What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 133 Stack Exchange communities.

Scenario

A new user posts a question quite innocently (but obviously without looking at the preview). It may be a fine question, but because of the user's use of < in the code sections (which of course are not formatted as code), most of the question is not visible.

Because the question is in a popular tag, perhaps ten people with edit rights read it within the first minute. Five of them immediately edit it. One of them (perhaps myself) goes ahead and cleans up the text part of the question also, and still finishes editing first.

Problem

So now the question is cleaned up (except perhaps the title and tags), but four more people are still editing it. Since their edits are pretty quick to make, the simultaneous-edit warning does not kick in (not that it helps much anyway), and the question is turned to wiki within three minutes of being posted.

Besides being unfair to the new user (which may or may not be a problem, since as I mentioned the user could have prevented this by merely looking at the preview), this also penalizes those who took more than three minutes to compose an answer -- or those who stopped to improve the question before answering.

This is not the first time I've seen this happen, or I wouldn't bother myself about it.

Question

What can be done to prevent this?

  • Should the warning check more often? Is this feasible?
  • Should there be a way to manually notify others that you are cleaning up the question so they don't need to? A comment could work, as long as no one had already gotten into the edit screen.
  • Could edits be marked as "formatting only" and not contribute to forcing the post wiki (like this suggestion)?
    • Could this be automatically detected if two consecutive revisions are the same (like revisions 3 and 4 in the example question)?
  • Should rollbacks not count toward wiki status?

See also

Five edits within the first five minutes caused a question to become CW -- probably a duplicate, but from a little different angle and posted before the simultaneous-edit warning was implemented.
(Also I'd written this whole question before finding it and it seemed like a shame to just throw it away.)

share|improve this question
    
What can be done to prevent this? The most important question is Should this be prevented? –  perbert Apr 9 '10 at 17:25
5  
@json: I thought I addressed this. You're not just penalizing the one asking the question, you're penalizing those answering it. –  mmyers Apr 9 '10 at 17:27
2  
I might consider this a problem if it had ever happened to me. But it hasn't; not once in the time I have been on SO. –  Robert Harvey Apr 9 '10 at 17:29
1  
Related/duplicate: meta.stackexchange.com/questions/6495/… (And my new answer there would completely remove the CW problem here.) –  Gnome Apr 9 '10 at 19:33
    
Another great edit war, which turned the question into CW: stackoverflow.com/posts/4566526/revisions –  marcog Dec 30 '10 at 22:07
1  
This is now irrelevant, as that question would never make it past our quality filters to begin with. Cure the disease, not the symptoms. –  Jeff Atwood May 31 '11 at 4:46

5 Answers 5

up vote 1 down vote accepted

I'm going to go with:

If you don't want your question bumped to CW because people tried to fix it, write it well in the first place.

Why? Because if there is a compelling reason for many-edits-->CW (and I'm wishy-washy on that question), then we don't want to provide lots of special cases that can become loopholes.

share|improve this answer
    
+1. You can have the greatest question for upvotes, but if you write like slop it's going to get pushed to CW. It's a lesson. –  Josh K Apr 9 '10 at 17:46
4  
I agree there is a compelling reason for CW over time, but not in the first 20 minutes of a post... –  RSolberg Apr 9 '10 at 18:07
2  
As I commented before: It's not that there's a compelling reason for many edits - the situation here is that the OP makes one mistake, and then 5+ people "help" fix the question all at the same time. Since they don't all submit their edits at the same time, the system sees it as 5 separate edits on the question, which forces it to CW. It's one mistake being corrected by 5+ users simultaneously, not a combative asker who keeps reverting the question after people edit it. –  Jared Harley Apr 9 '10 at 18:47
    
::goes and really looks at the questions in questions (heh!):: Hmm...maybe that does make the case for a grace period. Certainly I won't kick up a fuss. But writing well in the first place is the surest bet. –  dmckee Apr 9 '10 at 20:27
    
As Jeff pointed out, this is now actually enforced by the system, so this is really the correct answer. –  mmyers May 31 '11 at 4:58

How about making it so that edits within the first ~10 minutes of a question don't count towards the triggers?

The idea of a wiki is to have community-generated content, where many people edit the post to contribute, or update it as needed over a long period of time. In this way all the triggers make sense, except there is no bounding on the time density of the edits.

Improving a question (without the actual content changing) like this isn't quite in the spirit of what a wiki should be, and as you mentioned, going into wiki mode like that really isn't fair to the asker or potential answerers. The wiki-ness of a post should be evaluated after the content is up to community standards, which happens pretty quickly in the case of a new question.

Edit: For those who haven't seen this in action, here is an example I was just involved in.

share|improve this answer
5  
I like this idea - you could even push it to 15 minutes or so. You're right about the idea behind wiki, it's a long-term thing. A new asker who makes a mistake in formatting shouldn't be punished by getting no rep because he made one mistake (code formatted wrong) and 10 people tried to help at once. –  Jared Harley Apr 9 '10 at 17:57

The problem with the warning check isn't that it doesn't fire often enough, it's that it doesn't necessarily do anything - even if it fired continually and was able to catch edits submitted nearly simultaneously, the editor would have to see it and respond... Turning editing into a twitch game isn't really workable.

Comments can work, but only if everyone checks for them before submitting rather than before editing. Otherwise, you're back to the whole timing problem.

Changing the wikification rules might be a useful thing to do anyway, but still doesn't help with the problem of a later revision undoing helpful changes made by an earlier one. Personally, I'm less inclined to feel sorry for the OP (who now stands a better chance of getting an answer) than I am to pity the editors who wasted their time and now have to go and clean up after themselves.

IMHO, the only robust solution is to check the revision being edited against the current revision during submission, and force the editor to merge if the former is less than the latter. Even if that's nothing more than displaying the two revisions side by side and letting the editor copy from one to the other.

share|improve this answer
1  
Yes, that's about what I'd come up with -- some sort of merging page is what I've wished for for a while. –  mmyers Apr 9 '10 at 17:29

To prevent edit clashes you can do what Shog9 said (it was proposed several times before) or you can go with the locking strategy.

Possible implementation:
When the edit link is clicked, a flag in the database is set, that this post is edited. Every click on the edit link has to query the database, if it is already set or not. You cannot edit as long as it is set and you are informed about that.

These two options are used by version control systems for years. I prefer Shog's suggestion so.

share|improve this answer
2  
Possible, but the flag would have to either age away or check a heartbeat signal. Otherwise, you'd have Visual SourceSafe. –  mmyers Apr 9 '10 at 17:49
1  
Well, at least you got the idea, @mmyers ;) –  Ladybug Killer Apr 9 '10 at 17:52
    
The web is not stateful no matter how badly you want it to be. –  Earlz Apr 9 '10 at 19:00

There is a reasonable argument to be made that if your question requires extensive work by 5 other people, then at the end of that work it is no longer your question, but the community's question.

share|improve this answer
    
And the community's answers? Again, I realize there isn't a lot to say for the poster (although the question did not require extensive work by five people), but all subsequent answers are forced into CW mode also. –  mmyers Apr 9 '10 at 17:56
5  
The problem is not that the question is so bad it requires extensive work by 5 other people. The problem is that those 5 people all think they're the first one editing it, and when they submit their edits at different times, the system sees it as a completely new revision, rather than all of them being an edit of revision 1. This then punishes the OP, who made a mistake or two that 5 people tried to fix at once. –  Jared Harley Apr 9 '10 at 18:10

You must log in to answer this question.

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