Wikipedia talk:Edit warring/Archives/2012/September

Definition of "revert"

"A revert means undoing the actions of another editor." Um, if I do copyedit to an article, that's technically or essentially "undoing" the text I've changed, put there by another editor previously. So this definition seems a bit inadequate. Seems like it should be important to have a good def of this term. Perhaps it has been brought up before (I've not checked archives). What is the thought on this? Thank you. Ihardlythinkso (talk) 06:43, 16 August 2012 (UTC)

Any edit that removes existing text, even a cleanup edit, technically meets the definition of a revert. Most admins would not consider spelling or grammar improvements to be a revert for purposes of 3RR. If you are doing heavy cleanup on a controversial article, you should probably space out your edits so there is no possibility of a misunderstanding. Grammar can be fixed tomorrow just as easily as today. If cleanup edits are consecutive, they count as at most one revert anyway. EdJohnston (talk) 19:20, 16 August 2012 (UTC)

By "copyedit to an article", I wasn't thinking spelling or grammar corrections. (And I'm not sure what "cleanup" means, but I'm guessing I wasn't thinking about that either.) I was thinking about copyedits in the standard sense, rewrites to improve an article's readability. That often entails many edits. (See how Malleus edits an article, for example.) All such copyedits "undo the work of others", technically. The definition doesn't seem to cover this at all. I'm wondering if the definition doesn't seem inadequate therefore. Ihardlythinkso (talk) 05:33, 17 August 2012 (UTC)

Yes, under that interpretation if you did 4 edits to ancient material on one article in a day you would be violating 3RR! And 0RR would mean that existing material could never be edited! North8000 (talk) 13:08, 27 August 2012 (UTC)

Copyedits can be well-intentioned but to some people they could appear to change the substance of the article. Imagine someone copyediting an article on a politican who is currently running for office. Anyone who does this kind of copyedit should keep an eye on 3RR. Particularly if they restore one of their own changes after someone else removes it. In actual practice no admin is going to sanction for 'benign' copyedits. There is no practical way of defining these, that I can see. "I didn't break 3RR because I was only copyediting." This could create a new class of disputes. EdJohnston (talk) 14:16, 27 August 2012 (UTC)

The simple solution would be to include the term "recent". The first change done on material that is months or years old is not a "revert". North8000 (talk) 17:48, 27 August 2012 (UTC)

That makes sense. But it wouldn't cover the case where many times a user will add good content, but it isn't expressed well at all, and e.g. a copyeditor will go in to rectify that, so it's incorporated smoothly into existing article, including possible relocation too. Myself, I'm conservative and copyedit for readability, it's a delicate thing and takes skill to not alter meaning *at all*, and that is always my goal. Even though the original contributor may like seeing his add integrated better into the article, it's still undoing his ("crappier") work, technically, and ... fell under current definition of "revert"?! (That thought lead just now to another: How about User:North8000's criteria bar of "not recent" added, but for "recent" content, have criteria based on whether the rewrite altered meaning or not? [I.e. if a recent edit is copyedited, but the copyedit did not alter meaning, then it isn't a "revert".] User:EdJohnston, when you mentioned no practical way of defining these, perhaps these ideas are getting to a more refined definition. But I think you're saying without refined definition it is an administrator's judgement, with a refined definition it may lead to disputes whether qualified or not? [That makes sense, but wouldn't it also help an admin by lessening the load, other things being equal?] These are just ideas!) Ok, Ihardlythinkso (talk) 18:50, 27 August 2012 (UTC)

What you say is true but a separate issue which more complex to try to cover. Mine is just to fix a more simple glaring problem. Sincerely, North8000 (talk) 20:09, 27 August 2012 (UTC)

Understood. And I think it should be done. (Does there need to be proposed text here, and consensus? Or?) Ok, Ihardlythinkso (talk) 00:19, 28 August 2012 (UTC)

If anyone wants to change the policy, it would help if you could find an example in the AN3 archives where a rule like this would have changed the outcome of the case. Since the definition of a revert has been heavily discussed in the past, you might need to convince people that this is a serious problem, and that the cure is not worse than the disease. EdJohnston (talk) 15:39, 28 August 2012 (UTC)

In addition to what Ed said, I would like to see the precise proposed text. I hate evaluating policy changes without something concrete.--Bbb23 (talk) 00:15, 29 August 2012 (UTC)

"Changed the outcome" is not a criteria. You should not have a rule that defines routine editing as a "violation". People who want to follow the rules should not be having to commit technical violations of the rules in this policy and then have to depend on them not being enforced. And again, as written:

• Any 4 edits to exiting material in any article is a 3RR violation as it is written. So, if you go thorough an article and make corrections or tweaks on ancient material in 4 areas you are violating 3RR.
• EVERY edit to existing material in a 0RR article is a violation.

Adding one word ("recent") to the definition would solve all of these problems. While that still leaves "recent" open to interpretation, it does clarify that an edit to ancient material is not a "revert" North8000 (talk) 11:32, 2 September 2012 (UTC)

So the proposal is to change?

"A revert means undoing the actions of another editor" --> "A revert means undoing the recent actions of another editor" (Second paragraph).

"that reverses the actions of other editors" --> "that reverses the recent actions of other editors" (The three-revert rule; second paragraph)

Adding to 3RR exemptions; 8. Routine copy editing

Jeepday (talk) 12:36, 2 September 2012 (UTC)

Thanks. My (our?) proposal would be the first 2. I think that the third item is not required (if this change is made) and introduces other complexities. So the proposal is to change:

• "A revert means undoing the actions of another editor" --> "A revert means undoing the recent actions of another editor" (Second paragraph).
• "that reverses the actions of other editors" --> "that reverses the recent actions of other editors" (The three-revert rule; second paragraph)

Thanks again. Sincerely, North8000 (talk) 12:43, 2 September 2012 (UTC)

I think "recent" is sufficiently vague as to open an avenue for argument and uncertainty; how long is recent? I think it's better to leave in place the current strict and clear wording, leaving it up to the judgment of the administrator(s) evaluating the 3RR report. I have to agree with EdJohnson and Ihardlythinkso above, on creating a new class of disputes and increasing the workload on administrators. I'd like to see some examples of where this has been a problem, the reports require a diff showing the previous version reverted to - if this diff is several years old, then it may not be a valid report, it would really depend on the circumstances. Dreadstar ☥ 14:32, 2 September 2012 (UTC)

IMHO a situation where normal, accepted editing (e.g the two common situations described above) is a clear violation of the rules as-written IS a problem. North8000 (talk) 15:01, 2 September 2012 (UTC)

I haven't seen it happen. On the other hand, I have seen an editor who engaged in disruptive edit warring which resulted in AN/I threads and blocks, come back weeks or even months later to once again make the same edit (revert) that caused so much disruption - starting the cycle over again. I've seen that several times, so admins do need the flexibility to nip something like that in the bud. If it's normal, accepted editing, then that should be clear to the examining admin, and if not, it should be clear to the admin(s) looking at the unblock request - who should then take the time to educate the admin who was mistaken. I can't see many situations where wording that includes 'recent' would have been helpful, if any at all - and I can easily see where it would have been harmful. Dreadstar ☥ 15:12, 2 September 2012 (UTC)

I agree with all of Dreadstar's points and would not support the proposed change.--Bbb23 (talk) 15:31, 2 September 2012 (UTC)

Thirded. Any admin worth a damn knows how to tell the difference between edit warring and fixing up an article. We don't need to define it down to the nth degree. If a particualr admin is unable to do this they are the problem, not the policy. Beeblebrox (talk) 18:02, 4 September 2012 (UTC)

new template for users blocked for edit warring

All to often when users, even fairly experienced ones, are blocked for edit warring they respond by claiming they were acting in WPs best interest. They pften cite some previous consensus, a desire to preserve NPOV, or some other policy that they (wrongly) believed shielded them being blocked. Since this is such a regular occurence I have crafted a template to reply to such statements at Template:ewblock. Modifications are welcome but please not that I deliberately did not put a box or any shading behind it in the interest of not making it look like all the warnings and block notices they will almost certainly already have on their talk page by the time it is appropriate to add this one. The goal is to give them a brief, friendly review of what edit warring is and why we block for it. Beeblebrox (talk) 17:57, 4 September 2012 (UTC)

My comments:

1. I like the idea.
2. I'd skip the apology in the first sentence and reword the second to be something like "Many users find themselves confused as to why they were blocked for edit warring ..."
3. I'd change the phrase "as far as the edit warring policy is concerned there is no right and wrong in an edit war" to "as far as the edit warring policy is concerned there is no right or wrong".
4. I'd eliminate "is wrong" from the next sentence.
5. There are other exemptions besides vandalism.
6. I'd change "We don't allow edit warring because it never helps resolve an issue, and it always makes it worse" to "We don't allow edit warring because it doesn't help resolve an issue; it only makes it worse."
7. I'm not crazy about the first bullet point (tagging). Some might perceive such a change in the midst of a battle to be a revert, kind of an oblique way of getting in one's version. I'd get rid of it.
8. I'd add a few commas after introductory phrases.

--Bbb23 (talk) 00:00, 5 September 2012 (UTC)

Additions

I know this issue has been touched on before, but I don't know if it's ever been fully fleshed out.

Definition: "Undoing other editors—whether in whole or in part, whether involving the same or different material each time—counts as a revert."

Some (many?) admins take the position that an "addition" is not can "undoing" (I've never liked the word "undoing" for that and other reasons - I'd prefer "changing").

Let's take two examples of disputed text. The base text is "Johnny is a good boy" and the following two changes:

1. "Johnny is a good boy. (fact tag)."
2. "Johnny is a good boy. No, he is not. (source)."

Is Change #1 okay under 3RR? Is change #2 okay? If change #1 is okay but Change #2 is not, what's the policy basis for the distinction, or is all of this based on administrative discretion? Is there a way to clarify the policy but still allow for discretion?--Bbb23 (talk) 14:38, 14 September 2012 (UTC)

Neither is unacceptible. In fact, #2 is consistent with wp:NPOV.