The old question

Show $1 + 2 \sum_{n=1}^N \cos n x = \frac{ \sin (N + 1/2) x }{\sin \frac{x}{2}}$ for $x \neq 0$

(which has an accepted answer) was edited by a low-rep (<10) user in a way that completely changes its intention. Why was this edit approved? If the editing user wants an answer using a different method, they should ask their own question, not hijack someone else's.

  • 12
    $\begingroup$ Thanks for the heads up. The edit has been rolled back. Flagging the post might be a better way to handle this. $\endgroup$ – robjohn Feb 20 '14 at 17:03
  • 3
    $\begingroup$ Dear @robjohn, I did flag the post too. I just wanted to raise it here so that the approvers could explain why they did it. $\endgroup$ – user64687 Feb 20 '14 at 17:04
  • $\begingroup$ I see that now. Thanks :-) $\endgroup$ – robjohn Feb 20 '14 at 17:07
  • 25
    $\begingroup$ One irritating thing about the approval of edits process is that unless one is awfully quick, an attempt to reject is met with the announcement that the edit has already been approved. Attempts to improve usually meet the same fate. $\endgroup$ – André Nicolas Feb 20 '14 at 18:14
  • 24
    $\begingroup$ I agree with @Andre. It would be great if we can somehow prevent this race condition. (For example if by opening the review you have it assigned to you for some several minutes, unless you press cancel which returns it to the queue.) $\endgroup$ – Asaf Karagila Feb 20 '14 at 19:32
  • 2
    $\begingroup$ Here is link to the suggested edit mentioned in question. $\endgroup$ – Martin Sleziak Feb 21 '14 at 7:36
  • $\begingroup$ "I just wanted to raise it here so that the approvers could explain why they did it." Do you have any reason to expect the approvers to read this meta site? And if they do, what exactly do you expect them to give as an explanation? $\endgroup$ – Gerry Myerson Feb 21 '14 at 8:33
  • 8
    $\begingroup$ @GerryMyerson: one of the approvers is a high-rep user who (according to his "about me" text) is quite involved in "administrative" tasks on the site. So I imagine there is a good chance he is also a regular visitor to meta. As for explanation: I don't expect anything, but if someone thinks there is a good reason that I am completely missing to approve this edit, I would be glad to hear it. $\endgroup$ – user64687 Feb 21 '14 at 8:52
  • $\begingroup$ @AsafKaragila Hopefully the answer isn't "fight robo approval with robo rejection" :) $\endgroup$ – rschwieb Feb 21 '14 at 11:16
  • 4
    $\begingroup$ @rschwieb: No, I meant that once you open the review dialog the system recognizes that someone is reviewing the edit. So if you don't press "cancel" you have like ten minutes to approve, reject or improve (or cancel) and until that time the system prevents further users from deciding the fate of that review. This will invariably let people read the edits and examine them with leisure. $\endgroup$ – Asaf Karagila Feb 21 '14 at 16:26
  • 1
    $\begingroup$ @AsafKaragila Apparently my off-the-cuff jokes are so unfunny people can find a way to say "no" to them :( I really didn't mean it to address anything about the fix you had in mind... the idea of robo-rejection just amused me. $\endgroup$ – rschwieb Feb 21 '14 at 16:49
  • $\begingroup$ @rschwieb: Robots fighting robots to the death? Bender will not be amused. $\endgroup$ – Asaf Karagila Feb 22 '14 at 8:26
  • $\begingroup$ @Asaf it looks this is what happens see meta stackoverflow but not for edits. I wonder why not? (PS I remember being asked to review that edit, I think I tried to reject or skipped on the basis I didn't have the expertise.) $\endgroup$ – TooTone Feb 22 '14 at 15:11
  • $\begingroup$ It happened again here: math.stackexchange.com/q/688720/8157 . This approved edit changes formatting in such a way that it does not render. It should have been rejected. $\endgroup$ – Giuseppe Negro Feb 24 '14 at 17:05

Why was this edit approved?

The reason is more trivial than what we could expect. I wanted to reject, but once we click on "Approve", it seems we can't go back (the opposite problem may not happen because we have to click further on the reason). So it's the last bullet in azimut's answer.

In order to avoid this problem, maybe we could create an option. Each user chooses whether he/she want to have to confirm when he/she approves an edit.

  • 5
    $\begingroup$ Dear Davide, thanks for taking the time to explain. $\endgroup$ – user64687 Feb 21 '14 at 14:51
  • $\begingroup$ You are welcome. $\endgroup$ – Davide Giraudo Feb 21 '14 at 15:55
  • 1
    $\begingroup$ I don't understand why Davide was downvoted, when he was kind enough to explain that the approval happened through a simple slip of the mouse. $\endgroup$ – user64687 Feb 21 '14 at 15:57
  • $\begingroup$ @AsalBeagDubh Probably the downvoter disagrees with my proposal in the last sentences. $\endgroup$ – Davide Giraudo Feb 21 '14 at 16:53
  • 1
    $\begingroup$ Hmm, that sounds like something that could be doable with a user script. Let me look into it... $\endgroup$ – Ilmari Karonen Feb 26 '14 at 12:49
  • 5

I think many unjustified edit approves are provoked by the following causal chain:

  • You can earn badges for reviewed edits (Custodian, Proofreader, Reviewer, Steward), making many users keen.

  • To achieve a valid edit review, your review must be finished before (typically) two others do. So you have to be fast.

  • All users having enough reputation get informed in the status line about waiting edit reviews. So the potential competition is huge and you have to be very fast.

  • Blindly accepting an edit is much faster than actually thinking about it.

  • $\begingroup$ Dear azimut, I agree with your summary. Sadly, it risks making the edit approval process more or less a waste of time. $\endgroup$ – user64687 Feb 21 '14 at 14:49
  • 3
    $\begingroup$ @AsalBeagDubh: Yeah, it's definitely not optimal. I think the situation would be much less fierce if the notification is removed. Also, it doesn't make too much sense that this notification mechanism exists only for the the edit review queue, but not for the other 5 review queues. $\endgroup$ – azimut Feb 21 '14 at 15:00

You must log in to answer this question.