Should I approve spelling-and-grammar–only suggested edits?

Question

I've been seeing lots of suggested edits recently that are simple spelling and grammar fixes:

I've tended to reject several of these, because they're not really substantial edits. From an old blog post of Jeff's:

If you are going to edit a post, make sure you’re substantively improving it. Avoid making isolated, trivial edits, as they are the source of much friction. For example, don’t bother changing “its” to “it’s” unless you have several other edits to make in the same post. There has to be a legitimate case that your edit made multiple changes transforming the post from good to great — or at least substantively improving it.

However, there are some such edits which clearly provide huge readability improvements (not to mention taking a lot of effort to make the edits):

Considering the goals & repercussions of approving edits:

• We want to improve the general quality (and this includes readability) of questions & answers
• ...but, we have CHAOS to do this for us — but they're not on every SE site, and can't do everything anyway
• Users gain reputation (to a certain extent) from having their suggested edits approved

So how should I decide whether to approve these edits? Should I reject them all? Only approve really extensive ones like the 2nd and 3rd examples above? Or should I instead choose to "Improve" and submit the edit myself, to provide the same improvement in post quality without giving rep to users for "insubstantial" edits?

Answer 1

I would definitely approve those edits. The general rule is to approve any edit that improves the quality of the affected post. That helps to make the site better overall, and is a worthwhile goal in itself.

Most of the rather trivial edits are blocked from submission altogether with the requirement that more than 6 characters must be changed. Of course, if you see overly trivial edits that fail to fix glaring problems with the post (such as formatting code blocks, removing noise salutations, etc.), you should probably reject those—I do. You'll have to use your own judgment. You might also reject trivial edits on old questions (say, more than 6 months), because there's little reason to bump those old questions to the top of the list just for trivial corrections. But there's probably no good reason to reject valid spelling and grammar edits on recent questions.

If the editor made some useful corrections, but didn't catch everything, you should definitely use the "Improve" button to make the rest of the edits yourself. It's unreasonable to expect a single editor to catch everything him/herself. These are not edits from particularly experienced users; in fact, the suggested edits feature is kind of a "training ground" for future editors. They might miss things, and you should help out by going ahead to fix them yourself. But I would reject an edit, like I mentioned above, if they missed some particularly obvious or glaring problems with a post.

So I guess I disagree with Jeff's first guideline ("transforming the post from good to great") because I just don't think every edit should be held to the standard of "greatness", but I do agree with his revised guideline ("at least substantively improving it"). That should be necessary; the only quibble is over what you think qualifies as "substantial". I might be jaded from the number of actively bad suggested edits that I see. Nowhere near enough people vote to reject those, and these are the ones that need to be rejected. It's far less crucial for anything that falls into the "net positive" category.

But beyond all of that, I think you misunderstand the purpose of CHAOS. They're not intended to replace edits by regular community members. This is still a huge job, and it's important for everyone to pitch in. That's why we have a suggested edits feature in the first place. As Joel says:

Before they start to promote a site, they will spend a couple of days sweeping up a little bit just to get the place in order. This is the world's most superficial cleanup: they're just looking at the top 1000 questions in order by page views and, if necessary, editing the titles to make them more grammatical, fix spelling, and make them reflect the gist of the question better. [ . . . ]
They also have temporary diamonds while they do these edits so that their edits don't just go into a queue to annoy the real mods, but those are temporary and will go away when this editing project is done.

This is only a one-time, single-event type of edit. They're only fixing the top 1000 questions on the front page, and they're only making fairly superficial edits. They're not just a permanent team of janitors.

And there's nothing wrong with earning reputation for quality edits. It's only 2 points; you shouldn't let that become too much of a deterrent to clicking the "Accept" or "Improve" button. Things only become worrisome if a single user repeatedly submits trivial edits without substantively improving the overall quality of a post.

Answer 2

The first example is a textbook case of when I'd click Reject, and pester about the lack of a Reject and improve button. There are plenty of glaring flaws in that post: “Hi all” leader, lack of punctuation and capitalization, plenty of grammatical mistakes, plenty of misspellings including search fodder like “Telerik MVC 3”… All the suggested edit does is fix a few of these, and some of the changes it introduces are even wrong (all of the code tags).

The second example is a textbook case of when I'd click Improve. There's a post that's comprehensible, but hard to follow because of many grammatical mistakes. The editor has fixed most of them (there's one left: extraneous whitespace around some punctuation — which I consider too minor to fix on its own, but I'd do it if the post was being bumped anyway).

The third example is again a textbook case of when I'd click Improve. It didn't fix all the grammatical problems (e.g. the post should begin “I and a friend want…”), but enough to be helpful. (The edit was automatically rejected when the question was closed for being off-topic, so it's not worth bothering about it.)

So I think you've put the bar in the right place: reject the first example (for being insubstantial — and partly wrong), approve or improve the second and third example, which are substantial improvements.

Answer 3

We want to improve the general quality (and this includes readability) of questions & answers

This is the main goal we are trying to achieve when an edit is for grammar/spelling, rather than an accuracy fix. On a community run site, we can't just leave it to CHAOS to improve this. There are only six members of chaos across the entire network and given the number of questions/day on Stack Overflow alone, I wouldn't leave just that site to them.

Users gain reputation (to a certain extent) from having their suggested edits approved

This is the main point of contention.

From looking at your first example, the user should have:

• Removed the starting salutation.
• capitalised the is.
• Changed develop -> developed a
• Capitalised the start of both sentences and added full stops at the end.

There will be others too, but fixing that post doesn't answer your question.

They have, however, essentially highlighted some words and fixed the spelling of some words, which does make it easier to read. I would have approved the edit, or more likely, improved the edit. (Disclaimer, I don't have 2000 reputation anywhere yet, so I haven't ever put this to practice.)

So how should I decide whether to approve these edits?

If the edit to a post makes it clearer, easier to read or otherwise fixes the content, I would approve. However, be mindful of people who capitalise all the is in a post where key words are misspelt, I would reject those.

It would be ideal if every suggested edit solved all the errors within a post, but as long as there is a net positive to their actions, we are making the sites better.

Should I instead choose to "Improve" and submit the edit myself?

Certainly, I would suggest that if you have the time and you can see things that should be additionally improved, then improve them.

From having my own edits improved and checking back on them after, I have seen a couple improved and learnt to not just focus on one area when editing (I.e. the body and missing the tags, or title and missing the body).