Wikipedia talk:Wheel war/Proposal 1

This page concerns a proposed wheel warring policy to enforce sanctions strictly against wheel-warring on Wikipedia.

Proposals
The following proposals are proposed additions to Wheel war, and would be considered policy. Discussion over the drafting of these proposals is available at Wikipedia talk:Proposed wheel warring policy/Archive.

Philwelch's proposal
In the spirit of the 1RR, no sysop action may be reverted more than once in any given dispute without sufficient discussion.


 * Suppose Admin A decides to protect a page. Admin B disagrees with this decision and chooses to unprotect. Admin A accepts Admin B's reasoning. In this case, no violation has taken place.
 * Suppose instead that when Admin B chooses to unprotect, Admin A reprotects (thus reverting Admin B's unprotection). Admins A and B discuss the situation and come to agreement. No violation has taken place.
 * Suppose instead than when Admin A chooses to reprotect, Admin B unprotects again, and Admin A reprotects again. At this the situation has degenerated into a wheel war and both Admin A and B would be in violation of this policy.
 * If, in the midst of this wheel war, Admin C comes along and unprotects after Admin A, Admin C is also in violation of this policy. Unlike the Three-revert rule, the wheel-warring policy is applied "per side" or "per issue", not per user. Similarly, the time limit is not 24 hours, but rather however long it takes to settle the dispute and achieve Consensus—or, failing that, decide the issue via standard Wikipedia processes.

Edit wars over protected pages, while arguably wheel warring, fall under the policies regarding edit warring instead of this policy. Edit warring by admins, however, is in particularly poor form—the position of trust that admins hold requires them to be held to a high standard of conduct.

Violations of this policy are to be posted to Administrators' noticeboard/Incidents. Any administrator who violates this policy will have their administrator privileges stripped by the stewards pending further review, with obvious exceptions in the case of simple vandalism by a "rogue admin".

Statement by Philwelch regarding proposal
Just as the Three revert rule puts an end to edit warring, not by a vague ban on "edit warring" but rather by an objective limit, my proposed wheel-warring policy places objective limits on reversions of admin actions. Vague, unclear policies only lead to rules lawyering. Wheel warring admins try to say "I wasn't wheel warring!". Maybe not, but if they broke the 1RR on admin action, that's another thing entirely. — Phil Welch 00:57, 18 February 2006 (UTC)

Carnildo's proposal

 * Wheel warring consists of two or more administrators repeatedly using their admin abilities to undo each other's actions. Any admin engaging in wheel warring may be de-sysopped for up to 24 hours, much the same way that any editor engaged in edit warring may be blocked.

Theresa's proposal
Admins should not engage in Wheel wars. Those that do risk having their admin powers removed permanently by the arbitration committee or Jimbo. The best way to avoid wheel wars is to talk the matter over with other admins before taking any action.

John Reid's proposal
No admin may do any admin thing twice in the face of admin opposition.

Locke Cole's proposal
An adminstrator may only undo another admins action one time (this includes the originating admin redoing their action after having it undone). If any adminstrator exceeds this limit, their sysop access may be removed by Jimbo, a Steward, or a Bureaucrat (when and if they are able to remove sysop access in the future), for a period not to exceed 168 hours. Nothing in this proposal precludes Jimbo or the Arbitration Committee from removing an adminstrators sysop access.

Examples:


 * Sysop A blocks a user (Sysop A is the originating admin of this sequence of actions).
 * Sysop B unblocks the user (Sysop B has used their 1WW).
 * Sysop A reblocks the user (Sysop A has used their 1WW).
 * Sysop C unblocks the user (Sysop C has used their 1WW).
 * Sysop A reblocks the user (Sysop A has violated this proposed policy and is now subject to having their sysop access removed for a period of at least 24 hours).
 * Sysop D unblocks the user (Sysop D has used their 1WW).

Please note that 1WW is not an entitlement. Administrators should, ideally, never undo eachothers actions. This proposal only serves to set a hard limit on those actions when they do, unfortunately, occur.

Kelly Martin's proposal
This was originally proposed at Admin zero-revert rule on 2006 March 8.

The Admin zero-revert rule (or 0RR) is a proposed policy which would apply to all Wikipedia administrators.

The policy states that, if an administrative action (deletion, undeletion, blocking, unblocking, protection, unprotection, rollback, etc.) is reversed by the action of another administrator, it may not be redone within 24 hours by the same administrator who previously performed it.

The only exceptions to this rule are situations where not immediately repeating an administrative action would present a clear and immediate danger to a person, to Wikipedia, or to the Wikimedia Foundation, or when the action to be repeated has been performed by Jimbo Wales or at his explicit direction as an "Office Action".

This policy is not intended to apply to cases where the situation has clearly and significantly changed between the undoing and the redoing of the admininstrative action in question, especially where the stated reason for undoing the action no longer applies. For example, if an article unprotected after a vandalism spree is over, and the vandalism later begins anew, the same administrator who originally protected the article may protect it again. It is also not intended to apply to cases where consensus to redo the action has been established after discussion in a relevant forum, such as the administrators' noticeboard.

This rule specifically does not apply to:
 * repeated cycles of creating and deleting content which may be speedily deleted; the content must be undeleted by another admin to create a 0RR situation;
 * repeatedly blocking the same user within 24 hours, where the original block expires in the interim; the user must be explicitly unblocked by another admin (or the block explicitly shortened) to create a 0RR situation;
 * any situation in which an admin reverses his or her own action ("self-reversing").

An admin who inadvertently violates this rule is expected to immediately reverse the offending action as soon as possible. Failure to do so may result in the temporary or permanent loss of administrative privileges.

The purpose of this rule is to curtail wheel wars. Admins who feel that they have been improperly reversed by another administrator are required by this policy to either convince the admin who reversed them to reverse themselves and reinstate the original action, or seek the assistance of other admins to reach a consensus decision.

Admins who feel the need to reverse another admin's actions should carefully consider whether the action truly needs to be immediately reversed. Any admin who believes that another admin has acted incorrectly is strongly encouraged to request the acting administrator to reverse the action himself, and to wait a reasonable time for a response (preferably at least a day) before reversing. If the original admin is unavailable or does not agree that the action should be reversed the issue can be brought up at the administrators' noticeboard or another appropriate forum.

Attempting to "game" this policy (by repeating without discussion just past 24 hours), or deliberately and repeatedly reversing another administrator without discussion as a form of harassment, whether or not in violation of this policy, will be viewed as incivil and can result in the loss of administrative privileges.

See also


 * Three revert rule
 * Harmonious editing club

Kelly Martin's second proposal
No admin shall revert the action of any other admin without first discussing it with that admin. Failure to discuss a revert will result in desysoping. The only exception to this rule is when the reversion is performed by or at the direction of an official of the Wikimedia Foundation, acting in that capacity.

Straw poll
This is a straw poll to see which proposals have high levels of support. The straw poll begins 00:57, 18 February 2006 (UTC) and ends seven days (168 hours) from that date and time. This is sort of like approval voting—you can support more than one proposal, and if more than one proposal has sufficiently high levels of support, we'll deal with that later.

Support (1)

 * 1) Support my own proposal, of course. — Phil Welch 00:59, 18 February 2006 (UTC)
 * 2) Support one the condition that if a third admin really did not check the log first (didn't know there was a wheel war), s/he should not be punished. Voice _of_  All T 01:06, 18 February 2006 (UTC)
 * This is the problem with this proposal. Admins may very well violate it accidentally because they didn't think to check the logs. Theresa Knott | Taste the Korn 01:13, 18 February 2006 (UTC)
 * Admins *should* check the logs before they block or protect, or failing that, immediately afterwards. Upon discovering they've wandered into a wheel war, they should be able to revert themselves and escape sanction. I think it's important to prevent tag-teaming, which is why this provision is in there. — Phil Welch 02:49, 18 February 2006 (UTC)
 * 1) Support is clear and seems reasonable. Punishment for violation is left unstated, which is less good. Stevage 14:45, 18 February 2006 (UTC)
 * 2) Support. Any of them are better than none. People are nitpicking, and we need something to stop the nonsense. - Taxman Talk 23:24, 27 February 2006 (UTC)

Oppose (1)

 * 1) Oppose I'm not a big fan of the bit about "Admin C". This should be per-user, otherwise the admin who takes action first is protected from having their action undone by anyone. —Locke Cole • t • c 01:19, 18 February 2006 (UTC)
 * 2) Oppose -- Sorry; too complex. John Reid 05:05, 18 February 2006 (UTC)
 * 3) Oppose. I don't think as a hard rule this works. --Improv 06:18, 18 February 2006 (UTC)
 * 4) Oppose We have to enforce our rules and establish accountability at an individual level.  This attempt to regulate according to blocs is not in keeping with our principles or goals.  It's also just too hard to enforce.  Rossami (talk) 07:54, 18 February 2006 (UTC)
 * 5) Oppose as it doesn't make clear when a "dispute" has ended. Physchim62 (talk) 21:28, 18 February 2006 (UTC)
 * 6) Oppose : It turns it into a tic-tac-toe game, that can "lock" an article into a protected status without the user's recourse, OR ban a user without oversight until the issue is resolved? No thanks.  &rArr;   SWAT  Jester   [[Image:Flag_of_Iceland.svg|18px|]]  Ready    Aim    Fire!  20:14, 19 February 2006 (UTC)
 * 7) Oppose according to my logic below. &mdash;Simetrical (talk • contribs) 02:35, 20 February 2006 (UTC)
 * 8) I oppose any wording with strict limits. I'm sure that not every single one of our 800+ administrators act in a reasonable and fair manner 100% of the time.  In some rare cases, undoing another administrator's actions repeatedly is the only way for me to protect innocent users.  JYolkowski // talk 01:32, 24 February 2006 (UTC)
 * 9) Oppose, tortuous, nothing like the 1RR in spirit, letter, or effect. Alai 02:29, 24 February 2006 (UTC)
 * 10) Oppose. Too complicated.  There are several hundred admins; to expect them all to learn these complex niceties and hold them in their heads at all time won't happen, no matter how good an idea it would be.  - ikkyu2  ( talk ) 01:46, 6 March 2006 (UTC)
 * 11) Oppose. Excessively complicated, instruction creep, which even specifically fails to include edit wars over protected pages as admin actions.. --Mysidia (talk) 06:46, 7 March 2006 (UTC)
 * 12) Too complex. - Mailer Diablo 14:23, 8 March 2006 (UTC)
 * 13) Oppose --Ter e nce Ong 15:03, 8 March 2006 (UTC)
 * 14) Oppose. I agree with Rossami and Locke Cole about the "per side issue". Sjakkalle (Check!)  15:14, 24 March 2006 (UTC)
 * 15) Oppose. SlimVirgin (talk) 20:37, 13 April 2006 (UTC)
 * 16) Oppose as per Locke Cole. This gives too much priority to the bold.  And I say this as one of Wikipedia's bolder admins. Kelly Martin (talk) 23:34, 13 April 2006 (UTC)

Support (2)

 * 1) Weak support, doesn't contain anything I object to. JYolkowski // talk 01:36, 24 February 2006 (UTC)
 * 2) Support. Any of them are better than none. People are nitpicking, and we need something to stop the nonsense. - Taxman Talk 23:24, 27 February 2006 (UTC)
 * 3) Support. Defines wheel warring broadly, suggests a remedy, intuitively understandable.  - ikkyu2  ( talk ) 01:48, 6 March 2006 (UTC)

Oppose (2)

 * 1) Oppose -- viz "repeatedly"; cf can i do it until i need glasses; sorry John Reid 05:05, 18 February 2006 (UTC)
 * 2) Oppose. Also fails as a hard rule. --Improv 06:18, 18 February 2006 (UTC)
 * 3) Oppose. We need an unambiguous rule. &mdash;Simetrical (talk • contribs) 02:35, 20 February 2006 (UTC)
 * 4) Oppose. SlimVirgin (talk) 20:38, 13 April 2006 (UTC)
 * 5) Oppose. Too vague, and the threatened penalty too weak.  A 24-hour desysoping is too weak a penalty for wheel warring, especially since it often takes more than 24 hours to get someone desysoped in the first place.  Keep in mind that desysoping requires steward intervention and most stewards will require evidence that "a consensus exists" on the project to revoke administrative privilege; they cannot be expected to review the underlying rules of a project to determine if a violation of local policy exists.  In practice, the only way to desysop someone without an ArbCom order or a lengthy discussion is for Jimbo to do it himself, and we can't rely on Jimbovention to resolve all conflicts. Kelly Martin (talk) 23:40, 13 April 2006 (UTC)

Support (3)

 * 1) I support my idea. It has the advantage of flexibilty. It states no predefined punishments but allows the AC to decide the best course of action. Theresa Knott | Taste the Korn 01:09, 18 February 2006 (UTC)
 * 2) Support Fred Bauder 02:04, 18 February 2006 (UTC)
 * 3) Support - Clarifies existing policy. I want to note that my support is contingent on the spelling error in Theresa's proposal be fixed. :) Note however, that as this is effectively current policy, I don't know if codifying it is necessary. --Improv 06:18, 18 February 2006 (UTC)
 * 4) I can live with this one. JYolkowski // talk 01:35, 24 February 2006 (UTC)
 * 5) Support. Any of them are better than none. People are nitpicking, and we need something to stop the nonsense. - Taxman Talk 23:24, 27 February 2006 (UTC)
 * 6) Support. Leaves room for broad discretion and casuism.  Also note that it's not mutually exclusive to any of the other policies; in fact, it's compatible with all of them.  - ikkyu2  ( talk ) 01:49, 6 March 2006 (UTC)

Oppose (3)

 * 1) Oppose -- Too vague, too harsh; sorry John Reid 05:05, 18 February 2006 (UTC)
 * 2) *I can understand that you may think it too vague. (I prefer flexible to vague personally) But too harsh? It doesn't have any set penalties. Theresa Knott | Taste the Korn 16:01, 18 February 2006 (UTC)
 * 3) ** Um, ...risk having their admin powers removed permenantly by the arbitration committee or Jimbo... -- that may be appropriate in some circumstances; in others it might be about right to make the offender stand in the corner for 15 min. It's just my bias -- I like to see standards of conduct sharp and explicit, penalties for violations adjustable ad hoc. John Reid 08:17, 19 February 2006 (UTC)
 * 4) Oppose. We need a hard rule, sorry. &mdash;Simetrical (talk • contribs) 02:35, 20 February 2006 (UTC)
 * 5) Oppose fails to clarify existing policy.  Arbcom's operational definition of a WW, from at least one recent case, seems to be "first strike is decisive".  (i.e. that the "wheel-warrer" is always the original undoer of an admin action.)  If that's the definition to be used here, even stronger oppose.  Alai 02:33, 24 February 2006 (UTC)
 * 6) Oppose per my comments below. —Locke Cole • t • c 22:50, 25 February 2006 (UTC)
 * 7) Oppose. A penalty as harsh as desysoping is too harsh when the term "wheel war" remains vague. Sjakkalle (Check!)  15:12, 24 March 2006 (UTC)
 * 8) Oppose. Wheel war isn't defined. SlimVirgin (talk) 20:39, 13 April 2006 (UTC)
 * 9) Weak Oppose.  Mere restatement of existing policy, which has been demonstrated to be insufficient given the number of wheel wars which have occurred of late.  However, a formal statement that the ArbCom can desysop for wheel warring is a good thing to have on the books, even if rarely used.  I may change my vote if I end up opposing all of the proposals. Kelly Martin (talk) 23:43, 13 April 2006 (UTC) Too weak.  A more stringent policy is necessary in light of recent events. Kelly Martin (talk) 21:16, 19 April 2006 (UTC)

Comment (3)

 * This pretty much restates how things are now. I do think we should bave a 1RR-style wheel war (let's call it 1WW) policy though, but not as Phil Welch has defined it. A 1WW policy would discourage admins from engaging in repeated wheel war actions, but not allow a single admin to effectively hold the action they performed hostage (as Phil Welch's proposal seems to allow). —Locke Cole • t • c 01:26, 18 February 2006 (UTC)
 * 1WW is a good name for whatever washes out of here. John Reid 05:05, 18 February 2006 (UTC)
 * Agreed. —Locke Cole • t • c 01:00, 19 February 2006 (UTC)
 * This seems like a recipie for ArbCom cases to me: I would support it if it were clearer on the enforcement mechanism. For whatever choice we come up with, I would suggest a page such as WP:AN/1WW (or WP:ARBCOM/1WW or WP:BN/1WW, whatever), along the lines of WP:AN/3RR. Physchim62 (talk) 21:35, 18 February 2006 (UTC)
 * Wheelwars are sufficiently rare that WP:ANI would be enough. — Phil Welch 00:16, 19 February 2006 (UTC)
 * WP:BN seems reasonable (it's very low traffic at the moment). As Phil says, they're rare enough that a subpage seems like overkill. —Locke Cole • t • c 00:58, 19 February 2006 (UTC)
 * I think requests for enforcement would get lost on WP:AN/I (remember that were asking for steward action here); and having a seperate page would effective "name and shame" warring admins, and act as a deterrant in itself. Physchim62 (talk) 18:57, 20 February 2006 (UTC)

Support (4)

 * 1) Support -- Clear, unambiguous, simple, short. Doesn't address so-called tag-team ww; but are these the most common problems? If (say) 4 big-endian admins protect Boiled eggs and 5 little-endian admins unprotect, then we've run out of endians and the war is over, so long as none of the 9 re-enter the process. Penalties are deliberately not specified; violations should never be common enough to demand rubber-stamp justice. Let the punishment fit the crime. John Reid 05:05, 18 February 2006 (UTC)
 * 2) Support 1RR for admin actions (with the obvious caveats about self-reverts and subsequent actions as a result of consensus).  Rossami (talk) 07:58, 18 February 2006 (UTC)
 * 3) Support Fred Bauder 13:25, 18 February 2006 (UTC)
 * 4) Support. Actaully I like it better than my own idea. Theresa Knott | Taste the Korn 16:03, 18 February 2006 (UTC)
 * 5) Weak Support this should lead to better discussion of issues, but I'd like to see a little morte clarification. Obviously it means block the same user twice for example, rather than simply blocking, and there also needs to be some position on consensus derived actions. Steve block  talk 20:48, 18 February 2006 (UTC)
 * 6) Support. &mdash; Matt Crypto  14:53, 19 February 2006 (UTC)
 * 7) Support, simply because it effectively means that given equal numbers of admins on each side, the "default" state is unblocked, unprotected, undeleted until consensus is reached. All the other proposals keep things blocked, protected, or deleted until things are worked out, as far as I can tell. &mdash;Simetrical (talk • contribs) 02:29, 20 February 2006 (UTC)
 * 8) Support as per Simetrical. It's the simplest and seemingly the most effective proposal. Chairman S.  |  Talk  20:34, 20 February 2006 (UTC)
 * 9) Support as reasonable compromise between my position and doing nothing at all, with the proviso that it be less ambiguously phrased to prevent wiki-lawyering. — Phil Welch 04:28, 21 February 2006 (UTC)
 * 10) Support -- it seems like the fairest proposal, it's simple without any pretentious Wiki-legalise, and unlike other proposals, avoids being instruction creep. --Mysidia (talk) 14:45, 21 February 2006 (UTC)
 * 11) Support Brevity is the soul of wisdom in this case.  I support any proposal in this area over nothing, but this is my favorite. Xoloz 03:46, 23 February 2006 (UTC)
 * 12) Support. This really is in the spirit of the 1RR.  Would support additional non-binding guidelines on the undesirability and courtesy implications of "non-0RR" actions, and especially tag-teaming actions, without at least attempts at prior discussion, but wouldn't want to try and make that iron-clad.  Alai 20:34, 23 February 2006 (UTC)
 * 13) Support – clear and simple, sums it up perfectly – Gurch 22:12, 25 February 2006 (UTC)
 * 14) Support, so long as it is understood self-reverts to correct one's own mistakes are allowed (as we do for 3RR). I like the simplicity of this. Jonathunder 22:22, 25 February 2006 (UTC)
 * 15) Support This holds admins to the high standard they ought to have. Ashibaka tock 20:18, 27 February 2006 (UTC)
 * 16) Support. Any of them are better than none. People are nitpicking, and we need something to stop the nonsense. How about change this to say admin 'action' so it doesn't sound so goofy though? - Taxman Talk 23:24, 27 February 2006 (UTC)
 * 17) Support, simple and a third admin is able to step in if one of the others is being unreasonable or has gone insane, which seems to happen from time to time. It won't stop all wheel warring, as an admin can get their friends to do their bidding, but it will help reduce it greatly. Also, if one side is obviously wrong, to an outsider, they will probably run out of admins to vote for them sooner than the other side. Perhaps an addition should be made so that the default condition is that a user is unblocked and a page is unprotected, if different admins keep coming and reverting. I would not make that addition now. We should see if something like that is necessary after it is put into practice. -- Kjkolb 11:44, 1 March 2006 (UTC)
 * 18) Support, perfect, although ientical to Looke Cole's proposal. Werdna648T/C\@ 11:21, 7 March 2006 (UTC)
 * 19) Support as best proposal on the topic of the bunch.  young american  (talk) 21:56, 14 March 2006 (UTC)
 * 20) Support. Straight, simple and effective. Sjakkalle (Check!)  15:18, 24 March 2006 (UTC)
 * 21) Support a one-admin-revert rule as laid out here, though I would prefer the language be changed to "No administrator may twice perform any admin action in the face of admin opposition." I also suggest using 1AR and not 1WW, as explained in the Comments section below. JDoorj a m    Talk 13:08, 12 April 2006 (UTC)
 * Weak support, basically the same as my original proposal, below. Kelly Martin (talk) 00:03, 14 April 2006 (UTC)
 * 1) Support. This is defacto how I already thought things were.  Except I would like to see it per side instead of per admin.  I liked Phil's proposal, but way too wordy.  Basically if you take an admin action and you get reverted, it should immediately be kicked up into some kind of community process.  The privilege of individual admins to impose their judgment ought to end the minute two admins are in conflict.  (Same goes for IAR.) Jdavidb (talk &bull; contribs) 20:22, 17 April 2006 (UTC)
 * 2) Support. WHS Rich Farmbrough 23:19 19  April 2006 (UTC).
 * 3) Support. Better than no policy and seems to be gathering sufficient backing.--Eloquence* 15:31, 20 April 2006 (UTC)

Oppose (4)

 * 1) Oppose. Bad idea. Too inflexible. --06:18, 18 February 2006 (UTC) — Preceding unsigned comment added by Improv (talk • contribs)
 * 2) 3RR is inflexible and it's one of our best rules. Ashibaka tock 20:19, 27 February 2006 (UTC)
 * 3) Oppose. I would support of the wording was more clear. Does the first action count as against opposition?. Voice _of_  All T 21:44, 18 February 2006 (UTC)
 * 4) No. If Admin A does something she knows will annoy Admin B, it may be rude but it's not a ww yet. If Admin B undoes that action without talking it over with Admin A, it may be careless and below the standard expected of admins, but it's still not a ww -- if it stops there.
 * 5) Oppose, isn't clear as per Voice of All, and doesn't specify any enforcement for violators. —Locke Cole • t • c 00:59, 19 February 2006 (UTC)
 * 6) Oppose as per locke cole, also leaves issue of admins getting friends to chip in un-addressed. Why not make it "do any admin thing once ..."? No admin should ever be acting against voiced admin opposition, no? Stevage 11:36, 19 February 2006 (UTC)
 * 7) In a better world I would agree with you, Stevage. Ideally, no admin should ever undo another admin's action for any reason. (If B disagrees with A, they work it out and perhaps A undoes his own action, perhaps not.) But this is an imperfect world and admins are imperfect people. I think it's fair and reasonable for every admin to take one shot -- good, bad, cheap, whatever. It's not the ideal climate in which to work things out -- and I certainly hope admins will discuss things in sidebar (at least sometimes) without slogging through a stone-throwing war, however measured. But the true current state of affairs is recurrent ww and there's nothing to keep the whole raft from going over the falls. I think we can tolerate all the divisive, biased admins on Topic X weighing in, each throwing his single stone, and stepping aside. Eventually they will all have exhausted their one shots and more neutral admins can come along and clean up. Remember, 1WW is a hard, bright line. Crossing it is impermissible; simply stepping on it reflects unfavorably upon the stepper. John Reid 23:55, 21 February 2006 (UTC)
 * 8) The wording is too vauge. Do any admin thing twice? I thought it was a joke at first. (The opposition part is nice and clear for me, it's the "thing" that is unclear because it doesn't say something like "do any admin thing on one user or article twice".) BrokenSegue 04:28, 23 February 2006 (UTC)
 * 9) * Everyone already knows pretty much how to interpret the three-revert-rule. Would it be sufficiently clear if we reworded it to something more like "Admin actions are subject to a one-revert-rule"?  Rossami (talk) 06:29, 23 February 2006 (UTC)
 * 10) **The xRR are applied per 24 hour period, the 1WW would be per-dispute. — Phil Welch 19:20, 23 February 2006 (UTC)
 * 11) I oppose any wording with strict limits. I'm sure that not every single one of our 800+ administrators act in a reasonable and fair manner 100% of the time.  In some rare cases, undoing another administrator's actions repeatedly is the only way for me to protect innocent users.  JYolkowski // talk 01:32, 24 February 2006 (UTC)
 * Can you clarify why you're then supporting a different measure which prescribes temp-desysadmining for doing so? And doesn't this envisage a situation where you're the only admin "protecting innocent users", in the face of unspecified numbers of others re-doing the action in question?  Cut and paste opposes seem to imply a certainly lack of differentiation between the practical consequences of several very different measures.  Alai 02:28, 24 February 2006 (UTC)
 * If an administrator is in the process of blatantly ignoring all other policies, I can't count on them following this one. Improper blocks have the potential to cause enormous amounts of harm and ill will, and rectifying those immediately is of the utmost importance.  Ideally other people would be unblocking, but one can't always count on others immediately noticing these things.  JYolkowski // talk 22:00, 24 February 2006 (UTC)
 * Unless, say, one informs them. You didn't comment on the apparent inconsistency of your support for Carnildo's proposal.  Are you distinguishing between "No admin may do" and "may be de-sysopped"; or between "more than once" and "repeatedly"?  (I assume not the latter, given your claimed need to do exactly that.)  Alai 22:42, 24 February 2006 (UTC)
 * 1) Oppose. Too inflexible.  Also, so badly worded that no person who loved language could be happy to be governed by it.  - ikkyu2  ( talk ) 01:52, 6 March 2006 (UTC)
 * 2) Oppose. This would mean if admin 1 blocks and admin 2 unblocks and no one else gets involved, the unblock stands, so the person who started the wheel war by being first to undo the admin action gets his or her way. SlimVirgin (talk) 20:41, 13 April 2006 (UTC)
 * 3) Oppose. Too weak. Kelly Martin (talk) 21:16, 19 April 2006 (UTC)

Comment (4)

 * 1)  JYolkowski writes: In some rare cases, undoing another administrator's actions repeatedly is the only way for me to protect innocent users. This may indicate a basic difference in our understanding of how collaborative editing works. My theory is that there is nothing that I must do twice, even as a plain editor. Indeed, there is nothing I must do once. I feel compelled to respond on talk to questions directed to me; I may undo things I've done wrong; and of course I try to shoulder my fair share of the burden as and when I see it. But I must believe that the wiki will roll on with or without me. Were I an admin I would hold to the same view regarding admin actions -- no matter how strongly I feel that something must be done, if nobody else feels the same way, then I'm probably wrong. So, it's enough that I do any given thing once -- and if it is undone by another, then perhaps a third party will redo it. Certainly some editors -- and some admins -- do destructive things and we all have a responsibility to oppose this. But I cannot take the entire burden of policing even one word on one page; and I think it would be a dangerous attitude for me to cast myself in the role of the Lone Ranger. John Reid
 * Ok, I may be over-rhetoricking this, but can we call it something besides 1WW? I feel as though this says "you may wheel-war, but only a little bit."  So as to be in line with 1RR and 3RR, can we call this 1 Admin Revert, or 1AR?  Picky, perhaps, but if our goal is to eliminate wheel-warring, we shouldn't enshrine its use in the language we choose to use. JDoorj a m     Talk 13:06, 12 April 2006 (UTC)
 * Actually I felt 1WW might remind the person abusing it that what they're doing is wrong, even if they're allowed to do it once. —Locke Cole • t • c 22:10, 12 April 2006 (UTC)
 * As Theresa pointed out on the talk page, not all administrative reversions are wrong. If John Q. Bannedforvandalism promises me he'll be good from now on and some other admin blocked him, I think it's perfectly legit for me to unblock him so long as I make sure he keeps his promise. — Phil Welch (t) (c) 00:07, 13 April 2006 (UTC)

Support (5)

 * 1) Support as author. —Locke Cole • t • c 09:12, 18 February 2006 (UTC)
 * 2) Weak support I can agree with the points that say that this enshrines a "right" to undo other admins actions, and I can also see ways in which this would not stop the worst of wheel wars, but this seems the only proposal which has teeth and which is based on an already functioning model. Physchim62 (talk) 21:38, 18 February 2006 (UTC)
 * 3) Support This allows other admins to express their views and indeed show a majority or even consensus (i.e. Admin A blocks, Admin B unblocks, Admin A reblocks, Admin B reunblocks, Admin C blocks, if admin B comes in and unblocks it shows he's at the very least going against a majority....especially if Admin D E F and G are waiting to reblock.)  &rArr;   SWAT  Jester   [[Image:Flag_of_Iceland.svg|18px|]]  Ready    Aim    Fire!  20:16, 19 February 2006 (UTC)
 * 4) Mgm|(talk) 22:02, 19 February 2006 (UTC) Things degenerate into a wheelwar the moment one admin does something twice. Sometimes it's hard to figure out what's happening and an admin can be asked to undo, for example a block by the affected user. "Admin C" shouldn't be punished for intervening unless it was clear they were aware of the disagreement between the first to editors. - Mgm|(talk) 22:02, 19 February 2006 (UTC)
 * 5) Support for future development as well. — Phil Welch 04:30, 21 February 2006 (UTC)
 * 6) Fine by me. BrokenSegue 04:29, 23 February 2006 (UTC)
 * 7) Weak support as second choice. In effect much like option 4, but with "lengthened and worsened" wording and provisions.  Alai 02:37, 24 February 2006 (UTC)
 * 8) Support. Any of them are better than none. People are nitpicking, and we need something to stop the nonsense. - Taxman Talk 23:24, 27 February 2006 (UTC)
 * 9) Support. I like this better than John Reid's proposal. A bit less unambiguous, although the wording could be worked on.
 * 10) Support. This is a proposal in the same spirit of the 3RR for edit wars. That rule works pretty well, and I think this one will too. Sjakkalle (Check!)  15:16, 24 March 2006 (UTC)
 * 11) Weak support, I suppose, because at least with this one, admin 1 blocks, admin 2 unblocks, admin 1 gets to reblock, so that's fairer. But I would prefer none of these; just an agreement that we won't undo each other's admin actions, at least as a rule of thumb. SlimVirgin (talk) 20:43, 13 April 2006 (UTC)
 * 12) Weak support with caveats: one, I think there should be a time limit after which one can revert anyway (to prevent stonewalling or obstruction caused by nonresponsive admins), and two, I disagree with the 7 day maximum desysopping.  Indefinite and/or permanent desysoping must be possible penalties for wheel warring.  Kelly Martin (talk) 23:50, 13 April 2006 (UTC)
 * Regarding the 7 day max, this is why I included language allowing for the ArbCom (or Jimbo) to remove access as they see fit. Presumably 7 days would be enough to open an ArbCom case and impose a temporary injunction removing the sysops access until the conclusion of the case. And I may not have mentioned this explicitly in the proposal, but I read it as "7 days per offense"; so if someone wheel wars five times, that's 7 &times; 5 (35 days max). I'll have to think a little more about the stonewalling admin issue (but my gut says either it'll be worked out on AN/I, or it'll be taken to the ArbCom). —Locke Cole • t • c 00:07, 14 April 2006 (UTC)
 * Speaking as a former Arbitrator, I tell you that if you say the penalty is seven days for wheel warring, some idjit will argue before the ArbCom that he can't be permanently desysoped for wheel warring because the policy says the maximum penalty is seven days. Yes, I know, the proposal has the "no limitation" clause, but that won't prevent the argument, and it won't stop people from getting behind the arguer in support, too.  No, I think it's best not to set a maximum penalty.  Let's not make ArbCom's job any harder than it already is. Kelly Martin (talk) 00:12, 14 April 2006 (UTC)

Oppose (5)

 * 1) Oppose, no right to undo another administrator's actions should exist. It might be appropriate or necessary to do so, but that should be a rare action that has to be justified. Fred Bauder 13:28, 18 February 2006 (UTC)
 * 2) *It's not a right, it's a limit. Preferably no admin would undo anothers admin actions. But in the event this occurs, this sets a stringent limit of one time only. —Locke Cole • t • c 00:02, 19 February 2006 (UTC)
 * 3) Yes the problem is with the An adminstrator may only undo another admins action one time. This reads as if he allowed to do it once. I don't want a situation to develope where low level wheel warring is the norm. Theresa Knott | Taste the Korn 16:12, 18 February 2006 (UTC)
 * 4) *Note the words "may only", it's not "shall only" or "will only". If there's a way to word this such that it doesn't sound like an entitlement I'm all ears, but I did my best to not word it so it would be read as you seem be reading it. —Locke Cole • t • c 00:02, 19 February 2006 (UTC)
 * 5) I don't like that if there are only two admins in the dispute, the originating admin can redo his original action. I prefer John Reid's version (which is otherwise identical to this, wording aside). &mdash;Simetrical (talk • contribs) 02:33, 20 February 2006 (UTC)
 * 6) *The idea is to give preference to the admin who did the action being undone. If two admins disagree (that is, the originating admin and another admin) then they each "wheel war" once. If a third admin disagrees with the originating admin and undoes it as well, then there's basically a small/minor consensus against the originating admins actions. At this point the originating admin can't redo their action again without violating this proposed rule/policy. —Locke Cole • t • c 02:38, 20 February 2006 (UTC)
 * 7) **I realize that's the idea, I just disagree with it. What this policy would mean is that if Bob protects a page and then Jim unprotects it, Bob can immediately reprotect it, and it will stay that way until further admins are called in.  Granted, the latter might not take long, but it may, particularly if the reversing admin isn't paying attention/goes to sleep/whatever.  I could foresee disputed blocks or protections going on easily for 24 hours longer than necessary if only one objecting admin spots them quickly.  Protection is harmful, and blocking all the more so. &mdash;Simetrical (talk • contribs) 03:46, 22 February 2006 (UTC)
 * 8) ***This isn't meant to change your mind, but the reason I built that in was because I'd like to assume good faith on the part of the originating admin. Obviously if it degenerates into a back and forth wheel war, it'll take neutral non-involved admins to decide if the block/protect/etc was really necessary, but when it comes to just two admins, I think we should err on the side of the person who made the original call. If another admin objects (and it only takes one), then the original act can be overturned. (I should have explained this in my original reply, so sorry for being verbose again). And don't forget, this can also be used the other way around; an admin unblocking a long blocked user (or unprotecting a long protected page). —Locke Cole • t • c 05:54, 22 February 2006 (UTC)
 * 9) I oppose any wording with strict limits. I'm sure that not every single one of our 800+ administrators act in a reasonable and fair manner 100% of the time.  In some rare cases, undoing another administrator's actions repeatedly is the only way for me to protect innocent users.  JYolkowski // talk 01:32, 24 February 2006 (UTC)
 * 10) *I think you just described precisely why we need a 1WW rule with strict limits. Otherwise we'll have admins claiming they were doing what was right (and wheel warring in the process). If you're really correct, make your case on WP:AN or WP:AN/I and let a neutral non-involved admin undo the action if they agree. —Locke Cole • t • c 03:09, 24 February 2006 (UTC)
 * 11) Oppose. Too complicated.  There are several hundred admins; to expect them all to learn these complex niceties and hold them in their heads at all times won't happen, no matter how good an idea it would be.  - ikkyu2  ( talk ) 01:54, 6 March 2006 (UTC)
 * 12) *Actually it's not really complicated unless you're interested in enforcing it. To avoid violating my proposal you just need to remember that you shouldn't undo another admins action more than one time. Unlike some of the other proposals though, mine provides a remedy and assumes good faith on the initiator. Spelling these details out requires some level of complexity I'm afraid, but I think the tradeoff in good faith and providing for a remedy for violations is a good thing. —Locke Cole • t • c 05:23, 7 March 2006 (UTC)
 * 13) Oppose. For complexity, partly over this whole 1WWW thing and specified suspensions. An admin reversing another's actions is not the end of the world; more than one reversion very likely indicates a specifically abusive administrative action by either the reverter, or the admin who was reverted a second time.  The policy document can't know who, therefore the nature of the penalty and who's penalized should be left as a matter of discretion of the Arbitration Committee and Jimbo: In other words, we don't need a provision to specify 1WWW is automatic 168 hours desysop, that just takes proper discretion out of the picture. --Mysidia (talk) 07:02, 7 March 2006 (UTC)
 * 14) *... If any adminstrator exceeds this limit, their sysop access may be removed [...] for a period not to exceed 168 hours.. May. Not to exceed. Not enough discretion? I intentionally left it open to judgement by the b-crat/steward handling the complaint. —Locke Cole • t • c 07:17, 7 March 2006 (UTC)
 * 15) Oppose, too weak. Kelly Martin (talk) 21:18, 19 April 2006 (UTC)

Comment (5)

 * 1) This one reads to my eye like a more verbose version of my proposal, so far as the standard of conduct is concerned; and resembles TK's in the penalty section. I think we're gravitating toward some sort of consensus: 1WW. John Reid 08:30, 19 February 2006 (UTC)
 * 2) I propose an extension to this rule. After first violation is reverted, no other actions are allowed untill a consensus is reached. That is, in the situation given in the proposal, NO E Admin could come in and use his/her 1WW. Anybody doing anything after Admin D is subject to block and/or desysop. Now the question is what to do if B reverts A's actions? (i.e. B also violates 1WW). Renata 14:26, 20 February 2006 (UTC)

Support (6)

 * 1) Of course.  See talk page.  —Ilmari Karonen (talk) 19:37, 13 April 2006 (UTC)
 * 2) Yes, something is better than nothing. There's no urgent need for many admin actions, and we can wait to discuss re-implementing them if need be. - Taxman Talk 19:51, 13 April 2006 (UTC)
 * 3) Wordier than John Reid's proposal but I think it says basically the same thing.  I can live with it.  Rossami (talk) 21:18, 13 April 2006 (UTC)
 * 4) Support .  SlimVirgin's objection is based on the "sole defender of the wiki" theory of adminship.  We're a community, and I should think that most admins have made contacts within the administrative community who can be asked to review a situation and reimpose a reversed action.  The reason why the reversing admin is favored over the nonreversing admin is that the status quo ante ought to be the default state of affairs; by allowing the initial administrative act to be reversed we favor the state that existed before the first administrative act.  A "true 0RR" favors the state that exists after the first administrative act, and that rule gives insufficient protection to nonadmins who might otherwise fall prey to an admin who abuses administrative authority to, e.g. win an edit war, or to harass.  (Locke Cole's proposal, which is weaker than mine and much weaker than the full 0RR, also suffers from this defect, but not as much because the original admin has to be persistent in maintaining the rightness of his/her action to reinstate the status posteriori.)  The full 0RR basically prevents overly bold acts, and even irrational acts, from being reversed.  Note, for example, under a full 0RR as proposed by SlimVirgin, I could block any admin I wanted, and that admin could not be unblocked without "consensus", which as we know can be very hard thing to find on Wikipedia.  A true 0RR gives administrators a great deal more power to harass, and I think that would be a bad thing indeed.  My proposal was intended to protect nonadmins from abusive admins, more so than to protect admins from one another. Kelly Martin (talk) 23:59, 13 April 2006 (UTC)

Oppose (6)

 * 1) Oppose. Again, this has the problem of allowing the admin who starts the wheel war to be decisive, if only two admins are involved. Admin 1 blocks, admin 2 unblocks. Admin 1 can't reblock and if no one else gets involved, admin 2 got his or her own way by being the aggressor. 0RR should mean precisely what it says. Do not undo another admin's actions, unless the situation is urgent or the circumstances exceptional. —The preceding unsigned comment was added by SlimVirgin (talk • contribs) 20:47, 13 April 2006  (UTC)
 * 2) *The "only two admins" scenario seems like a red herring to me. It's not as if we didn't have WP:ANI, and user talk pages, and e-mail, and even (shudder) IRC.  You can always find another admin.  —Ilmari Karonen (talk) 21:41, 13 April 2006 (UTC)
 * 3) Oppose. Too weak. Kelly Martin (talk) 10:42, 20 April 2006 (UTC)
 * 4) Oppose, deopping should never be the response to a first offense of this scale. --Golbez 23:21, 20 April 2006 (UTC)

Comment (6)

 * I don't see how this would prevent the slow-simmering wheel wars - i.e. usual wheel wars are admins reversing each other every half hour or so - this seems to merely make it every 24 hours. Isn't the purpose of this to prevent wheel wars altogether? Just another star in the night T 20:30, 14 April 2006 (UTC)

Support (7)

 * 1) In light of recent events, it's clear to me that a MUCH FIRMER policy against reverting is necessary. Kelly Martin (talk) 21:20, 19 April 2006 (UTC)
 * Would you please enlighten the rest of us? What recent events are you referring to?  Rossami (talk) 22:19, 19 April 2006 (UTC)
 * The recent event of Eloquence deciding that he's entitled to unilaterally reverse without discussion a protection imposed by Danny, one of most trusted administrators. That was wheel warring.  He deserves to be desysoped (and was desysoped, although that punishment, while originally imposed by Jimbo, was subsequently reversed by Jimbo for reasons that frankly make no sense to me) for that.  Not every case is as severe as this one, but we need to have the policy so that it can be enforced when needed.  Kelly Martin (talk) 13:15, 20 April 2006 (UTC)
 * Those interested in the case can see what happened on Talk:NewsMax.com and in this wikien-l thread. Suffice it to say that I strongly disagree with Kelly's characterization of what happened, and am deeply disturbed by the underlying authoritarianism.--Eloquence* 14:57, 20 April 2006 (UTC)
 * Apparently, Eloquence is no longer one of our most trusted administrators? --Golbez 23:22, 20 April 2006 (UTC)

Oppose (7)

 * 1) Pending an explanation of the need for such a draconian policy, my initial reaction is to oppose this as overly strict.  While discussing a contested matter with me would be a courtesy, if I do make a mistake, I would not want it to go uncorrected just because I'm not on Wikipedia for a few days.  If strictly interpreted (which the wording certainly seems to imply that we should do), this would also make some of our established practices illegal.  For example, it would no longer be allowable to simply undelete and list a contested speedy-deletion or to unblock an anon IP when the block is having unintended consequences on signed-in users.  Rossami (talk) 22:28, 19 April 2006 (UTC)
 * 2) Abso!@#$inglutely not. I'm very disappointed with this proposal, and really opens the door for cabals who do controversial things, get reverted by an admin, then get a crat to dispose of them. Among many other things that scream that this proposal is bad. --Golbez 23:20, 20 April 2006 (UTC)
 * 3) Oppose; this is more severe than the emerging general consensus. John Reid 23:37, 21 April 2006 (UTC)

Comment (7)

 * Define "discuss with the other sysop"; do you mean literally leave a message on the talk page and wait for a response, etc.? Or do you mean simply leaving a message indicating why you're reversing their action (and then being able to reverse the action; but still discussing if the other sysop responds)? In principle, I agree with it; many sysops have ignored WP:BLOCKs clause regarding undoing other sysop' blocks; specifically:


 * And really, it's only become more common place to undo another sysops action without discussing it with them directly (which needs to stop, IMHO). But I agree with that it seems rather draconian. A temporary desysopping until the Arbitration Committee has had a chance to consider/review the situation would be more appropriate. If you make this change, I'll support. —Locke Cole • t • c 11:04, 20 April 2006 (UTC)
 * The Arbitration Committee is the wrong entity to be enforcing this policy anyway. Clearly a new entity is needed.  Kelly Martin (talk) 13:23, 20 April 2006 (UTC)
 * In any event, if you would change the language to defer to an entity (ArbCom, or something new) on the matter of permanent desysoppings I'd support this. —Locke Cole • t • c 17:38, 20 April 2006 (UTC)

Refactor
Please see talk. John Reid 04:42, 6 April 2006 (UTC)