I was a little suprised to see a Question of mine about a MathJax error showing up in Related Links (for another meta.Math.SE question), because as I remembered it, it had been closed as duplicate.

This error no longer happens to me with any frequency, and I recall reading a comment or two that explained the problem as network lag at the time. But the "original" question to which mine was a duplicate appears to have been "Removed for reasons of moderation" (perhaps because the User who asked it has been removed).

I point this out as an odd case, not especially to have it reopened, but to ask about our process (if any) when Questions linked as "originals" (to some closed as duplicates) are deleted. I'm tagging for discussion because I'm uncertain it rises to the level of "feature request", but if there's a Community consensus, we could ask about automatic reopening of cases like this.

Is there a Data Explorer query that can find such cases?

Related recent posts on meta.SO:

Report or automatically fix problematic duplicate patterns

Allow Trusted Users to modify Closed as Duplicate link

  • $\begingroup$ Why do you think the original was "removed for reasons of moderation"? (See my comment to Carl's answer below.) $\endgroup$ Commented Mar 18, 2013 at 12:31
  • $\begingroup$ In the mean time, I've re-opened your question. But are you sure that that question need to exist? Math.SE is now using Mathjax 2.1, and I doubt a bug about the integration with Mathjax 1.1 is still relevant. $\endgroup$ Commented Mar 18, 2013 at 12:35
  • 1
    $\begingroup$ @Willie Wong: it is not common for software development projects to delete old bug reports simply because they are no longer relevant. The bug report is tagged to mark it as outdated, unreproducible, etc., but it is kept for posterity. In our case, we have to use meta both as a discussion board and a bug tracker - but the bug reports should not be deleted, since then it is hard to see if there has been a regression or a recurring problem. $\endgroup$ Commented Mar 18, 2013 at 12:40
  • $\begingroup$ @WillieWong: "Removed for reasons of moderation" is what I recall seeing in place of the linked "original" (the one mine was closed as a duplicate of), next to the graphic slashed-backward-E (does not exist) for the missing question. $\endgroup$
    – hardmath
    Commented Mar 19, 2013 at 18:57
  • $\begingroup$ @WillieWong: As I pointed out(!) in the above post, it isn't so much a request for re-opening the MathJax bug Q as pointing out a process problem. I think the old question needs to exist (per Carl, to document a fixed bug, and also to illustrate the duplicate process problem), but it could be closed (perhaps as "too localized" rather than as "duplicate"). Thanks for your research reflected in the comments to Carl's Answer! $\endgroup$
    – hardmath
    Commented Mar 19, 2013 at 19:20

1 Answer 1


Normally, a question with an upvoted answer cannot be deleted. But in this case the person who answered the question (Jeff Atwood) deleted his own upvoted answer a year later, and then deleted the question some time later. This seems like a very strange situation - I think it would make more sense not to delete the original question.

Actually, it is even stranger than I thought. The question was deleted Jun 11 '12 at 15:05, but the answer was deleted Jun 11 '12 at 15:34 : http://meta.math.stackexchange.com/questions/2182/what-is-this-error-and-how-can-it-be-rectified . I was not aware that was even possible. My personal opinion is that moderators should not delete questions in this way.

  • $\begingroup$ Perhaps this is even a rarer case than I expected. I took a look at the Data Explorer mechanism, which amounts to writing something like a stored procedure in MSSQL Server syntax, something I should be able to do. However it's not obvious one can identify questions which were closed as duplicates, much less identify the "original" question that got linked to them. $\endgroup$
    – hardmath
    Commented Mar 17, 2013 at 23:01
  • 2
    $\begingroup$ @Carl: several points (a) Jeff Atwood is not a community moderator of this site (b) The question was not deleted by a moderator. It is deleted by the automatic clean-up engine for unanswered, negatively voted questions. (c) Since a short while ago (on the order of couple months), the SE software automatically checks for dangling links in case of human deletion of questions (it will pop up a notice to the effect of "deleting this will leave X which is marked as duplicate to this post orphaned, are you sure you want to proceed?")... $\endgroup$ Commented Mar 18, 2013 at 12:28
  • 2
    $\begingroup$ ...so things like this should no longer happen by human intervention. (d) As of fairly recently also, we can no longer close questions as duplicates of unanswered questions, this will help prevent the automatic software clean up from removing the "duplicate target". $\endgroup$ Commented Mar 18, 2013 at 12:30
  • $\begingroup$ @Willie Wong: correct me if I am wrong - wasn't the answer deleted after the question? $\endgroup$ Commented Mar 18, 2013 at 12:37
  • $\begingroup$ @Carl: no. It is the same known bug described here. When an answer is deleted and then the question gets deleted, sometimes the deletion timestamps for the answer gets whacked. $\endgroup$ Commented Mar 18, 2013 at 12:41
  • $\begingroup$ @Willie: thanks, that explains it. In this case I have voted to reopen the original question, since as a bug report I don't think it should have been deleted. $\endgroup$ Commented Mar 18, 2013 at 12:43

You must log in to answer this question.

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