I posted an answer at Does the following series converge?, then noticed it was wrong, deleted it, edited it, and undeleted it. (The system asks whether I want to "vote to undelete", which is a bit confusing, since I simply want to undelete and the answer does in fact get undeleted immediately without awaiting further "votes", but that's a different problem.)

After that, I noticed that the page is now saying "1 answer", though there are two answers, one that had been there before I'd posted mine, and my undeleted answer. I logged out to check whether this was maybe just an artefact of how it was presented to me as the author/deleter/undeleter of the answer, but after logging out and reloading the page it was still showing two answers but saying "1 answer".

  • $\begingroup$ The same happened to me here: math.stackexchange.com/questions/20946/… $\endgroup$ – t.b. Feb 9 '11 at 15:41
  • $\begingroup$ Do you also see an answer count of 2 in this question while there is only one answer and a deleted one? I take Jeff's answer below as an indication to just wait a bit and see if the count is corrected in a while. What do you think? $\endgroup$ – t.b. Aug 8 '11 at 17:18
  • $\begingroup$ @Theo: Yes, I see the same wrong answer count of 2. I'm still not sure how to interpret Jeff's answer (don't know why I wrote "I see"; perhaps I understood it at the time :-), but when read together with Jarrod's answer the next day, I thought they meant that this wouldn't happen once the fix had gone live and the numbers had "renormalized" themselves. Apparently that's not the case. I'll write a comment to Jarrod to see whether this is expected behaviour. $\endgroup$ – joriki Aug 8 '11 at 17:25
  • $\begingroup$ Thanks! I think I see the physics background transpiring in your writing "renormalization" instead of "denormalization" :) $\endgroup$ – t.b. Aug 8 '11 at 17:42
  • $\begingroup$ I decided to submit a new question: meta.math.stackexchange.com/q/2753 $\endgroup$ – t.b. Aug 9 '11 at 21:37

I introduced a bug in the undelete method; appropriate floggings have been issued.

A fix should be pushed out tonight.

  • $\begingroup$ Jarrod, this has happened again; see the link in Theo's comment under the question. Is this a new bug? Or just inevitable "renormalization"? :-) $\endgroup$ – joriki Aug 8 '11 at 17:27

Most counts are denormalized and will be auto-corrected every 24 hours.

  • $\begingroup$ I don't think I understand that answer :-) What does denormalized mean? And what do you mean by "most counts"? I think most counts I've been seeing other than this one have been correct. $\endgroup$ – joriki Feb 9 '11 at 16:00
  • 2
    $\begingroup$ @joriki: I think it just means that if you wait a while, the number will be fixed. What Jeff and the SE folks mean when they say "denormalized" is that for efficiency in running the software, certain numbers are not updated in real time, but (re-)computed every once in a while from the database. $\endgroup$ – Willie Wong Feb 9 '11 at 16:37
  • 2
    $\begingroup$ @jor this is still a bug, of course, it's just that the effects will not be visible to anyone in 0-24 hours from now with no action on anyone's part. $\endgroup$ – Jeff Atwood Feb 9 '11 at 17:36
  • $\begingroup$ I see, thanks for the explanation. $\endgroup$ – joriki Feb 9 '11 at 17:47

You must log in to answer this question.

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