11

I have answered a question in SO. It's correct and I know for sure and it was the only answer that was upvoted(three times).

Out of nowhere, a user comes in (with a reputation of 1), edits my answer which completely changes my answer(into a wrong one) and someone accepted the edit.

Luckily I was there when it happened, so I changed it.

If I hadn't been there, the answer would have invited downvotes or more edits. My question is, shouldn't the one who answered be the one that should be approving the edit?

  • Stack exchange is community edited and that usually works very well, the problem here is the reviewers incorrectly approving the edit. If it was wrong then rolling back was the right thing to do – Richard Tingle Mar 21 '14 at 12:16
  • Do you get a notification when your post is edited? – Chris Mar 21 '14 at 12:20
  • 1
    @Chris I did get the notification, but my question is what if I wasn't there? Who would take responsibility for approving a wrong edit or downvotes that might occur? – whereDragonsDwell Mar 21 '14 at 12:26
  • 5
    Thank you for bringing this up. What happened to you is exactly why I have a strong bias against approving edits to code. – Louis Mar 21 '14 at 12:48
  • Can you provide us with a link to the answer? – Toon Krijthe Mar 21 '14 at 13:29
  • @AmarDuplantier - I understand, my question was more curiosity then disagreement. The other side of your method however, would be that the millions (?) of answers who's OP's no longer visit SO could no longer be edited at all - which is a massive disadvantage. I think the current system is much better, the OP is able to revert bad changes the next time they are online - 6-8 hours of bad answer is better than a forever of bad answer. The problem appears to be more with the review system, than the edit system. – Chris Mar 21 '14 at 13:42
  • @ToonKrijthe Here's the link to my answer : stackoverflow.com/questions/22554779/… – whereDragonsDwell Mar 22 '14 at 3:18
  • @Chris I agree. The reviewers should be more scrupulous when it comes to editing an answer that is upvoted or one that contains some code. – whereDragonsDwell Mar 22 '14 at 3:43
  • @MatthewLundberg I didn't think it required any explanation plus it looked self-explanatory to me and the 3 others who upvoted it. It's just the reviewer didn't care to look at things like the question or the other answers, the reputation of the one who posted the answer vs the one who suggested the edit(especially when it's on code) or the fact that the reason given by the user for the edit doesn't even make sense for the code in the question. – whereDragonsDwell Mar 22 '14 at 5:51
  • @MatthewLundberg I am not against constructive criticism. All I am saying is the user probably didn't have a look at the question or simply doesn't know enough C# to realize if (total != null) is unnecessary and senseless when total is a non-nullable decimal. – whereDragonsDwell Mar 22 '14 at 6:15
  • @MatthewLundberg That wouldn't count as an explanation for my code. That's the explanation for the suggested edit. Gods! – whereDragonsDwell Mar 22 '14 at 6:22
  • @MatthewLundberg Yeah, I refuse, because I find it self-explanatory. – whereDragonsDwell Mar 22 '14 at 6:36
  • @AmarDuplantier You don't understand the point of StackOverflow. Whether your answer is useful to the questioner is secondary to its usefulness to future visitors. And in fact, your answer really is not of a very high quality. I'm sorry that your ego can't accept it, but this is true. You do not explain your code at all, other than "Try this," which is not an explanation. – Matthew Lundberg Mar 22 '14 at 6:42
7

The person who makes a post has a binding vote on suggested edits. So if you are in time, you can stop a post from being accepted - or stop it from being rejected, for that matter.

If you are not in time, you can roll the edit back. But there is very little you can do about downvotes you got that way, short of hoping that the downvoters will return and undo their downvote.

What you have probably run into is the problem that many people accept suggested edits too quickly. This is a known issue on Stack Overflow, and we've had several discussions on Meta on how we should address this.

| improve this answer | |
  • 1
    I also had this problem on my answer stackoverflow.com/revisions/14378462/8 where Synoli suggested a change which broke the code, never tried it himself, and got a successful review from a bunch of other people who hadn't tried it either. This lead to a downvote when the code stopped working and was wrong for days until I logged in again. I understand that this is a community site, but it's my name and reputation below the code that no longer works. Perhaps we could have extra stringent checks for answers that already have a lot of upvotes (and so are more likely to be correct)? – Adam Iley Oct 7 '14 at 22:31
  • @kybernetikos You could create a feature-request to that effect. You'll have to think carefully about the conditions and the wording to get it through. To get it upvoted, it usually helps if you have a few examples of where the current doesn't work properly. – S.L. Barth - Reinstate Monica Oct 8 '14 at 7:25
  • @kybernetikos I notice that the edit in your example was by a user below 2K, so it's possible the reviewers were sloppy. I have long believed that better review audits can prevent a lot of this. – S.L. Barth - Reinstate Monica Oct 8 '14 at 7:29
  • Changes are either wording/grammar/english or something that requires deep knowledge of the area. I'm reluctant to act as a reviewer for changes that require deep knowledge. Perhaps the best improvement would be for changes that involve modifications to code blocks to require review by someone with high reputation earned on questions with tags relevant to to the answer being editted. – Adam Iley Oct 8 '14 at 22:19
  • @kybernetikos You could propose to make the ability to review code edits a badge-based privilege. After they created the gold-based "dupe-hammer", SE is looking into other badge-based privileges: meta.stackexchange.com/questions/240700/… . – S.L. Barth - Reinstate Monica Oct 9 '14 at 7:53

You must log in to answer this question.

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