40

Recently I saw two cases where the users changed the whole question (body + title!) to a new question just because Stack Overflow prevented them from asking a new question within X time.

Example 1, Example 2.

There could be a script that checks the question doesn't get a radical change by users with new question restriction.

My suggestion is: edits of those users are treated as suggested-edits which need to be approved.

10
  • These edits would have been approved by a higher rep user... Commented Mar 24, 2012 at 23:01
  • 11
    @Ben. Nope, those edits are to his own post, no approval at all...
    – user173320
    Commented Mar 24, 2012 at 23:02
  • In that case I agree! Commented Mar 24, 2012 at 23:03
  • 4
    Ha, that was clever!
    – juan
    Commented Mar 24, 2012 at 23:34
  • 3
    Just roll them back, there's no scenario where invalidating the existing answers with a question edit is acceptable. If you don't have enough rep then just flag a moderator and they'll do it for you. Commented Mar 24, 2012 at 23:46
  • 5
    If after rolling it back, if they reinstate the changes flag the question for moderator attention: that constitutes abuse. Commented Mar 24, 2012 at 23:49
  • Am I missing something? Example 2 has gone through some reasonable changes, but is still the same question. It simply looks like the OP was struggling with the best way to ask the question and was trying to elaborate.
    – slugster
    Commented Mar 25, 2012 at 2:20
  • 7
    @slugster, Read the OP's comments on that question he admits it with... "ah ok first time I have ever had this 6 questions per day thing", "ugh your only allowed 6 questions in a 24 hour period now? God thats awful.", and "Ive updated this question to reflect what I need", etc. ... His first question was answered, and then he changed it to something related, but different. Commented Mar 25, 2012 at 3:26
  • Thanks @Brock, I did miss something. I only checked the review history of the questions and didn't check the comments.
    – slugster
    Commented Mar 25, 2012 at 3:39
  • 1
    The next time I have a question of my own, instead of using that "Ask Question" button, I'm just going to overwrite this post.
    – Pops
    Commented Apr 4, 2012 at 18:08

4 Answers 4

25

Yes, prevention is better than trying to clean up the mess afterwards. Your suggestion is a good one.

Proposal:

  1. If a user is banned from asking new questions, or throttled from asking a new question for a few hours...
  2. Then any of his edits, of his questions, are treated as suggested edits.
  3. Optionally, only if the edits are more than, say, 80 characters.
  4. And/or only if they are more than, say, 10% of the post.
10

This sounds like a good idea in theory, and I sympathize with the motivation, but I don't think it's actually something we should put into practice.

Why? Because we strongly encourage new users to edit their questions in order to improve them and bring them into compliance with our guidelines. This is especially true after their questions have been closed, but still true all the rest of the time as well. Experienced users will frequently make suggestions in the comments on how to improve a question, and conscientious newbies will follow those suggestions. Therefore, it seems like a really bad idea to complicate those users' ability to do this. In fact, if you ask me, not enough people do it already, so we certainly don't want to make it even more difficult for those rare users who do actually desire to improve.

This is a change that would actually backfire: you aim to improve the quality of questions from new users, but you'd do so in a way that would actually work to discourage new users from improving the quality of their questions.

There's no need to complicate the user interface or process. The existing system works perfectly well. Edits bump the questions up to the top of the questions page, and experienced users who monitor the question page will notice those edits. They can then roll back the edits themselves (or flag for moderator attention), and if desired, leave a comment explaining why they rolled back the edits to the user. You know, exactly what you did here. That worked just fine.

If, for whatever reason, the user rolls back your edits (or tries again), you should flag the question for moderator attention and ask them to lock it. Again, problem solved, and you only step on the toes of those users who have demonstrated an unwillingness to follow our guidelines, not those who are actually proactively trying to improve, just as we ask them to do.

3
  • I disagree. We shouldn't be cops. I don't need to monitor invalid edits. I believe there is a huge difference between change the question and improve the question. And new users know that difference as well.
    – user173320
    Commented Mar 27, 2012 at 13:14
  • 2
    @gdoron: Lots of improvements to the question require drastically rewriting it. Commented Mar 27, 2012 at 16:54
  • Would be interesting to see what % of substantial edits (see points 3 and 4 of Brock's answer) by restricted users were improvements and what were radical changes. Commented Mar 28, 2012 at 10:58
5

While I think @BrockAdams' answer has the right "criteria" for the "filter" to kick in, I also think @TheEstablishment has a great point about not getting in the way of users doing the right thing. I mean, the last thing we want is to stop a newer user from adding a much-needed code sample to their question.

So, rather than limiting the edits that might be damaging, I would propose that a flag is raised instead when this behavior is detected, very much the same way this is done for vandalism currently.

Now, I'm not a big fan of increasing the flagging load, but I think the advantage of catching this behavior while not actually getting in the way of good users far outweighs the added load on moderators and 10k+ users helping to deal with flags. This is especially true when you consider the likelyhood of this behavior actually happening. (whithout actual stats, it's hard to know, but I find it hard to believe that it's all that common)

The other advantage this has is that we can step in sooner, rather than later, and help to correct the user's behavior, rather than simply throwing up road blocks and encouraging them to find workarounds. Having a human be notified of bad behavior is, in my opinion, always preferable to simply hiding it in cases where things can be done to fix the problem.

6
  • +1. Maybe adding a confirm box that ask the new user if it's edit is a "question change edit" or improvement of the question. If he continue with the edit, now the flag comes in. What do you think?
    – user173320
    Commented Mar 27, 2012 at 13:35
  • 1
    @gdoron - I would rather not get in the way of users unless it's absolutely essential. Keeping the barrier to entry as low as possible encourages participation, and editing in general is a good thing. However, if there are stats that clearly show a certain set of criteria are highly (ie. 90+%) likely to be "bad" edits like this, then I would be OK with it. Even in that case, however, a flag should probably still be raised so that a human is made aware of that behavior.
    – cdeszaq
    Commented Mar 27, 2012 at 13:41
  • Raising a flag might be an acceptable compromise. That flag would need to go into the 10k-accessible query; they are all perfectly competent to judge its validity without requiring moderator intervention. Commented Mar 27, 2012 at 16:55
  • Even if 10K's can vote on a flag, it still requires a mod to make the final judgement, right? That's way more trouble than is warranted in this case. It's a perfect situation for suggested edits and SE's get processed much faster and by more people than flags. ... There is no need to increase the mod workload, or to further clog the under-serviced 10K tools. Commented Mar 27, 2012 at 19:52
  • @BrockAdams - You make some good points about resolution speed and the fact that mods need to deal with them in the end anyways, so suggested edits do make more sense. However, the edit in question shouldn't be delayed, but could instead be rolled back, so it would need to be a special sort of edit (which is why I didn't suggest putting it in the suggested edits queue)
    – cdeszaq
    Commented Mar 27, 2012 at 19:54
  • Yes, the delay is the issue here. The urgency with which the flag is processed is not that important. All it does is mark a possible problem. Commented Mar 28, 2012 at 16:01
1

I like the idea of automation as much as the next programmer, but I'm thinking flagging might be a good way to get a feel for the scope of the problem before we start making any big changes.

I would like to reserve the right to an expanded, comprehensive answer when I haven't just woken up.

You must log in to answer this question.