Should tiny edits be accepted or rejected in review?

Question

I regularly review suggested edits on the review page. One common situation is that a user will propose a tiny edit, like capitalising a single letter, or removing only the salutation from a post. In many cases there are still multiple further improvements that could be made, but that editor hasn't seen them - or can't be bothered.

I feel uneasy accepting such edits. On the one hand, it reminds me of the army of 'twiddlers' on Wikipedia, who spend their time making tiny inconsequential edits when they could be doing something much more useful. It also feels lame to award someone any points at all for capitalising a single letter in a title. I sometimes feel that this may be a kind of gaming of the system on the part of some users - suggesting multiple small edits for point gain rather than just fixing the whole post. I also see this with tag wiki editing.

On the other hand, it is a valid edit, and it does objectively improve the post.

There has been a prior ruling on part of this. There was some substantial discussion about the aggressive removal of salutations during the Rich B era (remember that wonderful chap?). I seem to remember that the conclusion was that removal of "hi! / thanks!" on their own was not sufficient grounds for a post edit, although I can't find that now (this post is relevant, but inconclusive).

I have been erring on the side of 'this edit is pointless' and voting to reject, but I notice I am often overruled by other reviewers.

Perhaps one answer would be to provide a third option, after accept/reject:insufficient?

What behaviour are we intending to reward here?

Answer 1

Edits should be accepted if they are correct or helpful, and rejected if they are incorrect or spam. The length of the edit is irrelevant. Who cares if the edit is tiny or major as long as it's improving the quality of the site's content? The reviewer's time has already been used in checking the edit, so nothing is lost in accepting it.

The How to Edit box next to the edit window encourages exactly these kinds of changes, so it would be pretty hypocritical and counterproductive to reject them:

How to Edit

► fix grammatical or spelling errors

► correct minor mistakes

The only valid concern I see here is that people could use spell-checking bots to "game the system" and gain rep without doing any work. But that's not solved by preventing small edits. It's solved by not rewarding rep for small edits. That's a problem with the site's design, not with the person going out of their way to improve the quality of its content. If the edits are constructive and helpful, accept them.

(And the "army of 'twiddlers' on Wikipedia" are the people who wrote Wikipedia. The vast majority of edits are small changes, that, together, produce the articles you read. It's a collaborative effort.)

If we're discouraged from making small edits, and discouraged from making sweeping edits ("always respect the original author"), then what's the purpose of including a wiki function at all?

Answer 2

I feel uneasy accepting such edits. On the one hand, it reminds me of the army of 'twiddlers' on Wikipedia, who spend their time making tiny inconsequential edits when they could be doing something much more useful. It also feels lame to award someone any points at all for capitalising a single letter in a title. I sometimes feel that this may be a kind of gaming of the system on the part of some users - suggesting multiple small edits for point gain rather than just fixing the whole post. I also see this with tag wiki editing.

I completely agree -- reject them, with extreme prejudice.

This is why we limit edit suggestions to 6 characters; there are still some hacky ways of getting around that.

Sam was against making this stricter, but I am going to overrule him on this and implement more strict checking myself; I would rather be too strict and reject some edits than have to deal with a continal stream of character-twiddling edit suggesting users.

(yes, there is an exception when a single character like a semicolon makes code wrong, but again -- is everything else in that post so perfect it cannot be improved? I doubt it.)

edit: we now have stricter enforcement of "no using whitespace to bypass the 6 char editing minimum for edit suggestions."