2

For quite some time now, I have had enough rep to be considered a "trusted user" or whatever, and as a result, I have been able to make edits to whatever I like (except tag wikis - I just got that privilege last night) without the need for other users to approve the edits.

Just now, I saw a post that needed a couple of minor spelling and grammar corrections. I clicked the link, but someone had already put a "suggested edit" into the queue on the question, so I couldn't do my own edit. I checked the other user's edit, and it was identical to what I had intended to do, so I approved it.

Instead of seeing the edit implemented immediately, I got a pop up message saying something like "Another user has to approve this edit before it is implemented".

This strikes me as absolutely ridiculous - I'm trusted enough to do whatever editing I please to questions and answers without approval from anyone else, but when I see a great edit and think "That's exactly what I was going to do", my judgment is no longer sufficient to correct someone's spelling and grammar?

It seems to me that we should either:

  1. Not allow new users to suggest edits

  2. Let a single trusted user approve and implement edits suggested by new users

  3. Let trusted users override pending suggested edits from new users, while allowing the new user to get the +2 if their edit makes the cut

  4. Something else that would have the result of allowing trusted users to avoid the need for a second opinion when approving an edit from a new user.

8
  • I don't like 1, think 3 would be rather difficult, but really agree with no. 2 Nov 13, 2015 at 4:22
  • 1
    SE used to require one review per edit, but increased it to two (three for SO) to stem the flood of low quality edits getting through the queue.
    – alexwlchan
    Nov 13, 2015 at 8:35
  • 2
    Why is this such an issue? Do we have a real problem with low-rep edits not getting through because of a lack of high-rep approvers? What problem do you want to solve here? Just "there is a mismatch between editing rights and approval rights" isn't a real problem, IMO. Nov 13, 2015 at 11:38
  • I'd prefer 5. allow editing posts without approval only to diamond moderators, unless the post is community wiki.
    – b_jonas
    Nov 13, 2015 at 12:09
  • 3
    @b_jonas That seeoms to go entirely against SE's policy of collaborative improvement of content (at least without making the review queues explode that is, imposing exponentially more work on everyone reviewing in contrast to just the individual work of the editors).
    – TARS
    Nov 13, 2015 at 12:30
  • @TARS: tag wikis (with that name) also require approval to edit. How does that line up with the collaborative thingy?
    – b_jonas
    Nov 13, 2015 at 12:35
  • @b_jonas They're also much rarer and slightly less important than posts, though. And at a certain (non-mod) reputation level even they don't require approval either anyway.
    – TARS
    Nov 13, 2015 at 12:36
  • 1
    @b_jonas Tag wikis have a higher threshold for automatically approved edits (presumably because they're rarer), but that threshold does exist.
    – alexwlchan
    Nov 13, 2015 at 15:54

2 Answers 2

9

While Jason Baker already gives quite an exhaustive rundown of the motivations for a multi-user review process, I'd like to approach your notion of a "ridiculous double standard" a little more from a mere conceptual viewpoint.

There is very well a difference between making that edit yourself and just giving someone else doing it your Ok. An actual edit comes with a much bigger personal responsibility and also with much more work. So if you do an edit yourself, you have to bring the motivation and dedication to perform that edit properly. This is not the same amount of motivation and dedication that is required for just saying "yes" to someone else's edit. Thus the bar for approving suggested edits is far lower than the one for actually performing that edit yourself.

And from personal experience on a site that is far smaller than StackOverflow (where I'm sure this problem is leagues bigger), and even smaller than this site, I have seen valuable high-rep users who normally perform very good edits themselves robo-review and approve downright invalid suggested edits (and I'm talking invalid here, spam, answers in edits, plagiarism). It is unclear if that's to get a badge, to clean the queues or just genuine insecurity about the validity of the edit and a reluctance to just "skip" it, but it all comes down to the fact that clicking that "approve" button is very easy and doesn't (directly) attach your name to it, it's someone else's edit and thus also someone else's responsibility.

If the edit is really good, then it will get through once another approval comes anyway. If you would have made the edit yourself, well, maybe you should have. And if the suggested edit was faster, then you'll just have to live with the waiting time. There is a way to somehow circumvent this review process. You could just click "improve" and make some minor change (like adding a space or what), this will immediately approve the suggestion. This might be seen as a bit of a leak in the system, as Jason does in his answer, but then again, improving the edit, be it only trivially, still requires you to actually do something and put your name on it, it still comes with the implication of an actual edit by you and thus already raises the commitment bar towards the levels of an actual edit. If you are ready to go to such lengths just to have the edit approved faster, then you're comitted enough to the edit to trust your actions anyway. So I don't think this is really a way to trick the system from a mere practical viewpoint.

7

This sort of feature would almost certainly have to be network-wide, so Super Meta may be a better place for this request. But here's my take anyway.

Don't allow new users to edit

Strongly opposed. There's a huge swathe of edits that you don't need a lot of experience with a site to handle; the spelling and grammar example you give is one. There's lots of value in letting low-rep users make some trivial edits. Even anonymous edits have value; 41.5% of anonymous suggested edits have been approved.

Sure we probably could let the high-rep users handle them, but why?

Give high-rep users a weightier approval hammer

Some variation on this has been suggested a bunch of times before, notably:

Three drops in a veritable tsunami of examples. The usual response given is a historical one; something like this was tried out on StackOverflow a yonk-and-a-half ago, and it ended up spamming the site with a bunch of really bad edits. And while I'm not personally convinced by an argument from historical experience, there's another argument I like, made by Servy in an answer to one of those Meta.SE posts:

The suggested edit queue has demonstrated time and time again that rep is not a very good indicator of the quality of the reviewer. There are lots of reviewers with a lot of rep doing a constantly poor job reviewing.

Maybe this isn't a problem on SFF; I don't know. I do know there have been times where I've disagreed with the outcome of a review, so having another set of eyes isn't an inherently bad idea.

I might support this more if it were tied to something more indicative of review experience; maybe you get a hammer with the review queue or editing badges , or something. But even then, I mean; I have three (almost four) gold review queue badges, all six review queue silver badges, and the gold and silver edit badges1, and there are still times I want another set of eyes on one of my reviews.

But there is one aspect of the suggested edits queue that I think is a bit silly, but I'll cover that next.

Let us override suggested edits

Unless I'm radically misunderstanding what you mean, we kind of already have this: it's basically what the "improve edit" and "reject and edit" buttons are.

In principle you could improve a suggested edit in an extremely trivial way, say by adding a comma somewhere; the original edit is automagically approved, and the user who suggested it gets the +2 rep.

Alternately, you could choose to "reject and edit", and just make all of the same proposed changes again; the edit is automagically declined, and the original proposer gets no rep.

I think that's a bit silly.

Something else

I'm unsure what, if anything, would work. A couple thoughts off the top of my head:

  • Reduce the reviewers required on trivial edits. Problem is there's no easy way for an automated tool to determine triviality, but still protect us from destructive edits being approved by lazy reviewers; it doesn't take much to introduce vandalism or profanity, and having only one pair of eyes on an edit like that leaves us very susceptible to robo-reviewers

  • Give certain badge-holders a suggested edit hammer. I don't know how difficult this would be to implement on the SE platform, but it seems reasonable enough on the face of it; we give a dupehammer to gold tag badge holders, so why not give a smaller hammer to someone who has, say, the copy editor badge?

    One issue I see is that edit badges aren't tied to contribution quality the way tag badges are; getting a gold tag badge is rather hard (unless you're Richard, who seems to attract them magnetically), because you need to demonstrate a volume of quality contributions. The edit badges satisfy the volume requirement (500 edits is rather a lot), but not the quality requirement (there is a minor quality baseline, since rolled-back edits don't count, but that's negligible). In theory, you could get that badge with 500 one-character edits, and I would not feel comfortable giving a hammer to the person who did that


1 Hashtage humblebrag

2
  • Well, while you can basically circumvent the 2-reviewers limitation by trivially improving, this already stops robo-reviewers still, since it's still more work and active engagement (and looks more iffy) than just blindly clicking approve.
    – TARS
    Nov 13, 2015 at 10:56
  • 2
    +100 for the quote that rep is not a good indicator of review quality. This is why we still need multiple reviewers, even if one of them has 100k rep.
    – Null Mod
    Nov 13, 2015 at 15:18

You must log in to answer this question.

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