Wikipedia talk:Edit warring/Archives/2011/April

Can a single revert make an editor part of an edit war and sanctionable for edit warring?
If there is an edit war ongoing, and an editor who has not previously (or recently) edited the article makes a (single) revert to it, can he be punished for edit warring? PS. This question is related to an ongoing AE report, see. There an administrator states that "an editor is also engaged in (and therefore sanctionable for) edit-warring if they contribute a single revert to an edit war that is otherwise carried on mainly by others.". If this is indeed a case, I would very strongly recommend adding this wording to EW as well as WP:BRD. For the record, I disagree with such interpretation of EW/BRD. --Piotr Konieczny aka Prokonsul Piotrus 18:13, 16 March 2011 (UTC)


 * Eh, it's not just that. The four edits (on four different articles) at issue where all made before the edit war started. In one case, the edit was made six months before other people started edit warring on the article. It's simply ridiculous. . On the other three they were a simple straightforward BRD reverts, and only afterward did other people start edit warring on it - I didn't and I ceased making any further edits to the article. Why am I being blamed for what other people did?Volunteer Marek (talk) 18:24, 16 March 2011 (UTC)


 * Guilt by association/collateral damage do come to mind... more seriously, I always thought that an editor was "edit warring" if they made more than one edit in a reasonably short period of time (1-2 days for regular edit warring, <1-2 weeks for "slow edit warring"). "1R" as in B"R"D or 1"R"R was, I believed, "the good thing". Seeing an editor called an edit warrior for a single edit to an article (not to mention an editor who was engaged in much constructive talk discussion) is something quite new, to me at least. --Piotr Konieczny aka Prokonsul Piotrus 18:27, 16 March 2011 (UTC)
 * Here is my understanding of the policy. According to WP:edit warring, when one arrives to a new article and see an edit war between other people ("the situation develops into a series of back-and-forth reverts"), he should not revert in support of any of the sides. However, I have seen a lot of people doing just that, and all of them thought they are not engaged in edit war, exactly as Marek. In "BRD cycle" someone suppose to revert your change, but he is not at fault if there is no ongoing edit war. However, if your previous change was revert, and he reverts you back, he is engaged in edit war. P.S. But if you arrive to a new article, which is not in a state of ongoing edit war, and revert to an earlier version, that does not make you involved in edit war. For example, this is not edit war, but this is edit war. Hodja Nasreddin (talk) 23:32, 16 March 2011 (UTC)
 * But I may be wrong. According to WP:Consensus, When reverting an edit you disagree with, it helps to state the actual disagreement rather than citing "no consensus". Hence, when a new editor comes to a new page, he may revert for the first time merely to state his disagreement and help to establish consensus. Is not it? Hodja Nasreddin (talk) 23:54, 16 March 2011 (UTC)
 * Indeed, a single revert is hardly not constructive - it indicates editor cares about the article, and also makes edit warrior takes note (because some of them may be ignoring talk page discussion). It is only when one makes 2+ reverts that they become involved in an edit war (as the policy states that edit warring is when editors "repeatedly override each other's contributions"). An editor who does not make more than one edit obviously makes no "repetitive overrides". --Piotr Konieczny aka Prokonsul Piotrus 19:42, 17 March 2011 (UTC)
 * This sounds simple and logical to me. Then we should probably write this down more clearly in this policy page, unless someone objects.Hodja Nasreddin (talk) 02:36, 18 March 2011 (UTC)
 * I generally agree, although a lot depends on the good or bad faith of the editor - for example, if they know that valid arguments against their edit have been made on the talk page, and they just come along and make the edit without addressing those arguments, that could be viewed as disruptive. In some situations, though, a new editor's arrival can actually resolve an edit war, if they are able to make a compromise edit to produce something that's (at least temporarily) satisfactory to all sides. --Kotniski (talk) 07:47, 18 March 2011 (UTC)
 * Since I hear no objections, I suggest you try to be bold and clarify the issue, we can see what happens (unless you guys think we should RfC this entire issue first?). Kotniski, I agree with you, but if the editor has ignored comments at his talk, it is easy to see they were being disruptive. Of course, there can also be the case when editors make token effort on talk, participating in a discussion in various unconstructive ways... but the main point here is to ensure there is no edit warring (but at the same time, allow the editors to make at least one edit in the article without being surprised by sanctions). In the ideal situation, the editors reach consensus on talk, aided by DR if needed, and slowly edit the article, improving it, without many reverts which would destabilize the situation (but an occasional revert per BRD is nothing bad). --Piotr Konieczny aka Prokonsul Piotrus 15:38, 19 March 2011 (UTC)


 * I think it is already there. Current version tells:
 * Making only one revert in an article is expected from an experienced editor who does not want to be involved in edit wars. It also tells:
 * Hence supporting one of the sides by a single revert is fine according to the policy, unless you were explicitly asked by one of the sides to revert. Hodja Nasreddin (talk) 15:16, 20 March 2011 (UTC)

I think that "an editor can be considered as engaged in (and therefore sanctionable for) edit-warring if they contribute a single revert to an edit war that is otherwise carried on mainly by others", provided that this editor has been properly informed about an ongoing edit war. That means that the mechanism of informing of new editors about ongoing edit war should be more clear and unambiguous. In connection to that, I have one suggestion. I propose to develop a special type of a message (for example, in a form of a template) that informs an editor that the article is currently is a subject of an edit war between two users/groups of users, and, therefore, the danger exists that by making his/her edits this new editor will join one of these party. Such a template is supposed to be added and removed only by administrators; it also can be automatically removed after some period (e.g. one week) by bots. If the edit war does not end for this period, administrators may re-insert the template again. The obvious advantage of this measure is that all users are informed that by joining a partisan edit war they cannot use their "personal" 3RR or 1RR limit, and a "collective" 3RR limit works instead. As a result, a possibility of gaming the 3RR rule does not exist any more in the articles that have been labeled as an edit war battlefield. In addition, such a message would prevent new good faith editors, who are unaware of the ongoing edit war, from joining the war by accident, and, accordingly, from being punished for that. In other words, any editor can make three reverts in the article that has no edit war template, and they can be sure they will not be sanctioned for that. However, if the edit war message is added to the article, that must be interpreted as a serious warning against making any change without familiarising themselves with the article's history and the talk page discussion.--Paul Siebert (talk) 18:40, 19 March 2011 (UTC)
 * I think this is a very constructive and reasonable suggestion.Volunteer Marek (talk) 18:46, 19 March 2011 (UTC)
 * Indeed. although your last sentence refers to "three edits" - note that we were discussing making a single revert. It is not uncommon for admins to sanction editors for three reverts, and it is not that difficult to argue that an editor who has made multiple reverts is edit warring. It is much harder to do so with a single editor. I like the idea of admin-added warning template that would make any reverts sanctionable, although it should be removed from an article after a period of inactivity automatically by a bot. Presumably, if such a template has not been added, single reverts would be not sanctionable, yes? --Piotr Konieczny aka Prokonsul Piotrus 15:31, 20 March 2011 (UTC)
 * Absolutely. Do you think we have to initiate RfC, or to take other steps to obtain a community approval for creation of such a template and for modification of the policy accordingly?--Paul Siebert (talk) 17:43, 23 March 2011 (UTC)
 * I think the template warning idea is a good idea. Maybe all we need to do is modify Template:Uw-3rr, which is already integrated into Twinkle and such. At Template:Uw-3rr/Sandbox, I've created a draft version that also advises: "If you continue a series of reversions started by others, you may be considered to be edit-warring even if you make only one or two reversions." After receiving such a warning once, editors should be considered warned that "team edit-warring" is just as sanctionable as individual edit-warring.  Sandstein   18:56, 23 March 2011 (UTC)
 * I think your draft looks good and could be implemented. Keep in mind though that this is a *new* addition to the policy and the template, and as such a grandfather clause should apply :). Previously it was not the customary, and certainly editors were not aware that this was the standard, way of interpreting the policy.Volunteer Marek (talk) 18:59, 23 March 2011 (UTC)
 * Hm, I don't think it's new policy, only perhaps a new application of existing policy, but that's a discussion for another page...  Sandstein   23:57, 23 March 2011 (UTC)
 * I probably misunderstand something, but my proposal was different. The template (something like this User:Paul Siebert/SandboxEW) is supposed to be added to the article (to the article's talk page), and it is supposed to inform all users who work on this article, not only those who are edit-warring. Therefore, the first sentence ("You currently appear to be engaged in an edit war") is too aggressive. It is added by admins, and it serves as an indicator that even a single revert may result in block if the group of editors exceeded the 3R limit. I do not know if it is technically feasible, but it would be good if this message was visible for anyone who tries to edit the article.--Paul Siebert (talk) 00:31, 24 March 2011 (UTC)
 * One appears to be for articles, and one for editors. They're not mutually exclusive.Volunteer Marek (talk) 06:28, 24 March 2011 (UTC)
 * Yes. Your proposal, Paul Siebert, states that "the standard three-revert rule is currently not in force for this article". It very much is, editors are still forbidden to make more than three individual reverts in 24 hours. The difference is that they are notified that they may also be blocked for edit-warring, rather than 3RR, if they continue the edit wars begun by others.  Sandstein   06:32, 24 March 2011 (UTC)
 * I would say, the difference is that, whereas your template affects concrete users (so the admins are supposed to notify each of them separately, the newcomers are not aware of the restriction, and this restriction doesn't work for them), the template I suggest is applied to the article as whole, and the users become notified automatically when they start to edit the article. As a result, no additional admin's actions towards any concrete user are necessary: the first user who exceeds a "collective" 3RR limit will be reported by another party, and, based on provided diffs any admin can almost automatically block them. Another advantage of such a template is that it makes canvassing impossible: any formally ununvolved user is automatically warned immediately after he comes to the talk page (or opens the editing window). I probably did not fully understand your proposal, however, the need to notify the users personally seems to be its major flaw. In addition, to notify the users who haven't yet committed any violation of the policy (3RR has not been exceeded by them yet) is against WP:AGF, and I don't think such a notification to promote WP:LOVE. --Paul Siebert (talk) 10:48, 24 March 2011 (UTC)
 * PS. Re the standard three-revert rule is currently not in force for this article"" I meant that the individual 3RR is not in force, and the collective 3RR works instead ("That means that even a single reversion may be considered to be edit-warring if you continue a series of reversions started by others. "). If that is unclear from the text, I'll try to clarify it. If you have any idea how to clarify it, feel free to edit this template draft directly.--Paul Siebert (talk) 10:55, 24 March 2011 (UTC)

Implementation?
I think some valuable ideas where proposed, but I think the discussion died before we implemented them. This is my proposed list of changes we should implement: If we are all fine with the above direction, I'd suggest we proceed to implement it. --Piotr Konieczny aka Prokonsul Piotrus 18:27, 15 April 2011 (UTC)
 * add a clarification to this and WP:BRD along the lines suggested by Paul above (to quote him: "an editor can be considered as engaged in (and therefore sanctionable for) edit-warring if they contribute a single revert to an edit war that is otherwise carried on mainly by others", provided that this editor has been properly informed about an ongoing edit war.") and Sandstein's (from template) "If you continue a series of reversions started by others, you may be considered to be edit-warring even if you make only one or two reversions.". Some other changes may be necessary to clarify issues and remove contradictions. In particular, we should make it clear that if the article is not templated and/or the user has not been warned (for that article), the user should not be blocked for making 1-2 reverts. I'd also suggest clarifying the user warning (Sandstein's), to make it clear that the editor might not have necessarily edit warred due to his/her own edits, but rather, due to joining (perhaps without even noticing it) a larger edit war. This is to avoid confusing this template with a "oh, this user has been warned about edit warring because they likely made many reverts" warning. This template should be more of "you made just 1 revert but note that this article is in a middle of an edit war and any further reverting may make you as sanctionable as others".
 * mainspace the Template:Uw-3rr/Sandbox and User:Paul Siebert/SandboxEW
 * Support. In my my opinion, Piotrus gave a correct shape to what I proposed.--Paul Siebert (talk) 01:03, 16 April 2011 (UTC)
 * Comment. Still considering after reading over several times. One concern I have is that even with templates and all, if editors stray into a conflict that the first warning in one of these cases be an advisory, not an official warning (i.e., strike one). Otherwise we risk institutionalizing a practice which assumes bad faith. P ЄTЄRS J V ►TALK 01:31, 16 April 2011 (UTC)
 * The existing policy is very much clear that a single revert by someone who did not edit the article for a long time does not constitute edit warring, no matter what others do (see quotes above). If you want to make this even more clear by modifying text of the policy, I do not have any objections. As about a warning template, any individual administrator can do it already in the areas of discretionary sanctions.Hodja Nasreddin (talk) 04:13, 16 April 2011 (UTC)
 * The existing policy still provides an opportunity for gaming a system in the same way as that had been done more than a year ago (do not force me to specify what concretely did I mean). A new proposal decreases such an opportunity, and therefore decrease the risk of these edit wars. --Paul Siebert (talk) 06:54, 16 April 2011 (UTC)
 * Yes, edit war is damaging. Yes, it must be restricted by rules. But we need very simple and clear rules to be enforced exactly. For example, 3RR is a good formal rule, and it should be properly enforced. It means: do not punish people who make only one or two reverts per article per day. If this rule is not enough, that's fine. Make it 2RR or 1RR rule (per article per day). But it means: do not punish someone who makes only one revert per article per week. Yes, I know, this is different now: anyone can be arbitrary blocked for any number of reverts, especially in an area of discretionary sanctions. I think this needs to be changed. But if this is not changed, there is only one practical solution: try to follow 0RR in the areas of discretionary sanctions, and do not be surprised when someone reports you to AE if you do not. Please do not interpret this as an argument in favor of edit warring. I personally was not engaged in a single edit warring episode during last year. . Hodja Nasreddin (talk) 00:16, 17 April 2011 (UTC)

Mmmmm... I'm having some serious second thoughts about it. The original intent of the proposal/discussion was sort of "in practice, you can get sanctioned for making a single revert, if other people have been edit warring over something, but this is not stated in policy, so we should clarify policy". But it seems like now we're forgetting that this "practice" ... is not stated in policy! In other words, perhaps "practice" should change to match policy rather than vice versa. I realize that it's easier to just change a few words on a Wikipedia page than to change practice, but easier doesn't mean better. I'm also coming around pretty strongly to the view that rules should be clear and transparent and, importantly, they are there just as much (if not more!) to protect regular editors from admin abuse (i.e. practice not following policy) as they are to give admins new cudgels to beat the peasants with (and yes, of course there are some peasants who well deserve it).Volunteer Marek (talk) 05:49, 23 April 2011 (UTC)
 * Well, first of all, we have to clarify that you cannot get sanctioned for a single revert. Second step is to allow admins to issue warnings after one revert, which I think is somewhat reasonable. I am not sure where we disagree, exactly? --Piotr Konieczny aka Prokonsul Piotrus 09:06, 23 April 2011 (UTC)
 * I am not sure I agree with the idea that noone can be sanctioned for a single revert. I would say, this works just for the first time, and as soon as the user has been informed that he has joined the edit war, he is supposed to be careful when he works on both this particular article, and on all other articles.
 * Another solution would be that the warning is supposed to be issued in each concrete case, however, that is being done not by admins, but by regular users, including the participants of the dispute. Upon getting the warning, the user is supposed to immediately self-revert, otherwise he can be reported by anyone, and subsequently blocked. Of course, a quick self-revert should be considered as a sufficient measure to prevent any sanctions. It is also necessary to specify that systematic misuse of such warnings may lead to sanctions as well: it is necessary to demonstrate that the revert had continued the chain of reversions started by others, and the total number of reversions exceeded the 3RR limit. No warning can be issued to the user who made unrelated reversion: for instance, if the subject of the edit war is the section "A", any reversion in the section "B" can hardly be a reason for sanctions, unless a strong contextual linkage has been demonstrated.
 * Taking into account that I am a proponent of providing of regular users with more rights, the second variant (a warning issued by non-admins) is more preferable for me.--Paul Siebert (talk) 19:04, 24 April 2011 (UTC)

Modification
Upon meditation, I came to a conclusion that 3RR in general probably needs in a modification. Most edit wars between two groups of users can be avoided if we simply expand this rule as follows:
 * "Please, be advised that the three-revert rule refers to a viewpoint, not to a user, so even a single reversion may be considered to be edit-warring if you continue a series of reversions started by others. In this case you may be blocked from editing without further notice. In connection to that, it is strongly advised that the user carefully read recent article's history and the article talk page, to make sure that the article they work on is not currently a subject of an edit war, and, therefore, there is no danger to join this edit war by accident."

In my opinion, most edit wars I was a witness of could be avoided had the policy been written in such a way.--Paul Siebert (talk) 14:49, 24 March 2011 (UTC)
 * This is a significant change of existing policy and probably contradicts WP:Consensus. Consider a common situation when five editors support one viewpoint and one editor supports another viewpoint. Such change would actually increase edit warring. Besides, a "revert" means undoing work by other editors, rather than supporting any viewpoints. A revert is not necessarily related to viewpoints.Hodja Nasreddin (talk) 17:08, 24 March 2011 (UTC)
 * Since a revert is currently defined as "any edit (or administrative action) that reverses the actions of other editors, in whole or in part, whether involving the same or different material.", a modification of another's text is also a revert.With regard to WP:Consensus, I see no contradiction, taking into account that consensus is defined as "a decision that takes account of all the legitimate concerns raised." As far as your legitimate concern has not been addressed, you always can insist on removal of the material that was added before consensus was achieved. By contrast, the present 3RR is in contrast with WP:DEMOCRACY, because in a situation when, e.g., 2 editors support one POV and 3 editors support the opposite POV, the last POV will inevitably prevail when all users strictly stick with the the 3RR policy in its present form. That is, therefore, a hidden form of poll, which is prohibited by policy. In addition, present policy encourage canvassing, which can be easily done taking into account the present level of communication technologies.--Paul Siebert (talk) 18:35, 24 March 2011 (UTC)
 * Counting several editors as one (as you suggested) contradicts WP:Consensus. Polls are not prohibited, but merely non-binding. WP:DEMOCRACY tells that consensus must be established by editing, and reverts (rather than edit warring) are a part of the legitimate process.Hodja Nasreddin (talk) 19:27, 24 March 2011 (UTC)
 * I see no logic here. Polls (in their open and clear form) are useful as a tool to avoid edit wars. What you propose, however, is to leave an opportunity to establish a (non-binding) community opinion via the de facto edit war. I don't think it is serious. Let me also point out that, had the rule proposed by me worked two years ago, there would be no need in creation of the notorious mailing list, or, more precisely, there would be no need to punish its members because they would be physically unable to game the system by using this list.--Paul Siebert (talk) 22:16, 24 March 2011 (UTC)
 * I did not propose any changes in the policy. You do. I am only telling that not every revert is edit war. Right?Hodja Nasreddin (talk) 02:48, 25 March 2011 (UTC)
 * Absolutely. However, if I start to work on some article and make a revert that is essentially similar to what have already been made twice by others I thereby support one party in the edit war. Of course, I can be absolutely uninvolved editor, but I equally can be an editor who monitored the edit war without joining it, and finally decided to tip the balance. Currently the policy absolutely allows that, and the interference of admins is needed each time when such an edit war between two parties starts. A possible solution would be to add a special restriction to the articles which are the edit war battlefield (see above), so it will be a user's responsibility to make sure that their actions do not lead to an escalation of an edit war, because in this situation he is not protected by 3RR any more. However, upon meditation I realised that nothing prevents us from expanding that rule to all articles, because for the articles where no group edit war is being waged this new rule is indistinguishable from the normal 3RR rule. Therefore, I simply see not problem with that modification.--Paul Siebert (talk) 03:15, 25 March 2011 (UTC)
 * Yes, this is not a good idea to edit any article in a state of active edit war between multiple editors. This is waste of time, if nothing else. I tried it a couple of times (e.g. in terrorism article) and will never do it again. But counting edits by several people as one contradicts WP:Consensus.12:48, 25 March 2011 (UTC)
 * As I already said, I do not see how does it contradict to WP:CONSENSUS. Please, explain.--Paul Siebert (talk) 19:21, 25 March 2011 (UTC)

I agree with the thrust of this proposal, but would simplify the change as follows:
 * "Edit wars disrupt consensus-based editing, no matter how many editors participate in them or how many reverts each editor makes. For this reason, even a single reversion may be considered to be edit-warring if an editor continues a series of reversions started by others. Before making a reversion, editors should therefore read the article's history and its talk page to make sure that they do not continue an edit war."

I don't think that it is helpful to invoke 3RR here, as 3RR is a special case of edit-warring that involves one user making threee reverts in a row - which is exactly not the sort of edit-warring we seek to address here. Rather, the edit-wars at issue here are the sort of wars in which there is a month-long slow back-and-forth over an article by many people, such as the EE naming disputes some here are familiar with.  Sandstein  06:09, 18 April 2011 (UTC)
 * Then to be consistent, you need to also change this part: "if a revert is necessary, another editor may conclude the same and do it (without you prompting them), which would then demonstrate consensus for the action", which is a very common practice. I think any policy changes are fine as long as they are clear and not open to gaming and interpretations. This new version suggests that anyone can be blocked/sanctioned for a single revert in an article he never edited before, on the discretion of a blocking administrator, which is kind of arbitrary. As a practical question, do you and other administrators want users reporting each other at AE for making single reverts in articles? Hodja Nasreddin (talk) 16:18, 18 April 2011 (UTC)
 * I don't want anybody to be reported, but I will certainly block for a single revert if that revert is the tenth in a chain of discussion-less reverts. I will not block for a single revert if that revert establishes a new consensus formulation arrived at via discussion. So, as always, it depends on the circumstances, but yes, a single revert may be part of an edit-war and therefore blockable edit-warring.  Sandstein   18:27, 18 April 2011 (UTC)
 * Consider an editor who went through many different articles (some of which where a subject of edit warring in a couple of last years) and made significant changes in each of them without talking. No one edited these articles for a long, so there is no one to talk about it at the moment ( this is my usual working mode, no matter if I do something in Biology or History ). Many of his changes are at least partial reverts to an older version, as usual with significant changes. Would that be something sanctionable? The answer is probably: "it depends" if the editor had a history of problems. Is that what you mean?Hodja Nasreddin (talk) 19:22, 18 April 2011 (UTC)
 * Agree with Hodja, what started out as a potentially constructive suggestion has degenerated into a "I can block anyone anytime for almost anything" (or to put it in other words "depends on the circumstances") arbitrariness and a bit of a ... power grab. The rules on edit warring and 3RR are there as much to restrict admin abuse and protect editors, as they are there to prevent edit warring.Volunteer Marek (talk) 18:34, 18 April 2011 (UTC)
 * This is actually important. Consider a hypothetical new user who reads this policy page. He thinks: "all right, four reverts are forbidden, so I will make only three". Then he is suddenly blocked for two or three reverts. He thinks: "all right, I will make only one revert per day". But someone collects diffs with his reverts in different articles and submits them to AE, and this user is sanctioned. It would make a lot more sense if he could read from the very beginning that we have de facto 0RR rule. But then we probably need a community discussion to replace 3RR rule by 0RR rule. Once again, it does not matter for me, because I know already that we have de facto 0RR rule. But one can not edit articles, especially in difficult areas, without occasionally making at least partial reverts. Hence we need at least 1RR but not 0RR. Hodja Nasreddin (talk) 20:15, 18 April 2011 (UTC)

Please correct me if I am wrong, but I see the discussion on this ("Modification") section as something separate from the discussion in the above ("Implementation?") section. This is why I split this ("Modification") section from the general thread in the first place. --Piotr Konieczny aka Prokonsul Piotrus 23:13, 18 April 2011 (UTC)
 * @ VM and HN. I am sorry, but you seem to misunderstand the essence of this proposal. Noone speaks about the admins' right to block anybody for anything. However, this proposed rule will force the users to be responsible, which means, among other things, familiarising oneself with the article's history and with ongoing discussion about the article's content before making reverts (addition of one's own text is not a revert, as a rule). If someone starts to work on the article, s/he is supposed to make sure that the article is currently not a battlefield of the edit war, and, accordingly, that by starting to change the text s/he would not continue a chain of reverts started by others. Obviously, that applies no restrictions on the total amount of reverts made by this particular user in all articles s/he is working on. Moreover, since both my and Sandstein's proposals clearly speak about a series of reversions started by others, any revert that does not continue a series of reversions started by others cannot be a reason for sanctions. An admin, or a reporting user, is supposed to demonstrate that continuation of the chain of the reversions did take place, and the burden of proof should be on those who applies sanctions, or who requests for sanctions. Moreover, in my opinion this new rule would make some other restrictions, e.g. 1RR and similar restrictions, redundant: any edit war becomes not a war between N users sharing one POV and M users sharing another POV, but between two viewpoints, and the number of the reverts made by each separate user becomes restricted naturally. In other words, this rule does not add new restrictions, but makes existing restrictions more logically consistent: since Wikipedia has no authorship, we can speak about an edit war between viewpoints, not users, and, accordingly, 3RR should be applied to the formers.--Paul Siebert (talk) 03:41, 24 April 2011 (UTC)

Codifing what constitues a revert
Hi, I've noticed there are many instances where whether an edit is a revert or not is very murky, so I have begun to codify what is a revert here. I of course need consensus to introduce the codifing officially so I ask that interested editors join me to fill in the question marks in the examples I have written out. As I will be posting this elsewhere as well, please talk on the talk page of the article. Thank you, Passionless   -Talk  04:09, 20 March 2011 (UTC)
 * I think that's a waste of time. It has been raised here repeatedly, and essentially there is no consensus as to what constitutes a revert in general. It only becomes evident when it's repeated. Tijfo098 (talk) 17:34, 24 March 2011 (UTC)
 * (Admins at WP:AE don't implement the policy as written, and the policy itself is self-contradictory between the general definition it gives and its 0RR case/example.) Tijfo098 (talk) 17:39, 24 March 2011 (UTC)
 * To give you a concise example of the difficulty in providing a definition, assume this is year 2000 and the article on Earth is blank. The 1st editor arriving writes: "The Earth is flat". The 2nd editor adds the word "not", as in "The Earth is not flat". The 2nd editor has reverted the meaning of the 1st editor's edit without formally changing any of its words, although it has altered a sentence (And you can even avoid that with a little creativity, e.g. "Only stupid people say 'The Earth is flat'".) Is this is a wikirevert or not? The attempt to define revert syntactically (as in who added or who deleted text) is obviously hopeless in a natural language. Tijfo098 (talk) 18:02, 24 March 2011 (UTC)
 * I did have a table on that point where I considered an obvious change in meaning to be a revert, but I deleted that chart as it was least important and most confusing. Passionless   -Talk  18:18, 24 March 2011 (UTC)
 * I have limited my scope since I posted that message to just trying to work on how to count reverts when they are made sequentially. If you make say four reverts to an article with no edits inbetween, you have of course only made one revert. But lets say in the middle of your four reverts someone else made a slight grammatical change, well now you have a problem because half the admins will count your reverts as one, will the other half will count you as making multiple reverts. The latter clearly goes against written policy as I have argued before-""A series of consecutive saved revert edits by one user with no intervening edits by another user counts as one revert." at WP:EW. To intervene means to "Come between so as to prevent or alter a result or course of events.", so as grammatical/minor edits do not "prevent or alter a result or course of events", this statement means that my three reverts should be counted as one regardless of the grammar edits inbetween them," I think we need to make admins all come at this situation in the same way so editors can know what is and what is not allowed. Passionless   -Talk  18:25, 24 March 2011 (UTC)

If I'm allowed a short lecture in ternary logic here, the above is an example of transition from "don't know" to "true" and then to "false" in (Kleene's) K3. You may be tempted to define revert as cycling back to a previous logical state ("don't know" in this case), but this is clearly not a satisfactory definition. Natural languages are even more complex than that. Often POV warriors are very aware of this (even if just intuitively), so they will "copyedit" a sentence or paragraph to alter its meaning. Tijfo098 (talk) 18:29, 24 March 2011 (UTC)
 * Changing the meaning of a sentence is of course a revert- you can look at it as if they deleted what was there(a revert) and added something wholly new. Passionless   -Talk  18:37, 24 March 2011 (UTC)
 * By that argument, at the limit, any change to text already written is a revert because it may alter its meaning if ever so slightly. Ergo, a paradox. So you cannot operationally define revert the way you propose. Tijfo098 (talk) 19:53, 24 March 2011 (UTC)
 * But you can make unlimited change to an article and it is only one revert, and only starts to count as multiple reverts once your edits start to undo the non-minor related edits of other editors who have edited since your reverting began. So this definition would not inhibit work, it only inhibits revert wars. Passionless   -Talk  20:02, 24 March 2011 (UTC)
 * Unfortunately, that's not sufficient. The policy does not require reverts to be of the same issue to count as reverts for XRR purpose. So, with your definition, two editors (A & B) making four edits each within 24hrs, but interleaved (ABABABAB), both break 3RR, regardless of the substance of their edits. Where the counting starts and where it stops are the key issues. As you define revert, 0RR makes no sense (as you would never be allowed to change anything if under that restriction), however other editors, admins in particular, see it differently; see the archives. Ergo, no consensus on this. You need to define some notion of semantic correlation (between edits) for your proposal to make sense in practice. This is a definable concept in sufficiently restrictive settings, search google scholar. But it's just another academic topic of no importance here because we have biographies for all the pop stars we see on teevee. Tijfo098 (talk) 20:15, 24 March 2011 (UTC)
 * The substance is important, B could have done several things that would not have interferred with the sequentialness of A's edits. B could have fixed grammar or made other minor edits that do not obviously change the meaning of the text, or B could have made 4 'reverts' to a section untouched by A, and both editors would have made only a single revert as none of their edits were to undo the edit of another editor since their reverting began. Passionless   -Talk  20:57, 24 March 2011 (UTC)

I'm glad I've convinced you. Your proposal however, does not capture that important issue&mdash;not in any concise way that could constitute a practical rule. If the 2nd part of your page (giant table) is part of the proposed rule set (and not merely a discussion/example of the fist), then see combinatorial explosion and WP:CREEP. Tijfo098 (talk) 23:34, 24 March 2011 (UTC)
 * I'll have you know that I enjoy combinatorial explosions, and that there was no change in my thinking between my previous two comments; one can make unlimited change to an article and keep making reverts even after someone has reverted their work, as long as they are not undoing the edits of others since they began to edit said article. As for my proposal I have given up on that and now only care about getting all admins to understand the policy, as it refers to sequential reverts, as many admins do not understand that the only way someone can make more than one revert is if they undo, in whole or part, the work of another since their first revert took place. Passionless   -Talk  17:09, 25 March 2011 (UTC)
 * My experience is somewhat limited, but it seems to me that there are quite enough edit wars that are simply a matter of pushing the "Revert" button. http://en.wikipedia.org/w/index.php?title=Tempo_%28chess%29&action=historysubmit&diff=423653213&oldid=422549298 This is a great example of such a case: the 18 edits in a one-week timespan equate to exactly nothing. The article goes back and forth five times (ten reversions total, or nine reversions and one initial edit) with two editors on one side and two editors on the other. I think enough edit wars follow that pattern that a "revert" can be simply defined as a "revert." The first edit was a good faith attempt to add to the article; the first revert a good faith attempt to undo what the other editor saw as damage. The second revert (re-inserting the first edits) is time to take the issue to the talk page; the third revert (removing the edits AGAIN after the author reinstated them with comments) is when it's obvious that they're at war. When the fourth, fifth, ..., ninth and tenth reverts occur... well, that's just plain ignorant.
 * My point is, that there seems to me to be enough blatant, obvious 3RR revert violation that actually involves the "Revert" tool that we don't need to split hairs over who added the word "not" or who changed the article to say "Only stupid people believe [old content]." Usually by that point, the edit war is probably well underway, and people could have been blocked under 3RR before any of that happened.
 * Correct procedure, as always, is BRD: once the first revert to your work is made, discuss it: once you've made the second revert, you're treading thin ice before you're fully edit warring. Jsharpminor (talk) 12:13, 12 April 2011 (UTC)

I am really frustrated at the 3RR rule because it has left stuff in that is out of the article scope
I have been told to stop reverting edits form a user who refuses to discuss his edits and keeps adding things outside Bradford to the Bradford article as reported here. Because he refuses to discuss or back down it means that if we obey the rules the article will continue to have the incorrect information. Surly it can't be the intention that wikipedia editors who break the rules always win edit wars? I have marked up warnings but this just seems to make the article worse - it is daft that we cannot correct it. -- Q Chris (talk) 10:55, 1 April 2011 (UTC)
 * If it clearly doesn't belong in the article, it shouldn't be that hard to get someone else to remove it, though. Have you tried our dispute resolution procedures? In your case, a third opinion might be useful. Heimstern Läufer (talk) 12:43, 1 April 2011 (UTC)

Talk page notices
From WP:USER: "Note that restoring talk page notices is not a listed exception to the three-revert rule."

While I am reluctant to see instruction creep ... this does seem like a valid exception, maybe it is subsumed under "vandalism" for those notices which WPUSER states should not be removed? Rich Farmbrough, 18:44, 29 April 2011 (UTC).