5
$\begingroup$

The community often brings old questions back up to the front page to allow for a fresh review by the membership of these questions. Recently on Cross Validation I have seen members suggest that an old question is a duplicate of a more recent question. If the two questions, call them A and B, are truly duplicates how can the older question A be called a duplicate of B? If anything B is a duplicate of A. I think this happens because the user is familiar B and is seeing A for the first time. So he calls A a duplicate of B without noticing the dates of the origination of the question.

I know that there have been many discussions here about duplicates but I don't think this issue has come up and I would be interested to here what other members think.

$\endgroup$
15
  • 4
    $\begingroup$ Glen_b's answer to a related question yesterday provides some reasons. $\endgroup$
    – whuber Mod
    May 18, 2017 at 13:01
  • $\begingroup$ Thanks, I wasn't on Meta yesterday, so I didn't see this. As I look at the question I see this more as a question of when to merge near duplicates and not exactly dealing with which is the duplicate when they are exact. $\endgroup$ May 18, 2017 at 13:02
  • 1
    $\begingroup$ In the examples I saw it was just stated that A appears to be a duplicate of B. The issue of which question was a better thread or needed merging did not seem to be at issue. I think my issue is that attention should be paid to the dates of the questions because if that is ignored the decision to close an old question may not be warranted. $\endgroup$ May 18, 2017 at 13:11
  • 1
    $\begingroup$ Have you reviewed the many related meta threads listed to the right of this page under the heading "Related"? I believe they address your concerns. $\endgroup$
    – whuber Mod
    May 18, 2017 at 13:36
  • 7
    $\begingroup$ Precedence in time is a factor (ceteris paribus) but it's not the primary criterion -- the aim is to arrive at a collection of good questions with good answers, and sometimes a newer question (or its answers) may be the better choice as the place to refer people - i.e. the one to treat as a canonical question. We can use closing as duplicate, merging, and even edits where needed to try to produce a good canonical question with good answers. $\endgroup$
    – Glen_b
    May 18, 2017 at 14:43
  • $\begingroup$ As I said at the outset that a lot has been written here about concerns about duplicates. Of the 12 related questions listed I even wrote one of my own in 2012. I understand the many issues involved and I agree with the reasons for eliminating or merging duplicates. I have only one major issue which regards overlooking which question came first. I have reviewed several of the 12 related questions but I do not see this issue mentioned in any of the titles. If someone could point me to one that addresses this issue specifically I would appreciate it. $\endgroup$ May 18, 2017 at 14:53
  • $\begingroup$ I also appreciate hearing any new opinions on the subject. I agree with the comment that Glen_b just gave above. I am not concerned about situations where the order is recognized. I am only concerned about closing a question because the order was overlooked. Perhaps this does happen very often and there are other remedies such as reopening a question or merging questions in an appropriate way. $\endgroup$ May 18, 2017 at 14:57
  • 4
    $\begingroup$ I don't think there is a Meta question specifically about this issue, but e.g. I wrote in this answer two years ago that "It [the canonical thread that other questions are duplicates of] does not have to be the oldest question", and that answer got +10 upvotes. Glen_b wrote the same in his yesterday's answer linked above. (I am sure this came up in some other Meta threads too.) So I guess that's the current consensus. $\endgroup$
    – amoeba
    May 18, 2017 at 15:16
  • $\begingroup$ @amoeba I would also agree about the choice of a "canonical question" although that can sometimes be very difficult to judge. I was just raising the possibility of not recognizing the time order. $\endgroup$ May 18, 2017 at 15:24
  • $\begingroup$ A discussion from GIS Meta SE (might help): gis.meta.stackexchange.com/questions/3477/…. $\endgroup$ May 19, 2017 at 15:52
  • $\begingroup$ That question is useful to the topic and your answer there seems to be the prevailing position here too. Yet all links mentioned thus far on this post do not address the issue of not recognizing the order of the questions when flagging it. I am beginning to realize that no one can really know for sure why the user flagged a particular question as a duplicate.since the newer question might be viewed as a better post. $\endgroup$ May 19, 2017 at 16:02
  • $\begingroup$ Michael, I am confused about what you are asking here. Your post contains one sentence with the question mark, and I quote: "If the two questions, call them A and B, are truly duplicates how can the older question A be called a duplicate of B?" The comments above explain pretty clearly, and you seem to agree, that in some situations older question A can be called a duplicate of a newer question B. So what is here to left to answer? $\endgroup$
    – amoeba
    May 24, 2017 at 14:45
  • $\begingroup$ The question asks why the first post could be considered a duplicate of the latter rather than the other way around. I specifically consider when the person posting doesn't recognize the time order. Many users argue that the weaker thread should be considered the duplicate regardless of the order and if they are nearly equivalent in terms of answers and comments they should be merged. $\endgroup$ May 24, 2017 at 15:57
  • $\begingroup$ Hi Michael! I noticed that Animuson's answer got 9 upvotes already, and seems in line with the discussion in the comments above. Is there any particular reason why you did not mark it as accepted? If it's intentional, perhaps you can edit to explain what exactly you are still waiting to be clarified; otherwise maybe accept to close this thread? $\endgroup$
    – amoeba
    Jun 21, 2017 at 10:45
  • $\begingroup$ @amoeba Please see my comment below the answer. $\endgroup$ Jun 21, 2017 at 13:59

1 Answer 1

13
$\begingroup$

Why can questions be closed as duplicates of older questions?

As others have mentioned in the comments, duplicates are not there to chastise users by yelling "someone else has already posted this!" They exist to direct users to where the information already exists on the site, which is why we discourage deleting duplicates which provide alternate routes to that information by leaving them as sign posts. Now, if you're going to use duplicates to direct users elsewhere, you want to make sure they are ending up at the best information, not just the earliest information, which is why we allow closing questions as duplicates of newer questions.

So what about the case where the asking dates were just overlooked?

Well, we can't possibly know whether or not that happened, so it frankly isn't worth considering. The dates the questions were asked just really do not matter at all. The only question you should be asking yourself is: "Which question would I rather end up on if I was looking for this information on Google?" Whichever one answers that for you is probably the one that should be the target. If that decision involves reversing the direction of the duplicate, you might bring it up on Meta so the community can evaluate whether or not the direction should be reversed.

What if both questions contain great information?

Evaluate whether or not the answers of both would make sense on one question. If the answers could easily be migrated either way with minimal effort, flag for moderator attention requesting they be merged to preserve the information of both questions. Merges usually go in the direction of the oldest question, but only for the sake of sanity (keeping the answer dates after the ask date - answer dates that occurred before the question was asked look weird). But that's also not always the case. Sometimes it makes sense to merge into the newer question. For example, if the newer one has 100,000 views versus the older's 100 it is obvious that the newer one is much more searchable and a better landing page for visitors.

$\endgroup$
3
  • $\begingroup$ I think you make some good points if your premise is that labeling a question a duplicate is intended to redirect the user to other questions that may provide the answer in a better way. But I know that duplicate is used as a criteria for closing a question. Aren't closed questions eventually removed unless they are voted to be reopened (which doesn't happen often)? $\endgroup$ Jun 21, 2017 at 13:58
  • 1
    $\begingroup$ @MichaelChernick Questions closed as duplicates are never automatically removed. Non-upvoted questions closed for any other reason are indeed automatically deleted; but duplicates are special. They stay. $\endgroup$
    – amoeba
    Jun 21, 2017 at 14:04
  • 1
    $\begingroup$ @amoeba Thank you for explaining that. I will accept the answer. I guess I should pay more attention to questions marked as on hold or duplicate. $\endgroup$ Jun 21, 2017 at 14:09

You must log in to answer this question.

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