What guidance should be given when edits are rejected?

Question

I'm hoping to get rid of the "too minor" edit rejection reason, in favor of a more direct way of indicating edits that fail to significantly improve a post. We're also fixing to warn editors when their edits are rejected. Between the two of these changes, I'm thinking the other rejection reasons are gonna get a whole lot more scrutiny in the near future...

So it seems like this would be a good time to revisit the guidance they provide, both to reviewers and editors.

Here are the existing rejection reasons in order from most-used to least-used, and my commentary on them:

• too minor - This edit is too minor; suggested edits should be substantive improvements addressing multiple issues in the post.

As previously mentioned, this should be dropped as an explicit rejection choice, in favor of being implicitly chosen by the system in response to the submission of an alternate edit.

• invalid edit - This edit is incorrect or an attempt to reply to or comment on the existing post.

This is a great reason hobbled by a terrible name - it is trivially easy to find suggested edits rejected for this reason that are neither incorrect nor an attempt to reply or comment. In practice, folks appear to use this for edits that add information they don't understand; while it is possible that some of these should be rejected for other reasons, the guidance provided by this reason to the editor is then useless. This reason should be restricted to attempts to reply to the post's author, and named accordingly.

• radical change - This edit changes too much in the original post; the original meaning or intent of the post would be lost.

Changing the original meaning or intent is a great reason to reject an edit; changing "too much" is terrible guidance - as subjective as "too minor" if not more so. Usage reflects this, with the edits rejected for this reason ranging from those that completely replace the question/answer with one entirely different, to those that make comprehensive grammar corrections. Changing the meaning/intent is the cardinal sin here - the name and guidance should focus on this.

• vandalism - This edit introduces spam, defaces the post in some way, or is otherwise inappropriate.

Another excellent reason ruined by making it too broad. Drop the ending "or is otherwise inappropriate".

• excerpt not helpful - This edit does not follow any of our tag wiki guidelines and is unlikely to help instruct future visitors in the appropriate use of the tag.

The only problem with this reason (which only appears for tag wiki excerpt edits) is that you have to read a blog post to figure out what sort of "helpfulness" it's talking about. Robert Cartaino suggested an alternate wording that addresses this - we should use it.

• wiki not helpful - This edit does not follow any of our tag wiki guidelines and is unlikely to help instruct future visitors in the appropriate use of the tag.

Again, if I have to read a blog post to remember what sort of stuff I'm looking for in a wiki edit then I'm probably going to get it wrong. Needs some short examples of what's good/bad in a tag wiki.

• copied content - This edit plagiarizes content from an external source without proper attribution.

This is the least-used rejection reason, which is a shame because it should be getting used all the time on tag wiki edits! In fact, this is probably more important for tag wiki edits than the one directly above. Robert has some good suggestions for improvement here too - we should use them.

I'm not going to suggest my own changes to these just yet (apart from including Robert's); first, I'd like to hear your ideas - and your opinions on whether or not my criticisms merit changes at all. So, thoughts?

###Related: What are the rejection reasons for suggested edits?

Answer 1

Big thanks to everyone who offered feedback here! I've reviewed it along with a heapin' helpin' of custom rejection reasons, and the following reasons are now live on all sites:

• spam or vandalism (all post types)

  This edit defaces the post in order to promote a product or service, or is deliberately destructive.

• no improvement whatsoever (all post types)

  This edit fails to make the post even a little bit easier to read, easier to find, more accurate or more accessible. Changes are either completely superfluous or actively harm readability.

• irrelevant tags (questions only)

  This edit introduces tags that do not help to define the topic of the question. Tags should help to describe what the question is about, not just what it contains.

• clearly conflicts with author's intent (Q/A only)

  This edit deviates from the original intent of the post. Even edits that must make drastic changes should strive to preserve the goals of the post's owner.

• attempt to reply (Q/A only)

This edit was intended to address the author of the post and makes no sense as an edit. It should have been written as a comment or an answer.

• copied content (Tag wikis / excerpts only)

  This edit copies a significant amount of content from an external source. Generic descriptions such as encyclopedia articles and ad copy do not provide useful guidance; try creating something useful to this community specifically, and be sure to attribute the original author. See: How to reference material written by others.

• lacks usage guidance (tag excerpts only)

  Simply defining what a [tag] is rarely helps those using it unless the tag's name itself is ambiguous. Excerpts should describe why and when a tag should be used.

• circular tag definition (tag excerpts only)

  Tag excerpts amounting to, "[tag] is for questions about [tag]" are pointless and usually rejected. Excerpts should describe why and when a tag would be used.

• causes harm (all post types, replaces "custom" - prompts for free-text input)
  Describe how this edit would make the post worse.

What this will look like in most cases:

###Notes:

We're walking a fine line here between documenting what is expected of edits and trying to enumerate badness. I've attempted to focus the new reasons on the worst problems observed in suggested edits, while re-positioning "custom" as a more obvious choice in situations not listed. I originally included another reason, "polishes excrement" - but dropped it after realizing that it asked reviewers to determine the value not of the edit but of the post being edited. While I do believe these edits should be rejected in many cases, the reason for declining them resists a generic description - therefore if it does not fall into one of the other categories listed, my hope is that folks utilize the "causes harm" field to indicate why such an edit would be counter-productive.

Special thanks to Gilles, bjb568 and Care Bear for their suggestions - while I've used none of them verbatim, I've incorporated ideas from all of them.

Answer 2

This post is of very low quality and should be deleted. Please only edit posts that can be salvaged.

Answer 3

Yay, at last!

Questions and answers on the one hand, and tag wikis (bodies and excerpts) on the other hand, need different sets of reasons.

The guidance shown to reviewers and the guidance shown to the author of the suggestion needn't be the same. I'll propose wording for both.

For questions and answers

• attempt to reply — This edit an attempt to reply to or comment on the existing post.

[question] To request some clarification or improvement from the author of the question, leave a comment under the question. If you want to provide a solution, use the answer box at the bottom of the page.

[answer] To request some clarification from the author or provide constructive criticism, leave a comment under the answer. If you want to provide an alternative solution, use the answer box at the bottom of the page.

Drop the “is incorrect” bit, which conflated two completely separate cases (attempts to reply, and modifying the post in a way that is factually incorrect). The guidance should be adjusted if the author doesn't have the comment privileges. I'm not very happy with my name proposal, I can't think of a better one right now.

• incorrect edit — This edit introduces a factual error or invalid formatting.

This suggested edit introduced a factual error or invalid formatting.

In all these years I have sorely missed a way to clearly say “your heart is in the right place but your head isn't”.

It's more informative if the reviewer leaves custom feedback explaining how the added material is incorrect, but there are many cases where the feedback would be simply “the post was correct, read it and its references”. Leaving custom feedback should be just as optional as leaving a comment when downvoting an answer.

I'm open to ideas of something that encompasses factual errors, invalid formatting (common on Stack Overflow), and also rarer cases such as edits that introduce a grammatical error under the guise of correcting one.

• radical change — This edit changes the core meaning or intent of the post, or denatures its style.

[question] This suggested edit changed too much in the original post. It added or removed too much material, or did not respect the original author. When editing questions, make sure not to accidentally remove the problem that the question seeks to solve.

[answer] This suggested edit changed too much in the original post. It added or removed too much material, or did not respect the original author. If you think an answer is too far from being correct, vote it down and, if possible, vote up correct answers or write an answer of your own.

“Changing too much” is reasonable guidance to give to the suggestor, but I agree with you that it isn't good guidance for the reviewers. What matters is whether the core meaning changes, not how many characters have changed. I'm not dead set on suggesting downvoting here, maybe suggesting a comment would be better? or both?

• vandalism — This edit introduces spam, defaces the post in some way.

I'm not sure if we should expand on the guidance here. I've seen folks misuse the vandalism reason in the past; hopefully having a clearly labeled reason for incorrect edits will help in this respect.

We're getting a better mechanism for too minor, good riddance.

“Copied content” makes no sense on questions and answers. If it ever happens (which as far as I remember is 0 times in over 10,000 reviews), write a custom reason.

Tag wikis

• incorrect edit; vandalism

These two are just like questions and answers.

• markup in excerpt — Tag wiki excerpts do not support any form of markup.

For excerpts only. I want a predefined reason for that, to alert reviewers that they are reviewing an excerpt and provide this information to many an uninformed editor.

• copied content — This edit copies content from an external source. It is not helpful or lacks attribution.

All content copied from an external source must be clearly marked as a citation and properly referenced. Generic descriptions such as encyclopedia articles and ad copy do not provide useful guidance regarding the tag as used in our community.

I'm not sure about this one: should attributed Wikipedia copypasta be lumped with unattributed copypasta or with “not helpful”?

• misguided — The proposed text does not adequately describe the meaning of the tag in our community.

The proposed text does not provide useful guidance regarding the tag as used in our community.

No, that's not a good name; I can't think of a better one at the moment. I dislike both the current name “not helpful” and the “RTFMBP” guidance.

I don't like Robert's proposed wording “lacks usage guidance”, because that's far from covering all cases. Some tags are about a precise concept (e.g. a specific product) and require no guidance. Some edits lack usage guidance because the appropriate guidance is already present. Conversely, an edit that says “the tag frobnication is for questions about frobnication” does provide usage guidance, but that doesn't make it helpful.

Do give this reason the same name for excerpts and bodies. I find it very annoying that they're currently in a different location because of the alphabetical sorting.

The blog posts currently used for wikis and excerpts isn't the best reference for this and the copied-content-with-attribution reason. It mixes two aspects: guidance on writing a good tag wiki, and explaining what the new excerpt feature is about. The blog post link should be replaced by a link to a new help center page that focuses on giving guidance in writing tag wikis and excerpts (a single page should cover both, as there is redundancy between the two).

I don't think we need anything like “radical change” for tag wikis. Removing useful material happens occasionally, but isn't very common. Completely rewording or reorganizing a tag wiki is not intrinsically bad.

Answer 4

How about renaming "invalid edit" to "not an edit", similar to how "not an answer" is defined?

The only other concern I have right now is that I believe "copied content" is underused because it's hard to know when that's the case. Renaming/clarifying it isn't gonna make a huge difference since the core problem is that you basically have to go to another site (Google!) and search for a phrase to see that the excerpt is plagiarized (y halo thar wikipedia!).

We should put more thought into that... Considering how exact the copies we get are, maybe we can do something to just check if Wikipedia has an article containing this exact text. Then again, maybe not. Either way, while it'd be nice to have a clearer description of the rejection reason that doubles as a call to action here, I don't think that's going to do much for solving the underlying issue.

Answer 5

Here's one I would love to see specifically addressed:

Adds excessive or unnecessary formatting.

It seems a little too common to see suggested edits that look like this:

how do I foo the bar in baz while

• standing on one foot

  1. on friday afternoon

                                          in
     

iceland?

Inspired by: Reject an already-approved suggested edit when rolling it back

Answer 6

I think either the too minor reason should be introduced back or the language of no improvement whatsoever reason should be softened (or the title of the custom box may be changed). Because currently, users who made sincere suggestions with some (but not significant) improvement get a very hostile return. Note that there are no other option to fit this situation -- i.e., the edit causes no harm to the post and improves it a little.

This matter has been raised in here with a list of other examples and voting results support the above claim and the explanations given for the rejection of the request are not satisfactory.