The previous poll ("Revert wars considered harmful") and discussion below indicate that the community is broadly in favor of the guideline "do not revert the same page more than three times in the same day".
For the purposes of this proposed policy, since it is more rigid than the current guideline, the rule would be no user may revert 6 or more times in one 24 hour period - these 6 reverts may be to one page, or spread across several pages. This rule is not intended to grant an allowance of 5 reverts per user per day, reverts should still be avoided.
The only exception to this rule will be in cases of clear vandalism — not content disputes, but cleaning up after a user who goes on a spree on vandalism, like the current case of User:Bird and his never-ending list of proxy IPs with which he is circumventing a hard ban.
Currently, when users engage in a revert war, sysops may decide to protect that page. A temporary page protection is supposed to allow the users to calm down and discuss the problem on the article's talk page. However, this prevents all users from editing the page in question, and does nothing to stop the warring users moving on to another article and engaging in exactly the same destructive behaviour.
This proposal is to turn the guideline into a policy, and to back this up with a 24 hour ban for any user who breaks it. This ban could be enforced by any sysop.
Proposed amendment by Eloquence, added on March 14[edit]
Any 24 hour ban decision needs to be confirmed by a quickpoll before it can be implemented. The policy will be refined on the basis of quickpoll results to better deal with exceptions and "gaming".
If you support this amendment, please add "with quickpolls" to your vote in favor. If you object to it, please add "against quickpolls" to your vote. Please also add your opinion on Wikipedia talk:Quickpolls.
Sysops already have the power to ban any anonymous user for 24 hours if they engage in destructive behaviour. This extension of their powers to ban logged in users would be clearly defined and their actions would be open to scrutiny.
Sysops would not be given the power to extend the 24 hour ban. Once a user returns after a 24 hour ban they have a clean slate and are given a second chance.
This ban could be put in place as swiftly as pages are currently protected by sysops; destructive users would no longer be able to disrupt the community for weeks before any action is taken against them.
If a sysop were to make a mistake, or abuse this power, the worst that any user would face is not being able to edit this website for 24 hours.
Any sysop abusing this new power would be an obvious candidate for desysopping.
Banning a user for 24 hours will give them time to calm down and to step back and examine their actions. The present system does little to slow a user down if they lose their sense of proportion and move from page to page as they are protected due to their warring.
When one or two users misbehave, the rest of the community is no longer punished by being prevented from editing certain pages and having to follow warring users around, cleaning up after them.
Sysops and users widely differ in what they see as "clear vandalism", so making an exception for clear vandals increases controversy. For example, note that Wik considers Lir a clear vandal.
Not making an exception for "clear vandalism" encourages individual vandals to get reverted (three times each!) by a wide variety of good Wikipedians. This promotes cohesiveness, shares the load, guards against burnouts, and displays a united community front.
When is a revert not a revert? When it is:
A partial revert - I revert part of an edit, but not the other part
A logical revert - I made a change, it got reverted, so I made the same changed but phrased it differently (hoping it would be more acceptable the second time).
A revert+edit - I reverted, and then I added some extra text myself (hoping it will address the relevant point).
Protecting the version disliked by the reversionista is a better alternative (see proposal and support below)
Banning prevents discussion, and discussion is very often precisely the force required to fix the revert war at source. While protection does not encourage discussion in all cases, often it does.
Extremely active editors may well revert six times in 24 hours for legitimate reasons without realizing it, particularly when the reverts are spread out over a large number of unrelated pages.
As phrased, there are too many ways to "game" the system
Setting per-user limits of any kind will set us up for endless, pointless games of "hunt the sock puppet."
Wikipedians in favor of this proposed policy[edit]
mav (revert wars solve nothing and only lead to ruined edit histories, high tempers, and increased attrition of valuable contributors. A 24 hour time-out should lesson the harm from edit disputes and encourage people to engage in consensus editing)
-Robert Merkel 11:32, 10 Mar 2004 (UTC) (I hate to have to make rules, but this one has become necessary).
4. Pfortuny 11:38, 10 Mar 2004 (UTC). See comment below.
—Eloquence 12:14, Mar 10, 2004 (UTC) (Tannin makes good points, but we can address these on a case-by-case basis)
Stewart Adcock 23:45, 10 Mar 2004 (UTC) (We should be able to trust any sysop to implement such bans on a case-by-case basis, when demanded by events.)
Good idea, Martin. Anything to reduce the anti-vandalism workload, is worth a try. --Uncle Ed 13:57, 11 Mar 2004 (UTC)
(Martin has voted against this, he added the arguments in opposition list. I wrote the proposal but have yet to vote. fabiform | talk 14:38, 11 Mar 2004 (UTC))
Hephaestos|§ 17:26, 12 Mar 2004 (UTC) I'll come right out and say one of the reasons I'm voting for this proposal is because if Wik, Lir and Anthony are all voting against it then it must be at least getting close to what we need here. Martin et al. have some very good objections, and these almost swayed my vote, but I think the overriding concern is that we need something in place now (yesterday, last month, last year). We cannot afford to allow this disruptive behavior to continue any longer. My feeling is that if this policy is put in place as-is now, any problems, glitches or nuances can be worked out on the fly.
J-V Heiskanen 15:11, 14 Mar 2004 (UTC) Hephaestos talks with my voice.
Bryan 22:38, 14 Mar 2004 (UTC) If this turns out to lead to worse heartache it can always be rescinded later, but edit wars are causing problems right now. We should be bold in experimenting with ways to fix them.
with quickpolls fabiform | talk 23:23, 14 Mar 2004 (UTC)
Angela - support with or without quickpolls. (note - I changed my vote from mildly oppose to support). I can't see any reason for one person to revert four times in the same day.
Tannin 11:48, 10 Mar 2004 (UTC) Much too inflexible to cope with reality. For example, two days ago someone moved a page ill-advisedly and I moved it back, then did about 30 reverts to fix all the links up again. That means I'd now be a banned user! A nice idea, but without more thought, this policy is a very bad one.
anthony 14:04, 10 Mar 2004 (UTC) The line between "clear vandalism" and "unclear vandalism" is far too fuzzy. This would allow subtle trolls to destroy wikipedia. Instead, I support some form of quickpolls, but the current implementation is not adequate.
Wik 18:15, Mar 10, 2004 (UTC) (This accomplishes nothing, edit wars will just go on with each side making 5 reverts a day. Furthermore people will mask reverts by making trivial changes to other parts of the article and by not indicating the revert in the edit summary. When two or more users have a content dispute that they can't resolve between themselves, someone else will just have to make a binding decision on the matter. That really should be obvious. But some people prefer to treat symptoms rather than causes.)
That's what this accomplishes: it removes those who are edit warring from the process and leaves others free to work on the article, instead of having those who edit war to get their version and only their version force the article to be protected indefinitely. The refinement tackles the dodging issue, by applying human judgement to the question. Jamesday 06:17, 12 Mar 2004 (UTC)
How does it remove them from the process? It just throttles them. There is no resolution of the dispute. --Wik 22:42, Mar 14, 2004 (UTC)
In its current form only, because I do not believe that an individual person should make a block, ban, unblock or unban decision. Jamesday 07:44, 11 Mar 2004 (UTC)
Strongly oppose, for the primary reason that "clear vandalism" is often unclear! Bird, at the start of his attacks, looked almost reasonable. A smarter vandal will be able to obscure himself in so much confusion that this rule will render us in effective to stop him. When we face a massive coordinated attack of smart vandals (and you know we will in the not-to-distant future) we don't want to be hindered by a six-revert-per-day rule and force the vandals to make us become hypocrites... -- Seth Ilys 20:13, 12 Mar 2004 (UTC)
I think we might be able to use the "Quickpoll" proposed above to deal with these situations instead of setting hard and fast rules. -- Seth Ilys 22:22, 14 Mar 2004 (UTC)
11. Ruhrjung 11:20, 10 Mar 2004 (UTC) Unless combined with a protection of the affected page(s) lasting much longer than these 24 hours, I fear a 24-hour ban is more of a teaser than teacher for the villians. — revised vote: something done is better than nothing! I've cast a vote in support of the wikipedia:quickpoll scheme.
Avoiding protection, so that others can continue to improve the article, is one of the objectives. Jamesday 07:44, 11 Mar 2004 (UTC)
We clearly have quite different frames of references with respect to time. I consider 24 hours a very short time. Far too short for other users to produce something the warrior won't feel tempted to disrupt. Had the ban been for a week or a month, and by default included neighbouring IP-adresses, it would have been quite another matter. --Ruhrjung 17:49, 12 Mar 2004 (UTC)
It can lead to abuse very easily.Muriel 21:40, 15 Mar 2004 (UTC)
I say it will you mite not be able to uther things then? Maybe just tel them to stop it -- zzo38 17:42, 2004 Mar 26 (UTC)
No. That means that each user can revert only 5 vandals every 24 hour period, and that's not counting any non-patent nonsense reverts. Anyway, this would probably cause "across the board reverts" - people could revert Wik without fear, because there's many (the general population) vs. a few (Wik, etc.). "Gang-up reverting" isn't a very good idea. ugen64 20:40, Apr 6, 2004 (UTC)
Oppose: If a revert war is under way then the page can be protected. Why need this?. Also what if someone is persistantly adding factually incorect statements etc to articles, and needs to be reverted G-Man 20:47, 6 Apr 2004 (UTC)
Oppose for all the reasons listed; the cure is worse than the disease. UninvitedCompany 21:07, 6 Apr 2004 (UTC)
Oppose, with quickpolls (that is to say, if this policy is forced through, the quickpoll ammendment would be a slight improvement). However, generally speaking, there are far too many different and legitimate reasons for reverting to justify an arbitrary limit of 6 (or any other number). Reverts and page protection routines should be dealt with individually and on their own merits. I would support simply a descision for page protection via quickpoll, but banning is never the answer to content disputes. 80.255 16:47, 16 Apr 2004 (UTC)
Oppose. Favors those with POV to come through, slant article, start revert war and leave the serious contributor on the down side of the rules. The rules should always favor wikicontent over nonwikicontent. Stirling Newberry 15:46, 27 Jun 2004 (UTC)
-- orthogonal 03:33, 14 Sep 2004 (UTC) All animals are equal, but some animals are more equal than others.
Oppose, I do not enjoy revert wars. I get involved in them when a user isists on POV. I prefer the topic be locked temporarily. Catchi? 23:19, 11 Apr 2005 (UTC)
Oppose. Jtkiefer 21:46, Jun 8, 2005 (UTC) The way I see it this would only place another layer of beaurocracy in the way of giving proper bans, I think this is an uneccessary extra block in the way of letting admins do their job.
Oppose. Not relatred to what it intends to accomplish, too time-delaying when many sysops won't even block and will remove someone elses block if the event happened even a few hours ago. Gene Nygaard19:13, 8 November 2005 (UTC)[reply]
Oppose. That would effect any user with a large amount of pages on watchlist too easily. On a real war where 2 parties revert each other accross a set of articles something similar might be benefitial but not in this form. Agathoclea11:32, 3 April 2006 (UTC)[reply]
Oppose: Martin: See below: I prefer allowing sysops to protect the version used by the set of contributors who have abided by the three revert guideline. I also question whether a limit of five reverts is too low. I also don't see the benefit in making exceptions for vandalism - pure vandalism will always be reverted by someone else. Also, I note that certain Wikipedians have felt free to unilaterally label other people as vandals, even where this is disputed, which would make this rule unnecessarilly controversial.
How does that address the desire to avoid protecting the article, so others can improve it? Jamesday 07:44, 11 Mar 2004 (UTC)
Some articles (e.g. DNA) are frozen for a long time, and editors might read talk pages, but people using wikipedia as a resource aren't going to. fabiform | talk 15:39, 11 Mar 2004 (UTC)
That's already failed. Those who are most prone to edit wars don't discuss and seek consensus, they edit war instead. That and the uneditability consequences of it is why we're having this discussion. Jamesday 06:11, 12 Mar 2004 (UTC)
I have changed my vote to abstain - having made the above protection policy (un)"official", I'm interested to see whether it's enough, or whether we need to do more. Martin 17:09, 14 Mar 2004 (UTC)
If you object, would the following change cause you to change your vote?
Whether the threshold has been passed is to be decided by a straw poll which requires at least five sysops participating and at least an 80% agreement that the threshold has been passed. If that 80% is subsequently not met, the block should be removed until it is. Sysops only because one concern raised elsewhere was that the process would be subverted through the use of multiple accounts created for the purpose. Human judgement should be used to decide what is a revert and what is legitimate dealing with trolls or masked reverts through trivial changes elsewhere in the article.
The purpose is to remove the scope for unilateral action and allow for human interpretation of whether a series of edits is legitimate, such as dealing with vandalism, or an attempt to work around the limit, as postulated by Wik. Jamesday 07:44, 11 Mar 2004 (UTC)
Pete/Pcb21(talk) 09:06, 11 Mar 2004 (UTC) (seems like a reasonable well-defined formulation of what I was grappling for in the chat below).
This might address the issue of the differentiation between good and bad reverts being so subjective if the decision were not left to one sysop. Even if it's a "good revert", it doesn't need to be made 4 times, so there is no need for a modification. The rule should apply in all cases. Angela
Tannin. At least I'd probably change my vote. I think there needs to be a bit more thought and refinement yet, but this suggestion is clearly heading in the right direction.
Stirling Newberry 15:55, 27 Jun 2004 (UTC) After having thought about it, there are just too many individual variables. Revert wars, the real kind, seem to come at the end of deteroriating relations between two users. That should be the deciding factor - how people have behaved over the course of the dispute. Otherwise, this is an incentive to goad a reasonable poster into anger and then ban them. That's a school yard game of long standing and institutionalizing it is a bad idea. There is also the very difference between pages with many contributors and pages with a few. In the many contributors section, "someone else will revert" is a reasonable attitude, however, in many cases where a page, particular in a topic that has a zealous net following, one could easily have several dittoheads against one reasonable poster.
Support the proposed policy but wouldn't support it with this change[edit]
Darkelf 01:41, 13 Mar 2004 (UTC) — The refined policy as I understand it would make the 24 hour bans useless again. A committee will take too long to decide and by that time the revert war will be in full force again.
I oppose this refinement SamSpade 20:36, 6 Apr 2004 (UTC)
Comments, suggestions for refinements, arguments against[edit]
Refactored conversation
We still need to warn people better that (multiple) reverting is bad.
There should be a 'semi-automatic' way of dealing with this problem. Ban lengths could be proportional to number of reverts. If a user reverts 60 times in the last two days, punished, say, (60/6/2)*24 = 120 hours (he was 5 times "a single bad behaviour")
This suggestion is inflexible and unrelated to quality of reversions. Why not just use a string of 24-hour bans?
This rule is trivial for a troll to "game" by making trivial (bad) edits, which a good user wouldn't be allowed to revert.
This rule isn't aimed at all trollish behaviour. We need seperate rules for that (and we need them badly).
You shouldn't feed the troll by following him and reverting every troll. Rather, call for community support.
Proposal: A user can be banned for 24 hours if X sysops post agreement, with reasons, on some notice page, within some short timespace. X would be at least 3 but small enough to enable a rapid response.
Proposal seconded, thirded, and fourthed; given glory laud and honor.
No good. This would delay implementing a ban and add an extra layer of bureaucracy. Still, it's not that bad: small values of X, and shared responsibility.
Unless a single admin can ban a user, there is scope for this bureaucracy. Yes, it should be easy to find 4-5 sysops. Not advocating a software change- there can just be a page for it.
Sysops currently have the technical ability to ban a logged in users, but absolutely no right to do so, except in the case of re-encarnations of hard banned users.
Too open-ended, and dangerously broad. Might allow a small number of sysops to effectively create policy as well as enforce it. Open to abuse. Voting for a fast ban could split the community. Sysops shouldn't have such broad powers.
Proposal was badly worded. Try instead: "X admins can ban a user if, in their collective opinion, the user has breached established policy." In effect it would be a rapid-fire arbitration committee, with more limited powers (temporary bans only). Binding decision would only be made by the real committee.
Wikipedia:Administrators says "Sysops are not imbued with any special authority, and are equal to everybody else in terms of editorial responsibility." This goes against that.
Well, currently any one sysop can ban a user for violating policy established by the whole community. This would make it require X sysops to agree, and is more democratic. We have already tried having whole community input (the old /ban pages). They were a disaster: lots of talk but decisions still had to be taken by Jimbo.
Yes, even one sysop has the power, but that's supposed to be used only for clear vandalism and reincarnations of hard-banned users. Extending this power to breaches of established policy is a significant change.
An X sysops model may be too simplistic. Suppose that X agree, but several disagree. Rather, we need ways to develop a consensus faster in an emergency.
This is reasonable and urgently needed until some of the ideas on the mailing list (in the Rethinking Vandalism and Transition to WikiDemocracy threads) come into being.
What we need is to delay all edits so they won't take place for several days, during which you'd only see the update while editing. Edit wars wouldn't be exciting.
This is a horrible idea. It's a WikiWiki. Fast fast. What you propose is a fundemental change which would drain the life and vitality of everything good about our rapid-fire editing processes. - Fennec 14:35, 12 Mar 2004 (UTC)
This wiki should be an encyclopedia, right? Encyclopedic articles don't need to be hastily changeable. Quite the contrary, indeed.--Ruhrjung 15:27, 12 Mar 2004 (UTC)
The instant gratification of a visible edit is a big part of what attracts people to the project. I fear this route would lead to a dramatic slowdon in "new blood". Pete/Pcb21(talk) 15:55, 12 Mar 2004 (UTC)
In the case of Wikipedia, I think three strikes are a sufficient criterion; three reverts by one user on a page within 24 hours, and the page is protected. Three revert or edit wars accumulated by one user, and he's banned for a week. Low enough thresholds to stop vandalism early, but high enough to allow room for a newbie learning curve. -- Seth Ilys 20:21, 12 Mar 2004 (UTC)
Seconded! We need to act quicly. Additionally: it should be possible to make a script to scan articles history to find out the alternating same users' edits (like that at Expulsion of Germans after World War II), isn't it? Perhaps that script could automatically protect page at 3rd round of revert? -- Forseti 09:40, 26 Mar 2004 (UTC)
May I ask from where this 24-hours period originates? While I'm all for measures against the extremely un-cooperative behavior certain regular wikipedians seem addicted to, I've found most of the ideas presented in this context to be far too lenient, and thereby maybe rather counterproductive. If the offenders get satisfaction from our indignation, then a 24-hours period is little more than the time to sleep, eat, and return from school/work/whatever. Would there be any support in this community for more perceptible sanctions, as for instance a bann lasting two or four weeks?--Ruhrjung 15:33, 14 Mar 2004 (UTC)
But Erik, you are trying to have edit wars solution be based uniquely on 2 or 3 wik-like people, while there are many other very regular and sometimes very respected contributors, who actually sometimes get into edit war and reversion war themselves in the heat of the moment.
For Wik-like people, yes, a temporary ban may be a relief. Now, for regular editors, I think it is a bad solution.
Suggestion :
What about a different policy depending on whether editors are listed as "frequently involved in wars" people ?
For those "problematic users", for example, though not mandatory, any 3 reverts session could grant either softban or page protection or slow-editing for 24 hours. Without the sysop doing the ban or the protection or the slow edit having to justify himself or to argue he did not commit sysop abuse.
However, none of these three actions would be mandatory. People could still consider applying one or another, depending on the person.
Now, for people not listed as "problematic", only page protection could be applied, eventually, after a certain number of reverts.
As for listing people problematic, I can just suggest a poll. If over 75% wikipedians agree a person is problematic, well, he may undergo harder punishement than others.
This will allow people like Wik to be blocked after 3 reverts. So, satisfy some wikipedians.
This will allow regular users only to see only article protection occuring for 3 reverts in most cases, so might satisfy all those against the ban for 3 revert rule.
This should satisfy both those willing to stop editors like Wik, and those saying he is part of the project.
This will sent him the signal "better behavior" -> "lighter punishment"
Similar to real life when a person gets heavier and heavier punishement as he multiplies offenses.
I have created Wikipedia:Quickpolls to add a review stage to the proposed 24 hour ban policy. This page would serve as a quick way to decide whether the policy has been violated in spirit and in letter. This way we address the problem of "gaming" the policy in various ways, and can adapt the policy over time rather than have to inflexibly apply it.
I would like to invite all who have voted against 24 hour bans to reexamine their arguments in light of this proposed system, to change their vote accordingly if they want to, and to add their support/opposition on Wikipedia talk:Quickpolls.—Eloquence 15:45, Mar 14, 2004 (UTC)
