32

Currently, once an edit by a user without full edit privilege is suggested, it can be either approved or rejected.

There is a known edge case that causes suggested edits to get auto rejected: (aka edit conflict)

  1. User with full edit privilege click "edit" and start typing. Let's name this user Joe.
  2. Since the user from previous step did not yet send the edit, users without full edit privilege can still click the "edit" as well, oblivious to the fact someone else is editing the same post in those very seconds. Such user click "edit" too and start typing. Let's name this user Kenny.
  3. Kenny is faster than Joe: he sends his edit first, creating new suggested edit. Joe is still typing.
  4. Joe is finally done typing. He sends his edit to its merry way oblivious to the consequences.
  5. As result of the full edit made by Joe, Kenny's suggested edit is kicked away and rejected. Boom.

This is all known and reported many times. Recent example where the user got it hard with at least two auto rejections in short period of time. Some concrete examples: one, two, three.

For edge case, it became pretty thick edge. This is not just minor hassle: enough rejections, and the user is banned from suggesting for a whole week: and remember we're talking here about good edits made by good users.
The above used to count for the ban, but no longer do. Thanks Gilles for requesting such a change!

I know it was already suggested to lock edits and similar solutions, but I suggest whole different approach: instead of auto-rejecting the suggested edit, dismiss it, i.e. cancel it like it never existed same way a bounty can be revoked by moderator.

In the edit suggestion page, it will show "Dismissed" instead of Approved or Rejected maybe with a tooltip explaining what caused this. This will soften the blow from those edge cases, taking the edge out of them. :)

12
  • 7
    It is possible that rejections by Community are not factored into the automatic ban... Do you have any proof that Community rejections count? Commented Mar 12, 2013 at 13:38
  • @MartijnPieters possible but not very likely.. if that's the case no need for what I suggest of course. :) Commented Mar 12, 2013 at 13:39
  • 2
    Why is it not likely? I know the devs of SE are all human, but don't dismiss the possibility they already thought of excluding auto-rejections by Community from the ban script. Commented Mar 12, 2013 at 13:50
  • As Yannis love to say... proof or it does not exist. ;) Commented Mar 12, 2013 at 14:37
  • I am eagerly waiting for some official word on this ... else I would soon find myself being banned :( Another example
    – Apurv
    Commented Mar 12, 2013 at 15:24
  • @Apurv for now you can simply wait five minutes before suggesting an edit, this is usually enough time for the other editors to do their part. Commented Mar 12, 2013 at 16:09
  • Here The edit was suggested 20mins after the question was asked. Now a 2k+ user edited the question kicking out the suggested edit. This edit process surely needs some change
    – Apurv
    Commented Mar 15, 2013 at 4:39
  • @Apurv agreed. Rejecting the suggested edit in such cases is wrong in my opinion, while can't have them approved either "dismissed" would be more suitable. Commented Mar 15, 2013 at 7:35
  • @Sha Wiz Dow Ard, the link in you Bounty announcement just gets me a bunch of waffle. Commented Aug 18, 2013 at 8:10
  • @Bill I know, I mentioned 10K only meaning it's deleted. Here is a screenshot though so <10K can see what I mean: i.sstatic.net/nswMu.png Commented Aug 18, 2013 at 8:27
  • @MartijnPieters Rejections due to concurrent edits used to count for the ban, but no longer do. Commented Aug 18, 2013 at 8:49
  • @Gilles thanks, didn't notice that! Prefer to have it embedded into the post though. This reduce the importance of my request, but would still be nice because of the other reasons I have. :) Commented Aug 18, 2013 at 9:22

3 Answers 3

6

This feature is now live starting July 1, 2024. You can see this staff post on Meta Stack Overflow which describes the change in its entirety.

We don’t have “dismissed”, but we now have “cancelled” for suggested edits that are canceled by the Community bot due to a conflicting subsequent edit.

We can also now see the number of edits that are canceled in the editor stats section within the suggested edit review queue:

screenshot of reviewer & editor stats on a completed review inside the suggested edit queue

6
  • While this appears to be true, I can't see how the screenshot is related. Where you got it from, and what does it have to do with the feature request, or change announced on MSO? It's a screenshot of some approved edit suggestion, without bot or being cancelled. Commented Jul 21 at 8:47
  • I want to say that we can see how many suggested edits are canceled in the editor stats. Commented Jul 21 at 8:49
  • 1
    This adds very little, and takes more than half of the answer space, plus the example isn't good as it's for editor with 0 canceled edits. Anyway, thanks for pointing out it's done. :) Commented Jul 21 at 9:35
  • @ShadowWizard Don’t know how to find a cancelled edit in this site. In Mathematics Stack Exchange, there are many canceled edits. Commented Jul 21 at 9:36
  • Why black out the names of the reviewers when they are visible in previous revisions? Commented Jul 22 at 16:52
  • @Tyma Mostly because the identity of the reviewers is completely irrelevant to the point of the screenshot, but I thought cutting the names out of the frame completely made it less understandable as a whole. I totally understand that it's a matter of preference, and that they're still visible in old revisions; my point was more about drawing attention rather than protecting identity per say.
    – zcoop98
    Commented Jul 22 at 20:51
5

Yes. I think it should be like this, too.

Here is the standard process that I would wish for:

  • If the reviewers choose to approve or improve edit, the edit should be approved.
  • If the suggested edit has been conflicted with another edit, the system should auto-mark it as dismissed.
  • If the reviewers choose to reject or reject and edit, it should be marked as rejected.
4

To add to User447378's answer, I would also like to propose the following:

Good (approved) edits:

  • Approve

  • Improve edit

Neutral (dismissed/invalidated) edits:

  • This edit conflicted with a subsequent edit.

  • This item is not/no longer reviewable. (Edit conflict due to rollback)

  • Post was locked - suggested edit cannot be processed. (Post locked - Example.)

Bad (rejected) edits:

  • Reject

  • Reject and edit

By using this method, it will make it more explicit that Reject and edit outcomes do count towards the ban, unlike the other auto-declined reasons.

3
  • 2
    Thanks, nice suggestion as well. Just curious, how did you find this 10 years old request? (You appear to be new to the network, so wonder how of all questions you landed here. :)) Commented Apr 19, 2023 at 10:04
  • 2
    @ShadowTheSpringWizard This user is not new to the network, and this specific issue is one of the things that led to their previous year long suspension, and has directly led to their new year long suspension.
    – Rory Alsop
    Commented Apr 20, 2023 at 21:21
  • @RoryAlsop sorry to hear, now this makes more sense. Can't see how such thing can lead to suspension, and sadly mods aren't allowed to discuss suspensions in depth, but all good, thanks for explaining what I asked. Commented Apr 21, 2023 at 5:14

You must log in to answer this question.

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