14

I've just seen this in the edit queue:

https://stackoverflow.com/review/suggested-edits/3444674

If it's not the worst edit I've ever seen, it's certainly not far off. I was writing a custom rejection reason, and got the 'This edit has already been approved' message; so I poked around - it looks like it was approved by Community, because Johnsyweb went through and corrected it. He's got considerably more patience than I do.

But that means that the user earned himself rep for getting an edit approved. Is that by design?

Edit

It looks like this was caused by an edit clash, rather than a bad edit. I apologise for jumping to conclusions.

  • 15
    When someone in the review queue improves an edit, they have a choice to uncheck a checkbox that says the suggested edit was helpful. If they don't uncheck it, the suggested edit will give the rep. Frankly, for 2 rep, it makes little difference. – Oded Nov 26 '13 at 13:17
  • Ah, I'd forgotten about that checkbox - thanks for the reminder. I was a little more concerned that getting the rep is going to encourage the user to carry on adding extra edits of similar quality. – andrewsi Nov 26 '13 at 13:21
  • Much of the time, the fact that an edit brought a bad post into attention is enough for me to not uncheck the box. If the edit was completely out of whack, I will uncheck it. – Oded Nov 26 '13 at 13:22
  • 4
    I would definitely not have marked that as helpful. The removal of apostrophes and lower casing of words such as "I'm" and "I'll" is distinctly unhelpful. – Martin Smith Nov 26 '13 at 13:27
  • 5
    I actually suspect an edit clash here. A user who merely edited the code, when another user already corrected all the capitalization problems before it. So I'm not sure the editor here actually made this terrible suggested edit. – Bart Nov 26 '13 at 13:38
  • 2
    @MartinSmith what Bart said, 12 seconds ahead of me. :) – Shadow 10 Years Wizard Nov 26 '13 at 13:38
  • @Bart - is there any way of finding out if that is actually the case? – andrewsi Nov 26 '13 at 13:44
  • 2
    @andrewsi: Check the revision history. Revision 3 was submitted 2 minutes before the edit was suggested; which definiately hints towards an edit collision (e.g. both users clicked "Edit" from revision 2). – Matt Nov 26 '13 at 13:50
  • @Matt - Ah, thank you. That does look like it's actually the case. – andrewsi Nov 26 '13 at 14:05
5

I uncheck "useful" checkbox if I have to undo more than about 1/3 of original edit.

Like, if someone sees a totally unformatted list and then formats it as code instead of a list, it might be helpful - after all it indeed is a bit more readable now. But if he sees list formatted as a list and turns it into a code block, and fix some other problem, it's not a helpful thing as he destroyed as much as he improved.

If specific edit seems to revert some parts of the post, and older revisions is only slightly older, I try to take it into consideration and not count it as a part that I have to undo - as said in the comments above, it might be unintentional, race condition problem. On the other hand, if some of earlier corrections was overridden by later edit, it means editor omitted them even if he could and should fix them, so not giving him 2 rep is not a terrible mistake on reviewer's part.

In this specific situation both decisions, improving with and without "helpful" mark, can be found justified.

| improve this answer | |
5

I will reference the "terrible edit" part. Consider this flow of events:

  1. User A is posting a post. Time is 10:00 for sake of example and simplicity.
  2. User B with less than 2K rep click "edit" at 10:01, one minute after post was posted and start typing, fixing stuff happily.
  3. User C with more than 2K rep click edit at 10:02, being able to do that since there's still nothing in the queue.
  4. User C saves his own edits at 10:03, sending the fixes as new revision. There's still no pending edit to be kicked away since User B still did not submit.
  5. User B send his suggestion at 10:04, oblivious that it's using the old revision as the "base" for the edit thus overwriting whatever user C has made.

Such edits better just be rejected in my opinion, then 2K user can pick the good part of the edit (whatever the user was actually trying to fix) and make new edit. Trying to manually "reconstruct" the previous good revision is just time consuming and risky.

| improve this answer | |

You must log in to answer this question.

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