76

This is going to be pretty short and somewhat to the point. I have a big problem with what goes on in the suggested edits section of /review. Obviously there is a lot of mindless approval going on, but I think that the other big problem here is the massive amount of "too minor" edits that are slipping through.

The argument has already been made that these edits still benefit the site, which of course is true. But, a lot of them technically are still too minor.

I propose a simple little "minor" check box next to the approve button that allows us to approve a minor edit but with a diminished reward the the editor. (1 or 0 rep for approval) I can't be the only one who's getting sick of having to reject people all day.

A couple more good examples of what @Bo Persson is talking about:

8
  • 18
    You don't have to reject, you can edit the post into better shape and untick the "suggested edit was helpful" box (if applicable - I always do if I have to edit more) Commented Oct 8, 2012 at 9:38
  • 10
    Interesting idea, but if there are already a lot of mindless approvals, I fear that people would just mindlessly ignore the check box and just approve.
    – yannis
    Commented Oct 8, 2012 at 9:39
  • 1
    Good point Ben, and @YannisRizos it's true that would probably happen, but I think it could still help those of us who are actually reviewing. Commented Oct 8, 2012 at 9:40
  • 6
    It doesn't help much if you try to reject it either stackoverflow.com/review-beta/suggested-edits/687660#./687660
    – Bo Persson
    Commented Oct 8, 2012 at 12:46
  • 17
    @Ben: good idea in theory, not so good in practice. Mindless approving takes nearly no time or effort; actually improving requires both. By the time an improver finishes up, the bad edits are usually already approved. See What can we do to stop bad edits getting accepted?. My answer to that question proposes some solutions.
    – Pops
    Commented Oct 8, 2012 at 21:05
  • 1
    I think this is still an issue. I'm currently approving edits which are nice, but pretty minor. I'd really like a way to mark them as minor.
    – Nathan S.
    Commented Mar 4, 2014 at 18:28
  • 1
    @Yannis So have it automatically checked.
    – bjb568
    Commented Aug 15, 2014 at 3:31
  • Relevant on SO meta, more recently (shameless self-plug): meta.stackoverflow.com/questions/420349/… Commented Dec 20, 2023 at 3:18

4 Answers 4

71
+50

There are a couple of problems with this:

  1. The meaning of "too minor" varies widely between individuals. Are trivial changes always too minor, or only too minor when they ignore other, more damning problems with the post? The edits you provide as examples illustrate this nicely - even the last one isn't problematic so much because it's minor but rather because it introduces more problems than it fixes!

  2. Implemented as you describe, it's nothing more than a rep-denial mechanism. "I like your edit, but not enough to reward you for making it". There's no way to enforce its use in cases that actually justify it, meaning the actual application is likely to be seen as capricious. See also #1...

Note that there's an existing facility for preserving edits while marking them as "rejected": a checkbox on the "Improve" page. However, this suffers from both of the problems noted above and has the additional problem of edits being approved while a reviewer is improving them!

Fortunately, all of these problems can be solved... Here's my solution:

  1. Prevent pending edits from being reviewed by more than one reviewer at a time. This is impossible to implement perfectly, but it is possible to hold reviews for a short period of time after one reviewer has started reviewing, thus ensuring that reviewers have a little bit of time to evaluate edits without being sniped by other, faster reviewers. see here.

  2. Replace the "Improve" button with a new "More changes needed..." button. This will produce a pop-up menu with two options:

    • Approve suggestion then edit will open the editor with the suggested revision, behaving as the current Improve does but lacks a checkbox. Submitting the edit approves the suggestion.

    • Discard suggestion then edit will open the editor with the current revision, implicitly rejecting the suggested edit with the following reason:

      Your edit did not correct critical issues with the post - view the revision history to see what should have been changed.

      A link to the revision history should be provided if possible.

      with this UI:

      Approve, Improve Edit, Reject and Edit, Skip

  3. Get rid of the "too minor" reject reason entirely. If it's really too minor, reviewers should demonstrate that by providing a not-minor edit. If the reviewer opts to build upon the edit instead of starting over from the current revision, then it isn't too minor! .

36
  • 14
    I like number three.
    – user102937
    Commented Aug 13, 2014 at 17:26
  • 1
    Excellent ideas, one question though: rejections via "More changes needed" button would count towards the edit ban same way like the "Reject" button itself? Commented Aug 13, 2014 at 17:36
  • 4
    We don't lack review manpower nowadays (maybe we lack good reviewers, but that's a whole different kettle of fish). So we could go without “too minor” altogether. If an edit is really a waste of time, reject with a short but explanatory custom message. But “Accept and improve” must stay. Commented Aug 13, 2014 at 18:19
  • 3
    This is the most sane thing I've seen in regards to this, since day 1. Let's do this.
    – user50049
    Commented Aug 14, 2014 at 14:44
  • 1
    Ah, now this is what it should have been all along! Keep “improve”, and fix “reject and improve” where we had to manually undo bad edits. Just a minor point: please don't make us click a menu with two options, put two different buttons “Improve” and “Discard and edit”. Commented Aug 14, 2014 at 19:49
  • 4
    @Gilles The bad reviewers aren't going to care whether the suggestions are minor or not, but the good ones likely are. If they see that a continually increasing number of suggested edits are less and less valuable, they're more likely to be dis-heartened and stop reviewing, because after all what's the point of reviewing suggested edits if no meaningful percentage of the edits matter.
    – Servy
    Commented Aug 14, 2014 at 19:51
  • 1
    I suggest a more consistent wording of your choices after you click 'More changes needed...' I suggest labeling them "Apply suggestion and edit" and "Discard suggestion and edit". Commented Aug 14, 2014 at 20:08
  • 1
    Why in earth did this took 2 years? We knew for 2 years that the whole review thing is fundamentally broken. It's still not enough to properly fix it, but I'm happy to see that finally a serious change is coming. I sincerely hope that more will follow. All those monkeys who are consistently degrading Stack Overflow since 2 years have completely killed my will to review and edit content.
    – user138231
    Commented Aug 14, 2014 at 21:53
  • 2
    This is a wonderful idea, however the one problem I have is about the approve/deny mechanism. Sometimes, a person has made a few good corrections, but there's a ton more work to do. This would be 'too minor' (as there's a ton more work left), but why should I be forced to correct these items again, as 'rejecting' it would discard the edit?
    – hichris123
    Commented Aug 21, 2014 at 2:02
  • 2
    If you are providing an edit, you are providing an example, @air. That's really preferable to a canned reason, IMHO.
    – Shog9
    Commented Aug 29, 2014 at 15:45
  • 6
    I'm not reviewing, but if I were then I'd be bothered by 3: now suddenly reviewers are forced to edit themselves (or skip, putting the minor edit in the hands of robo-reviewers)? (cc. @RobertH)
    – Arjan
    Commented Aug 30, 2014 at 7:53
  • 4
    But wouldn't "too minor" still be the best reason to reject suggestions that are just that? I guess it would be weird to select a different reason, if one encounters it but one doesn't feel like editing themselves. (Maybe especially for posts that are beyond repair, or plain off topic, but from which someone still just removed some "Thanks". But then, maybe those who review are willing to improve edits.)
    – Arjan
    Commented Aug 30, 2014 at 15:26
  • 2
    If the changes you're preserving are really worth all that trouble, @Jonathan... Then why not just Improve Edit? I've found that posts which require additional extensive editing are rarely worth preserving a few typo fixes - chances are, I'll be re-writing the entire sentences surrounding them anyway. If that's not true, you're effectively taking someone else's work without attribution, which is widely seen as a faux pas - the fact that we ever allowed this was probably due more to the tacked-on nature of the feature (Improve + Reject was not originally possible) than anything sane.
    – Shog9
    Commented Sep 10, 2014 at 23:23
  • 3
    Everyone misses something sooner or later, @Jonathan. The question is, does the suggestion make a substantive enough improvement that simply discarding it would be a waste, or is it so trivial / pointless / unhelpful that throwing it away is no great loss. Discussing this in abstract is always somewhat difficult, so I'll provide an example of what I'm talking about: worth preserving vs. not worth preserving.
    – Shog9
    Commented Sep 11, 2014 at 0:38
  • 2
    Agree that changing code is controversial, @Jonathan - but that's a separate discussion. FWIW, I'm firmly on the side of "approve what improves" when it comes to reviewing edits, regardless of how.
    – Shog9
    Commented Sep 11, 2014 at 15:23
23

Minor edits (that only include positive changes) should be approved but with no rewards given, and possibly, all edits should be considered minor unless otherwise stated (by the reviewers). I have never subscribed to the idea of edits being too minor because it leads to subjective review decisions (some reviewers are understandably more lenient than others even though rejecting minor edits is generally a safe decision to make, given the current policy).

This will solve the following issues:

  1. time wasted for deliberating whether minor edits are too minor.
  2. time wasted for improving minor edits so that they are no longer minor
  3. pple losing sleep over minor edits being too rewarding
  4. helpful edits being rejected for being too minor
16

This proposal assumes that getting free, quick rep is the goal of people who suggest minor edits. (Otherwise, reducing the rep reward wouldn't be an effective solution.) Let's say that we do reduce the reward all the way down to zero, not just one.

That wouldn't solve the problem, because people would still get the regular two-rep reward from reviewers who didn't use the "too minor" checkbox. We wouldn't actually be getting rid of the rep incentive at all; we'd only be reducing the probability of receiving rep for any given suggestion. If anything, it'll make the problem worse, as people start to submit a higher volume of suggestions to compensate for the reduced probability of reward per suggestion, and individual suggestions get even worse because the suggesters don't want to increase the amount of time they spend on rep farming.

Getting away from the rep argument, this would also teach people that submitting minor edits, while not ideal, is still acceptable. That seems contrary to the usual goal of getting people to fix all the problems in the posts they edit.

9

I already mentioned this idea in a previous discussion. So I'm in favor.

With any minor edit, I try to save the effort by adding some extra edits, but it isn't always possible. And there are still a lot of posts needing good edits so anyone who wants to earn their 500 times 2 rep have all the possibilities in the world.

// Private rant space: Then again, there are still a lot of bad edit suggestions. (adding thank you, change I to i to mention a few)

7
  • 1
    "adding thank you, change I to i to mention a few" sounds unlikely. Are you sure those aren't concurrency issues? Commented Oct 8, 2012 at 10:57
  • 1
    I have seen several of them. 2 today, maybe concurrency, but they counted as suggested edits. Commented Oct 8, 2012 at 10:58
  • 4
    I suspect they simply edited a version where the "thank you" was still there, and somebody edited it out while they were editing it, making it seem like they added it back in. Commented Oct 8, 2012 at 11:00
  • 1
    Here is an example of a seemingly added thank you: stackoverflow.com/review-beta/suggested-edits/770684 Commented Oct 8, 2012 at 11:57
  • 8
    There are people who do actually actively add "thank you" to posts. Commented Oct 8, 2012 at 19:15
  • 2
    @NullUserExceptionอ_อ, do we have a spare giant I? Commented Oct 8, 2012 at 20:02
  • And here's an examnple of a I turned into a i: stackoverflow.com/review-beta/suggested-edits/774981 Commented Oct 9, 2012 at 7:48

You must log in to answer this question.

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