37

I'm throwing this feature request out in response to the following two questions.

The current system for suggested edits seems to work perfectly fine in all cases but one: when a user (or group of users) decides to search for possible typos and edit posts en masse. I entirely agree with the following logic.

  • I do not believe it is possible to fix all the issues that might exist in the posts this way.
  • If you don't have edit privileges, you're causing work for users. Now, nobody minds this, unless you're not fixing all the issues in a post and it's clear there's no thought going into the process at all.
  • Editing bumps things to the SO home page. Again, nobody minds and this is by design, but if you're fixing one tiny issue, is that a reason to bump 100 questions to the home page? I'd say if you're fixing all the issues with a post - including flagging what should be flagged, closing what should be closed etc - great. If you're not, you're bumping a whole lot of stuff...

There are a few more issues to consider, as well.

  • Even though a single user can (theoretically) suggest an unlimited number of edits, reviewers are limited in the number of edits that they can accept or reject per day.
  • Most behavior on the Stack Exchange network is rate-limited: why shouldn't suggested edits be, as well?
  • A single user's suggested edits can take up a disproportionate amount of the reviewers' time.

I therefore propose placing a limit on the rate of each user's suggested edits. There are a number of ways we could do this.

  1. Limit the number of suggested edits per user per day—to approximately 100? This is a high enough threshold for any reasonable number of daily edits, and besides, after 100 suggested edits get approved there's no rep in it for the editor anyway.
  2. Limit the number of suggested edits per user per hour.
  3. Limit the number of outstanding (unreviewed) edits per user at any given time.

I personally like the last solution because it would encourage people to spend more time on each suggested edit and it would optimize for the valuable resource that is being spent on each suggested edit: reviewers' time.

| improve this question | |
  • 2
    That'd be more than one Strunk & White worth of edits per day. Do we have evidence that a non-trivial number of people are editing prolificly enough to warrant this? – Pops Dec 20 '11 at 20:24
  • 2
    @PopularDemand the one thing I find odd about this is that except for answering I can't think of anything else that isn't rate limited. – Some Helpful Commenter Dec 20 '11 at 20:27
  • So what you're saying, @ConradFrix, is that all Chris has to do to get this implemented is to whip up a bot that looks for particular misspellings on SO and serially edits them? That... actually sounds doable. – Pops Dec 20 '11 at 20:29
  • 1
    @PopularDemand You have a point, I don't know how many people this would apply to (I may make a Data Explorer query for it when I have time, though). I'm making this request not because I think that it is a very high priority, but because it would serve as a limiting mechanism when a few users do decide to start editing posts too quickly. Now that you mention it, the issue may be simply to prevent people from making too many posts too quickly, which would indicate that not enough effort was spent on each edit. – Chris Frederick Dec 20 '11 at 20:31
  • 3
    @PopularDemand yep. By the way this query gives counts of suggested edits per day that were above 50 for the day. The requested limit has been met 18 times. If you lower the threshold to 50 its happened 82 times – Some Helpful Commenter Dec 20 '11 at 20:35
  • Is there a 200-rep cap on edit rep? I'd never heard of it (I guess I don't edit profusely enough to have found it by accident, either). – Flimzy May 31 '12 at 16:52
15

Four and a half years later:

This is now implemented, although not quite as requested: Rate limit suggested edits

  • at-most 20 pending suggested edits per editor on beta sites
  • at-most 5 pending suggested edits per editor on graduated sites

Hopefully, it will suffice all the same.

Back in 2011...

I could have sworn this was already in place, but looking now I can't find any documented evidence of it. In a way, I suppose that makes sense - there's no limit on how many edits you can do once you have the editing privilege after all.

I do know that if you get too many rejections, you'll be prevented from further suggestions for a period of time... So perhaps the lesson here is, if someone's spamming the site with useless edits, start rejecting them.

The "suggestions per hour" (note: not edits per hour - let's keep these straight here, since suggestions have an entirely different set of problems from normal edits) idea seems reasonable to me, as it does seem to encourage putting more thought into editing. Let's say, 20 suggestions per hour (sliding window) - that shouldn't really slow anyone down under normal circumstances, while allowing for an average of 3 minutes per edit.

| improve this answer | |
  • I have nine Shogs. Your edit is invalid. – Tim Post♦ Dec 20 '11 at 20:38
  • In the specific case prompting this discussion, despite several users spending all their edit votes on rejections, the useless edits seemed to continue until the user simply left the site. – jscs Dec 20 '11 at 20:42
  • 2
    The only problem with this solution is that the edits themselves may be valid (as explained here). An edit should be accepted or rejected on its own merits, not simply rejected to rate-limit a user. – Chris Frederick Dec 20 '11 at 20:42
  • 1
    @Chris: if the edits are valid, then what exactly is the point of rate-limiting them? This reminds me of "Jon Skeet posts an answer an hour - can't we slow him down to give the rest of us a shot?" – Shog9 Dec 20 '11 at 20:47
  • 1
    @NineShogsShogging Two things: (1) they bump questions to the front page much more quickly than even Jon Skeet ever could; (2) they take up a disproportionate amount of reviewers' time. – Chris Frederick Dec 20 '11 at 20:50
  • 2
    @Problem: if the root problem here is a disagreement between some users and some moderators as to what constitutes a valid edit, then that needs to be addressed head on. – Shog9 Dec 20 '11 at 20:50
  • 1
    @Chris: they only bump if they're getting approved. They're only getting approved as folks approve them. That's the whole point of having a queue, y'know? They can sit there and rot if folks don't have time to deal with them. Of course, if they're useless edits, then they're potentially clogging the queue - but if they're actually useless, then they should probably just be rejected (at which point they don't bump anything (and in sufficient quantity stop the editor from further editing)). – Shog9 Dec 20 '11 at 20:53
  • @Nine: I agree, but my point was that even if we had a consensus that some batch of edits should be rejected, it seems to take a lot of concerted effort by reviewers to trigger this suggestion prevention mechanism. – jscs Dec 20 '11 at 20:54
  • 6
    @Problem: for the user in question, I suspect a lot of this was due to early edits being accepted fairly quickly, with most of the rejections coming in as pile-on after meta discussion. That's not really the problem this was intended to solve - normally you'd expect to see rejections more evenly distributed. IOW, if you're rejecting valid edits simply because you think the editor is doing too much editing, you're abusing the system as much or more than he is. – Shog9 Dec 20 '11 at 21:01
  • @Nine: Agreed again; I made a point of actually checking the edits that I rejected and I believe that they were indeed Too Minor. A rate limit or cap on suggestions might not have much affect, but it seems that it would allow reviewers as a group a little more time to evaluate a pile of suggestions, and might prompt the suggester to slow down, take a look at the results of the review, and perhaps be more thorough. I don't understand why you think rejections should be distributed, though. – jscs Dec 20 '11 at 21:24
  • 1
    There is an upper cap to the number of absolute suggestions allowed at any one time. I don't think there's a per day thing. – Grace Note♦ Dec 20 '11 at 21:42
  • @GraceNote Interesting; I didn't realize that all users shared a finite suggested-edit queue. Do you happen to know if and/or how often the queue gets completely full, blocking all new suggested edits? – Chris Frederick Dec 20 '11 at 22:28
  • @Chris I think it only ever can happen on Stack Overflow in terms of feasibility. It's very rare, though, far as I know. – Grace Note♦ Dec 20 '11 at 22:48
3

I actually don't think this is necessary. Due to the 200 daily rep limit there's really no point in doing much more than 100 edits anyway. Which is probably why the query I wrote bears this out.

Also since according to the to the suggested edit faq you can only get a certain number of rejected edits and keep doing it, so its not like your going to do 200 edits in the hope that half get accepted because that won't work long term.

What about abuse?

There are strict limits enforced. If a user (anonymous or registered) submits many rejected edits they will be automatically banned from suggesting edits. The fixed size queue also helps protect us from abuse.

| improve this answer | |

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