6
$\begingroup$

Should questions ever be marked a duplicate of another duplicate question? Specifically, I'm looking at these three questions, each one marked as a duplicate of the next:

  1. ?+?+?=30 (1,3,5,7,9,11,13,15)
  2. What three odd integers have a sum of 30?
  3. What five odd integers have a sum of 30?

Shouldn't the first one just be marked directly as a duplicate to the third one? Or maybe a better solution in this specific case would be to delete the first question altogether? (Please do, because it's depressing that my answer to that question is one of my more upvoted answers.)

$\endgroup$
  • 3
    $\begingroup$ It is certainly allowed to do this. Sometimes a new Question is much more clearly a duplicate of a previously closed-as-duplicate post than of the "root" Question. $\endgroup$ – hardmath Nov 19 '17 at 22:25
  • 3
    $\begingroup$ I will gladly cast a delete vote on the first question, but I am confused. Why don't you just delete your answer if you find it depressing? $\endgroup$ – Asaf Karagila Nov 20 '17 at 6:33
  • 1
    $\begingroup$ @AsafKaragila yeah, I could have just deleted my own answer, but that felt wrong considering it had so many upvotes. And really, I think the fact that the question itself was still around was what was bumming me out. :P $\endgroup$ – Mike Pierce Nov 21 '17 at 17:08
  • 1
    $\begingroup$ I think a question should be marked as duplicate only if it actually is a duplicate. Meaning, the exact same question, with the same intent. Not just the same problem but the same fundamental question about the problem. Procedure is different than concept is a different than proof, etc. And even then, only when the question has been adequately answered in the first place. Also, proper titling is important. No one should be faulted for not finding questions that are obscure or difficult to find. $\endgroup$ – CogitoErgoCogitoSum Nov 25 '17 at 0:02
  • $\begingroup$ @hardmath - compound error. $\endgroup$ – Myridium Aug 7 '18 at 6:32
4
$\begingroup$

If you're searching for something and finds a question that was closed as a duplicate, I guess it would be nice if the link was to a question that actually has answers you could use.

That will probably most often mean a question that is not closed as a duplicate, but sometimes questions are answered before being marked s duplicates, and I'm pretty sure you can also find instances where (one of) the answers on the duplicate is better for some people than the answers on the question that was marked a duplicate of.

So it's not simple (and absolutely not something that can be automated) to determine if you should link to a question closed as a duplicate or the "original" for that (and the chain might be even longer).

And even though questions closed as duplicates have "[duplicate]" in the list of possible duplicates when you vote to close a question, it might be the only one that came up, and it's a lot more (relatively speaking) work if you have to walk of duplicates yourself, and as said for a questionable gain.

$\endgroup$
2
$\begingroup$

This is the response to the general question about duplicates of duplicates. (I.e., I am not commenting on the specific questions linked above as an example.)

I can see several situations where I might consider choosing as a duplicate target a question which already is closed as a duplicate:

  • One reason is to add a link to question with useful answers. For frequently asked questions, there are usually several copies floating around. If a question is closed with more than one duplicate target, this might help users to get links to several interesting answers - more of them than if only one duplicate link is given. (Admittedly, there is not much difference if the alternative link is only mentioned in comments - which adds the post among linked questions.)
  • If I think that Q1 is duplicate of Q2 but I disagree with the earlier closure of Q2 as a duplicate of Q3, then Q2 is the most logical duplicate target for me. (Of course, in such situations I should also try to reopen Q2 if possible.) Here I can include even more specific example. Personally, I would prefer not to close question about $\sum_{k=1}^\infty \frac1{2^k}$ as a duplicate $\sum_{k=1}^\infty x^k$, for the reasons I tried to outline in this answer. However, it seems that many other users have different opinion on this. Still, if there as a new question about this, it might be more useful for the asker if they are directed via duplicate link to the question about the special case $x=1/2$ rather than to the more general question - even in the case that the special case is closed as a duplicate of the general case. (Probably ideal situation is that the asker is given both links. However, if we assume that it is possible that the duplicate from special case to the general case might be cancelled in the future; it is better to have among duplicate targets at least one question which makes sense even after this - hypothetical - change.)
$\endgroup$
  • $\begingroup$ It looks like some of my concern, and some of the disagreement among users stems from there being several useful versions of the same question floating around. Do you know if there's ever been a discussion on either Math.Meta or Meta.SE about consolidating duplicate questions into a single parent question? $\endgroup$ – Mike Pierce Nov 23 '17 at 17:32
  • $\begingroup$ @MikePierce Such thing is done quite often and it is called merging. (If you want to see some specific examples, you can find a few in this chatroom. Here is one random example.) At least if this is what you meant by consolidating. $\endgroup$ – Martin Sleziak Nov 23 '17 at 17:38
  • $\begingroup$ However, merging two question is not always an option, even if they are duplicates. For example, if notation or formulation is different enough, answers from one question would not work on the other one. Or if merging would lead to duplication of answers, then it would be contraproductive. $\endgroup$ – Martin Sleziak Nov 23 '17 at 17:38
  • $\begingroup$ I don't agree that it would be contraproductive to have duplicate answers on a merged question, or even that a difference in notation/formulation should be a barrier to merging two questions. Some substantial editing may need to be done to the questions and answers though to make a cohesive post. Like, for the example you give of the two summation questions, why not word the merged question something like "How do you evaluate $\sum_{k=1}^\infty x^k$, ... in particular, what if $x=1/2$?" Then some of the merged answers would respond to the specific question, and others to the general question. $\endgroup$ – Mike Pierce Nov 23 '17 at 18:03
  • $\begingroup$ This could require heavy editing though. Is there some "limitation" (official or otherwise) on the extent to which users' questions and answers should be edited for something like this? (Apologies if this stream of questions have gotten off-topic) $\endgroup$ – Mike Pierce Nov 23 '17 at 18:05
  • 1
    $\begingroup$ @MikePierce Probably if you want to know more details about migration, that would be better for a separate question (or perhaps chat). I do not have much time at the moment and I am not sure I would be able to address all what you mention in the limited space in the comments. (But maybe somebody else notices your comments and will try to say more.) $\endgroup$ – Martin Sleziak Nov 23 '17 at 18:30

You must log in to answer this question.

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