7

A recent experience with a bad edit approval has led me to wander into the meta world of suggested edit reviews. I've read quite a few posts, including in particular:

There seems to be a pretty widespread consensus that between the problem of poor edit proposals and inappropriate edit approvals, we should be focusing on the latter. There also seems to be a broad set of thoughtful proposals as to what should be done, many of them summarized in Improving the Quality of Reviews: Project Honeypot. Finally, I understand that prior to the introduction of the review badges, there was a problem with the edit review queue size growing to an excessive level.

Can someone provide a summary status on SE's view of the current situation? Is there a belief that the current set of badges combined with the "honeypot strategy" is working adequately? If not, is the situation similar to the VTC problem, where further action is delayed due to both SE prioritization and lack of agreement on what approach to take? If I want to provide some additional ideas for reducing poor reviews, is there a particular question I should be answering?

  • How ironic that a nonsense answer was posted to this question. – BoltClock's a Unicorn Dec 7 '13 at 7:27
10

Not sure there's anything even close to a cohesive "company view" on this topic, but here's my $0.02:

Edits are awesome

Editing is what separates us from the animals forums. The ability of anyone viewing a page to step in and fix problems they see is incredibly powerful.

Of course, you don't get great power without great potential for abuse...

Generally, folks get too wound up about minor mistakes

You'll find all sorts of discussion and hand-wringing here on Meta. Some of it is warranted; a lot of it isn't. Yes, folks make mistakes. That's why everything is logged and anyone with the ability to edit a post has the ability to fix someone else's edit.

The review system for suggested edits acts as a sanity-check, not an editorial board

The biggest danger with allowing anyone to edit comes from folks who are overtly malicious or utterly clueless, and would cheerfully fill the site with spam, vandalism and various other forms of abuse if given the chance.

By and large, the review system does a pretty decent job of preventing this.

FWIW:

As you can see, reviewers are not at all hesitant to reject edits, and the results reflect this: the roll-back rate for suggested edits is significantly lower than that for 3rd-party edits as a whole.

Matching up reviewers with topics they know well is hard

This is probably the Achilles heel of suggested edits and, well, /review in general: it's site-wide, so the chances of someone with an in-depth knowledge of the subject reviewing any given edit aren't great. This doesn't matter for edits that just straighten up formatting or fix spelling or attempt to introduce blatant spam... But anything more interesting stands a reasonable chance of getting botched.

If a bad edit is approved, the damage is limited - folks can always roll it back. But if a good edit is rejected, there's no guarantee anyone will ever suggest it again. If the editor doesn't check back and the original author isn't around anymore, that edit gets buried in the system and forgotten about.

So what can we do to improve this?

I don't know.

I suspect there's a place for some additional oversight here; the equivalent of an "editor in chief" role. But I haven't particularly liked any of the ideas I've seen toward this end thus far.

I wish there was a way to put edits in front of topic experts before deferring to folks in other areas of the site. Frankly, I wish this was the case for pretty much all of community moderation. But there's not very much support for that in the system, and building it isn't trivial. Right now, Stack Overflow is probably the only site that really needs something like this, so the appeal is fairly limited.

By and large though, I think the system works reasonably well. When I review reviewers, I'm generally pleased with their work, and with the improvements made as a result of it.

  • Thanks, Shoq9. A couple of follow-ups. The 33% rejected edits in the last 90 days is a little hard to believe and there isn't any link behind it. Is that figure really correct? Also, can you recommend a question to post additional ideas on? – Peter Alfvin Dec 7 '13 at 2:00
  • Same query for both of the last 2 bullets, @Peter - I've edited to clarify. And... Not without knowing what those ideas are. Find an existing problem-statement and address it, or post your own + idea. – Shog9 Dec 7 '13 at 4:08
  • “Matching up reviewers with topics they know well is hard”. Yes, but in my experience (and I know what I'm talking about), very few suggested edits require subject knowledge — it's not a big deal to hit “Skip” now and then. I found the close and flag reviews a lot more demanding of subject knowledge; without the tag filters I find the close queue nigh-unusable due to the amount of skipping. – Gilles 'SO- stop being evil' Dec 9 '13 at 0:39
  • Audits seem to have improved this slightly, but my informal feeling is that there are still a sizable portion of bad suggested edits that get accepted, and (especially on SO) a sizable portion of good suggested edits that get rejected. I don't have a good solution for that other than educating the reviewers — and I work in security so I am keenly aware how pie-in-the-sky this is. – Gilles 'SO- stop being evil' Dec 9 '13 at 0:41

You must log in to answer this question.

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