6
$\begingroup$

The title is probably a little hard to grasp, so let me explain what I mean with an example. I'll make it a very specific example, which will ignore other closing criteria for the sake of simplicity.

  1. A question (Q1) gets asked: Why is potassium permanganate solution purple?
  2. This question goes unanswered for days, maybe weeks or even longer. Long enough to be off the radar for most of our users.
  3. Another question (Q2) gets asked: How is the permanganate ion responsible for the color of the solution?
    This is essentially the same question as Q1, which may or may have not been linked to it via comment. By the ordinary user it cannot be closed as a duplicate.1
  4. Suppose Q2 now attracts an answer2 (or more), because it wasn't possible to close the question (or maybe not even desired, see 1).

What should now happen to the two questions?

I see essentially two scenarios. The questions are still duplicates.

  1. We usually close questions as duplicates of other questions because they have been answered already. And most of the times the newer question will get closed in favour of the older one.
    Option: We can still close Q2 as duplicate of Q1 and carry on normally.

  2. We apply similar logic and link the unanswered duplicate to the answered one, which would be in the spirit of the redirect.
    Option: We close Q1 as duplicate of Q2.

There are probably other scenarios which I did not consider. If you can think of something more clever or simply want to add to the discussion I am looking forward to your answer.3


Notes: 1 Moderators can do it, so you can custom flag the question. Whether it is beneficial to close the question in the first place is another thing to discuss, but essentially it is possible. I only use this option if the same question gets asked in rapid succession by the same user. 2 Answer in the sense that the algorithm considers the question as answered, i.e. a score of 1 and above, or accepted. 3 An answer is much preferred over a comment.

$\endgroup$
6
$\begingroup$

Well, if you really want an answer... Closing as duplicate can be tricky, and perhaps this post is needed.

I think that while closing we shouldn't care much which one is older. It's rather quality that should be taken into account. If there's a new question with good answer and old, unanswered that's particularly dupey, than closing the old one is a very valid idea. Still, it's not like it has to be done. In theory it might bring even better answer, who knows? Still closing is probably the way to go.

If new question and answer are of poor quality then it's probably better not to do that. Not all posts are good enough to become duplicate targets. Perhaps such question should be even closed and eventually deleted. Or it could be closed as duplicate of old, even unanswered one... but even if old one is a perfect dupe target, with good answer it's quite possible that new one shouldn't be closed as dupe, because it would mean we keep it and sometimes it's better to get rid of it. Closing as dupe of unanswered should maybe be reserved for cases where great unanswered question exists, one which already is more informative then many answers.

Tell me if I should write some more or clarify my old user musings ;)

$\endgroup$
  • $\begingroup$ Yes, I really want an answer ;) And because it's tricky, I asked here. Thanks for your take, I think there is mostly consensus about the idea, so I guess this post all served its purpose. $\endgroup$ – Martin - マーチン Dec 22 '19 at 23:12
6
$\begingroup$

I don't think there seems to be much disagreement in the situation where one question has answers and the other doesn't. The obvious choice is to close the unanswered as a duplicate of the answered. The only thing I would add is that if the questions are exactly the same, then I would merge the poorer question into the better question, since the answers don't matter. There are some caveats about accepted answers (since merging removes green ticks), but merging happens so infrequently that those can really be dealt with on a case by case basis.

The bigger question is about what happens if both are unanswered. I think that there are valid arguments for several approaches. For example:

  • the worse question should be closed as a dupe of the better (makes people more willing to provide a good answer)
  • the older should be closed (the newer is around on the front page, so more likely to get visibility)
  • the newer should be closed (avoids stepping on people's toes)

I don't have an answer to that yet, but I lean towards Jan's suggestion of using question quality.

Of course, this only really applies to questions that are pretty much the same. If one is broader than the other, then the narrower one should obviously be closed.

$\endgroup$
  • $\begingroup$ Merging is an extreme measure, and that is probably stepping on people's toes more than closing the wrong one - and yes, that to be handled on a case by case basis (I'm not even sure how reversible the whole process would be). $\endgroup$ – Martin - マーチン Dec 22 '19 at 23:02
3
$\begingroup$

The idea of closing questions as duplicates exists to prevent fragmentation of answers, i.e. if two questions are functionally the same (but e.g. sufficiently different in wording that they do not show up when searching for the other) we/Stack Exchange want(s) to collect the answers in one place so that they may be directly compared.

In an ideal world, a question rapidly receives an answer so the newer question will always be a duplicate of the older one (where there are answers). We are in a real world, though.

If we are in the lucky situation that we have one (unanswered) question off everybody’s radar and one answered, newer question, the person who finds that old question should close it as a dupe of the newer question. This site doesn’t really function according to the ideas of academic or patent precedence where it concerns questions; only answers are assessed according to a higher standard where a later answer must offer an improvement or something different to an earlier answer of the same question.


Of course, the other two cases exist and I will add my opinion on those without them being mentioned in the OP:

If we have two questions without answers and we notice, in my opinion we should attempt to close the worse question as a dupe of the better question; however, I fully accept that that is not an objective standard to go by.

If we have two answered questions, we might consider any of the following:

  • better question has precedence
  • more highly upvoted question has precedence
  • higher number of answers has precedence
  • higher vote total or highest vote count of highest-voted answer has precedence
  • better answer has precedence (extremely murky)
  • merge the answers onto the better question if the content allows.
$\endgroup$
  • $\begingroup$ I am honestly not really concerned with both questions having no answers. This part will usually be figured out by quite ordinary handling of flags. At least from my point of view, I'll do what I think is right for that/these questions. Those decisions, if controversial can be discussed and reverted, so no harm will be done. $\endgroup$ – Martin - マーチン Dec 22 '19 at 23:07
1
$\begingroup$

As a minor update I decided to post the question that prompted the post:

As you can see, I decided to close the old one in favour of the new one.


I think those cases are especially rare, and it appears there is consensus that questions without answers should be directed towards questions with answers, but when there's doubt, we always know where to ask (here).

$\endgroup$

You must log in to answer this question.

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