9
$\begingroup$

These two identical questions were asked by the same user (the first one was migrated from MO actually):

The second one has been closed as a duplicate. But now the first one also has two close votes for being a duplicate of the second one, and appears in review queues. What's the correct thing to do here? I checked "leave open", but I feel that this is something that shouldn't happen...

$\endgroup$
  • 3
    $\begingroup$ Loops should not be possible anymore, I expect that the last close vote would fail. If it doesn't and a loop is actually created, retag this as a bug. $\endgroup$ – Mad Scientist Nov 26 '13 at 20:00
  • 9
    $\begingroup$ I tried clicking "close" (currently, there are 2 close votes) and marking it as a duplicate, but I got a message This closure would result in the 'duplicate of' navigation only leading in a circle. So, I think there will be no problems here. $\endgroup$ – Vedran Šego Nov 26 '13 at 20:18
15
$\begingroup$

The loops are still possible. Look at this two questions (by the same author, but that's not really important; timing is the key):

  1. https://math.stackexchange.com/q/604492/78926

  2. https://math.stackexchange.com/q/601602/78926

Being one of those who voted, I can explain what happened:

  1. Both questions were voted as duplicates of each other. However 1 got closed before 2.

  2. I wanted to vote "close" on 2, but the "duplicate" reason was unavailable. It didn't matter, since I wanted to pick the usual "no effort" reason (I checked if the "duplicate" was available purely out of curiosity).

    The key thing is: it already had 3 "duplicate" reason votes.

  3. TZakrevskiy voted after me, presumably also as "no effort", so the final result was 5 closing votes:

    • 3 votes for "duplicate of 1" (which was already closed as "duplicate" of 2"); and
    • 2 votes for other reasons.

    Since 3 > 2, the published reason is "duplicate of 1".

Some of the possible solutions:

  1. Display all of the closing reasons (this would generally be a good idea), omitting "duplicate" in the latest-closed question if circular duplicity occurs.

  2. Detect loops and cancel "duplicate of X" votes if X gets closed as a duplicate of the question being observed.

Personally, I prefer solution 2, but I also think that the "display all close reasons" should be implemented regardless of this problem.

Addendum

I have just noticed that the author has deleted the question from the first link. So, the second one now has a notice that it was closed because it is a duplicate of the first one, with a link to the "This question was deleted by the author" page. I think that this kind of deletes should also be handled in some better way.

$\endgroup$

You must log in to answer this question.

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