 25, 2017 at 7:21
  • It is not only the cost of implementing it, there is also a cost in performance which I think could be considerable but you're right, it is an engineering problem to be solved by the SE devs.
    – rene
    Commented Jul 25, 2017 at 8:52
  • @MichaelKjörling If it has not accepted answer, is it still a dupe?
    – luchonacho
    Commented Sep 6, 2017 at 13:35
  • @luchonacho The nuances probably warrant a question of their own (I'm sure it'd be a duplicate of some question -- oh, the irony) but yes. Duplicate basically means "any answers to that are also answers to this". The proposed duplicate has an upvoted, albeit not accepted, answer. I didn't see any significant difference between the two questions, hence I proposed that yours is a duplicate of the older one. If you feel differently, then I recommend that you edit this question to highlight how they are different.
    – user
    Commented Sep 6, 2017 at 14:25

0