Wikipedia talk:Quickpolls

Review of quickpoll process
The quickpoll process was originally implemented as a 30-day trial. We have now had 30 days to evaluate whether this process is beneficial. It seems appropriate for us to consider the experience and decide whether we want to continue using quickpolls.


 * To review the poll and discussion surrounding the initial design and implementation of quickpolls, see Wikipedia talk:Quickpolls/Archive1.
 * To review the actual quickpolls and related comments, see Quickpolls/Archive.
 * To review general quickpoll discussion during the trial period, see Wikipedia talk:Quickpolls/Archive2.
 * To review the current quickpoll policy and related discussion, see Quickpolls policy and Wikipedia talk:Quickpolls policy.

Question: Do you think we should continue using quickpolls?

Yes

 * 1) Quickpolls would work great if everyone accepted the three revert guideline. Sadly, somebody always wants to make an exception, thus negating the rule. It should be an absolute rule -- anybody who reverts three times in 24 hours for any reason whatsoever should get a 24 hour ban.  Tuf-Kat 18:54, Apr 20, 2004 (UTC)
 * 2) William M. Connolley 19:38, 2004 Apr 20 (UTC) Yes. Also, the 3-revert rule should become absolute, with ignorance no excuse. Being banned for 24h is not such a serious penalty that we need to agonise over it too long. Currently we are far too wishy washy. Suggestion: 2nd ban gets 2 days; 3rd 4; 4th 8... etc.
 * 3) Ruhrjung 19:58, 20 Apr 2004 (UTC) Yes, users of Wik's and 172's kind seem almost to have observed the three-revert rule thanks to the Quickpoll scheme. But modification is called for to hamper the usage of Quickpolls as a tool for one side against the other in an edit war or long-lasting POV-strife.
 * 4) Cgranade 23:12, Apr 20, 2004 (UTC)
 * 5) Eloquence* 01:43, Apr 21, 2004 (UTC) - improvements are welcome, particularly for the edit war policy
 * 6) Bryan 05:28, 21 Apr 2004 (UTC) Wikipedia needs some way to counter the people who think they can get their way simply by out-enduring the patience of the people who disagree with them. 24-hour bans seem like a reasonable way to "handicap" such tactics to me.
 * 7) Delirium 06:35, Apr 21, 2004 (UTC)
 * 8) Halibutt 14:47, 21 Apr 2004 (UTC) Totally agree with Ruhrjung.
 * 9) Totally agree with Halibutt, Bryan and Eloquence. --Uncle Ed 18:11, 21 Apr 2004 (UTC)
 * 10) Generally useful. Jamesday 21:13, 21 Apr 2004 (UTC)
 * Yes, but the same outcome should occur for both sides of the edit war. The current system that allows one person who broke the 3-revert rule to be banned and other person not is grossly unfair. Angela. 20:25, Apr 22, 2004 (UTC)
 * 1) VV 08:30, 23 Apr 2004 (UTC) Although QPs are flawed and biased, past ad hoc approaches were more so. And in my case a recurring issue was thrust into full view, where some stopped to see for themselves who was telling the truth  and experienced first-hand the sort of behavior I had faced .  This might have changed the minds of some who were otherwise following ideology or incomplete information, the best outcome I could reasonably hope for in this environment.
 * 2) Tuomas 23:22, 23 Apr 2004 (UTC) Quickpolls do not solve the underlying flaw of Wikipedia, and in addition they increase a tribal mentality which is bad when a focus on wordings and arguments is needed. But they make an important difference if they improve the social climate. Single users and sysops who ignore to cooperate harm wikipedia much more than flawed Quickpolls. The three-revert rule can be seen as a codification against uncooperative behavior at its worst. Quickpolls give hope that the wiki-community actually cares, and don't accept everything from wik, 172 and their ilk.
 * 3) John Gohde, aka Mr-Natural-Health 02:42, 25 Apr 2004 (UTC) -- Editors are against Quickpolls because they are changing the balance of power. Ergo, the quickpolls are working great!
 * 4) Comrade Nick I love the quickpoll process, because it's democratic!
 * 5) Arno Yes. Its decidedly efficient. One gripe - not everyone may find out about a particular quickpoll in time in order to be able to vote.
 * 6) Everyking 19:52, 5 May 2004 (UTC)

Yes, but only for certain cases
The quickpolls policy currently allows quickpolls when: If you support quickpolls for some of these situations and oppose them for others, please indicate in your vote the situations for which you support quickpolls.
 * 1) someone violates the three revert guideline
 * 2) a sysop repeatedly misuses a sysop capability
 * 3) a signed in user goes on a "rampage" of some type
 * 4) a signed in user confesses to deliberate trolling


 * 1) Only for cases 3 and 4. In my mind, quickpolls are very useful for situations like we had with Plautus satire, or maybe Bird.  In other words, situations where there's a consensus that someone is a major disruptive influence, and something needs to be done faster than the arbitration process can work.  They shouldn't be used for routine disputes.  I dislike the current situation where there seems to always be a quickpoll going on, usually involving conflicts that really aren't worth the attention of the whole community. Isomorphic 17:57, 20 Apr 2004 (UTC)
 * 2) Only for cases 2, 3 and 4. I'd like some method to enforce the 3-revert rule, but with Quickpolls, enforcement turns into a popularity contest and becomes asymmetric. -- Cyan 18:15, 20 Apr 2004 (UTC)
 * 3) Agreed with Cyan. It is frustrating to try to be fair and vote to ban two users only to have one of them not punished because the community perceives they are "right".  I think it increases the perception that we are unfair, rather than reducing it. Jwrosenzweig 18:46, 20 Apr 2004 (UTC)
 * Yes, quickpolls would work great if everyone accepted the three revert guideline. Sadly, somebody always wants to make an exception, thus negating the rule. Tuf-Kat 18:54, Apr 20, 2004 (UTC)
 * 1) Agree with Cyan and Jwrosenzweig.  It seems to work reasonably for cases 2, 3 and 4. Warofdreams 18:54, 20 Apr 2004 (UTC)
 * 2) for cases 2,3, and 4.  I don't think it works at all for the three revert rule. john 20:34, 20 Apr 2004 (UTC)
 * 3) Only for case 2. anthony (see warning)
 * 4) Case 2, for now. (see UninvitedCompany's comments, below) Martin 23:03, 20 Apr 2004 (UTC)
 * 5) Quickpolls is a powerful tool that allows the community to deal with a troublesome few. But it should only be permitted for specific cases. -- llywrch 00:22, 21 Apr 2004 (UTC)
 * 6) I think in case 2-when dealing with a sysop a quickpoll would be appropriate to rein in those who have authority. In other instances quickpolls are usually just a venue for people to make personal comments rather than vote on actual behavior. I am voting in this poll because others said it would be ok but for the record-I have not be on three months. GrazingshipIV 01:06, Apr 21, 2004 (UTC)
 * 7) Cases 2 and 3. SweetLittleFluffyThing
 * 2, 3, 4, --Voodoo 23:20, 21 Apr 2004 (UTC)
 * 1) Cases 3,4 Wikimol 20:46, 22 Apr 2004 (UTC)
 * 2) Case 1 possibly yes, not sure about the others. I highly respect the desire to back sysop decisions by community opinion, but I don't think Quickpolls are the right way. Philosophical background: Quickpolls add still more bureaucracy, and we already have way too much of that. People violate policies all the time, so what? Other people correct the damage done. Methinks we need fewer policies, not more (to be quite honest, three would be enough: Respect your fellow Wikipedians / Respect NPOV and the GFDL / When in doubt, apply common sense). However, the edit war problem must be solved, and if indeed the quickpoll process was helpful in that respect (but I am not sure about that), I guess it may be justified. Kosebamse 15:10, 25 Apr 2004 (UTC)
 * 3) All cases specified.  --Merovingian &#8597; Talk 18:36, Apr 28, 2004 (UTC)
 * 4) Cases 2 and 3.  Case 1 is trivial, Case 4 can be dealt with a ban. silsor 18:18, May 1, 2004 (UTC)
 * 5) *(William M. Connolley 18:27, 2004 May 1 (UTC)) I'm baffled. In what sense is case 1 trivial? Review the votes on 3-revert cases.
 * 6) **Review the definition of trivial. silsor 21:23, May 2, 2004 (UTC)
 * 7) ***(William M. Connolley 21:45, 2004 May 2 (UTC)) I don't understand you: what do you mean by case 1 is trivial? If you look back over the history of those quickpolls concerning 3 reverts, you find lots of arguments and discussion - few if any were trivially solvable.
 * 8) ****I think the idea is that violations of the three revert rule are fairly trivial, and ought not involve the effort of a quickpoll. john 21:56, 2 May 2004 (UTC)
 * 9) *****(William M. Connolley 22:17, 2004 May 2 (UTC)) Which is why I said, review the quickpolls on 3-revert cases: you'll quickly find that what you say isn't true.
 * 10) ****Of course, they were extremely acrimonious, and led to much bad blood, and so forth. That doesn't mean that the violation itself is not a trivial one for which a 24 hour ban ought not be the punishment.  In fact, I think the very acrimony of the discussions blew the issues way out of proportion.  Edit wars should lead to page protection, period, except in cases of clear vandalism. john
 * 11) *****(William M. Connolley 09:40, 2004 May 3 (UTC)) Personally I agree: violate 3-revert-rule and you get banned for 24 h, no excuses. But excuses have been accepted: ignorance, or defending against "vandalism".

No

 * 1) I do not think they are useful. Cases like P.Satire and Bird are better dealt with "in vivo" and "in real time" by one/several sysops, for my taste. Pfortuny 18:05, 20 Apr 2004 (UTC)
 * Admins do not have the authority to ban except for simple vandalism. That's why we started quickpolls in the first place – so admins could get the authority to quickly block a serious problem user. Isomorphic 18:56, 20 Apr 2004 (UTC)
 * Well, that's why I think an automated thing would be better. Nobody makes a decission, it is the computer. And for me, the above rule could do with a revision (if Bird were not simple vandalism, then we are not speaking a clear language). Pfortuny 07:20, 21 Apr 2004 (UTC)
 * In several cases it's been noticed that the nominated party hadn't actually met the requirements for action. While I've some support for automation, it does seem useful to be able to discuss whether someone actually qualifies for the proposed action. Jamesday 21:16, 21 Apr 2004 (UTC)
 * 1) They've proven to be rather useless, neither quick nor particularily effective. Scrap Quickpolls.  &#8212; Jor (Talk) 18:41, 20 Apr 2004 (UTC)
 * 2) They are mostly used to attack opponents, even persons who have not broken any rules. Even Ed Poor has been the target of such ridiculous attacks. The whole thing is unpleasant. Nico 18:50, 20 Apr 2004 (UTC)
 * 3) I believe that this experiment has shown that the time and effort required of the community to revert "rampages" and "trolling" and to deal with "excessive reverts" is rather less than the time and effort expended on quickpolls.  While the community may benefit from some mechanism for dealing with problem behavior, quickpolls isn't it.  UninvitedCompany 19:16, 20 Apr 2004 (UTC)
 * 4) Take this as experience, and start over with this experience as a guide to designing a new system. We can do better than this. -- Cimon 20:06, Apr 20, 2004 (UTC)
 * 5) agree with the above, we need a new and better system, one w clear rules and guidelines, not mob 'justice', and popularity contests of death Sam Spade 20:14, 20 Apr 2004 (UTC)
 * 6) *Realistically, if we get rid of QPs, we end up with no system at all again, which brings us back to unlimited edit wars and developers being the only control instance on sysop abuse. Do you have any alternative suggestions?--Eloquence* 01:45, Apr 21, 2004 (UTC)
 * I've said it above: make it automatic. How? That's what needs to be thought of (but not here+now, I think). Pfortuny 07:20, 21 Apr 2004 (UTC)
 * 1) I do not think they are working well--they only serve to exacerbate existing tensions between users rather than to resolve conflicts in the long term. In certain cases, responses should be much quicker than what a quickpoll allows for. It is also difficult to distinguish between valid users and sockpuppets, etc. I hate the idea of debates between users being the first thing we see at the top of a page, and I am worried that QuickPolls is turning into a popularity contest, rather than addressing issues. I therefore vote to end them. Oh, and did I mention that I hate the three revert rule, because it fails to consider the context of the reverts. There are cases where more than three reverts should be allowed and even encouraged. The three revert rule is not sacrosanct per se: it was established to deal with a specific problem, and if a troublesome user is taking advantage of that, that user should be revrrted. Danny 11:28, 21 Apr 2004 (UTC)
 * 2) They're not very useful and half the time, it's whether the user is well-liked or not, not a policy-enforcing tool. ugen64 20:52, Apr 21, 2004 (UTC)
 * 3) I thought it was a good idea at the start, and it certainly needs a system to replace it, but at present, they're simply not working - at all. Ambivalenthysteria 14:09, 22 Apr 2004 (UTC)
 * 4) I was going to write exactly what Ambivalenthysteria said. moink 19:56, 22 Apr 2004 (UTC)
 * 5) NO QUICKPOLLS. Never.  Not only is three reverts sometimes very justified, but it doesn't actually do anything to the real problem users who know how to play the game.  Start going after the f***ing trolls, not people who are trying to correct and article constantly butchered by wackjobs and idiots. Lord Kenneth 01:50, Apr 23, 2004 (UTC)
 * 6) Voting isn't a natural wiki process. Quickpolls are another artificial complexity on top of what, to the casual amateur participant, is a already an impossibly arcane tangle of bureaucracy.  Let's cut all the red tape, and get back to basics.  Wikipedia is a wiki - that is its unique strength.  I think we should stop trying to micro-manage everything, and allow the naturally democratic decentralised nature of wiki processes to come to the fore again.  If we only had enough wikifaith to allow Wikipedia to "find its own level", then I'm convinced that most of the problems that concern people would turn out not to be problems at all.  GrahamN 00:32, 24 Apr 2004 (UTC)
 * Quickpolls weren't votes though, at least not the way I understood them. They were polls - an attempt to get a sense of what the community thinks. Isomorphic 22:25, 1 May 2004 (UTC)
 * It seems to me that "what the community thinks" would emerge naturally, if only people would show a little patience and tolerance. GrahamN 18:51, 2 May 2004 (UTC)
 * 1) No. Mark Richards 00:09, 25 May 2004 (UTC)

Suggested modifications
This space is to consider possible changes to improve quickpolls


 * Most of the problems actually helped through quickpolls have been new users who are not contributing in good faith. In light of this, why not give greater discretion to sysops in dealing with this?  Bearing in mind Don't bite the newbies, we could still permit discretionary bans and blocks for new users (less than 30 days old?) who are disruptive and who have not made any good-faith contributions.  UninvitedCompany 19:16, 20 Apr 2004 (UTC)
 * I may not be correctly interpreting current policy, but I think that admins already have this authority provided the user has made no good-faith edits. The problem was with disruptive users who made at least some legitimate edits. Isomorphic 21:16, 20 Apr 2004 (UTC)
 * I believe that admins can only block for vandalism now, not, for example, for a pattern of POV edits indicative of an agenda, trolls on the policy pages, or the like. Presently we define vandalism narrowly.  Perhaps that should change.  If not, we can just use SoftSecurity as we did before Quickpolls. UninvitedCompany 21:33, 20 Apr 2004 (UTC)


 * Someone above suggested that instead of Quickpolls, we should just allow admins more leeway. That goes too far, since we wanted quickpolls precisely so that admins wouldn't make unilateral decisions.  However, we could say that only an admin can start a Quickpoll.  An admin sees a problem and decides "this user should be blocked for a while," but since the user isn't a simple vandal, the admin has no authority to block.  So, the admin starts a Quickpoll to get authorization. This strikes a balance between our old system and the current Quickpoll system.  It would cut down on people starting Quickpolls over any random dispute, since it requires at least one admin to believe that action is needed. Oh, and if we did this, it should also be with the understanding that it is NOT going to be used for enforcing the 3RR rule.  What happened to the days when we just protected pages to stop edit wars?  That was easier.  Isomorphic 22:01, 20 Apr 2004 (UTC)
 * Given the rising number of sysops, maybe it would be an acceptable idea to keep controversial pages semi-constantly protected, and to make changes to the articles only after establishing consensus on the respective talk-page? It would in the long run probably mean a lot more work for sysops, though. --Ruhrjung 22:13, 20 Apr 2004 (UTC)
 * Well, I don't like the idea much, but I must say that since it was unprotected, the great majority of edits to terrorism have been reverted. That would seem to support your suggestion. Isomorphic 22:57, 20 Apr 2004 (UTC)


 * As I mentioned on wikipedia talk:quickpolls policy, polls should be expired after 24 hours (not 48), or immediately the action finishes, since they're meant to be fast. Leave a little note with the taken action behind.


 * You might have them only startable by people physically able to take the proposed action. So, only a steward or developer could start a poll to desysop someone. Martin 23:23, 20 Apr 2004 (UTC)~
 * I think this might be a good plan. It doesn't actually add to the power of admins, developers, stewards, etc. (because the community gets final say) but it should weed out a lot of the constant pointless accusations. Isomorphic 01:01, 21 Apr 2004 (UTC)
 * I too like the sound of this, with the proviso that the sysop/developer who starts the quickpoll shouldn't be involved in the problem (i.e. if I'm in a revert war with a user, I can't quickpoll them just because I'm a sysop). fabiform | talk 07:24, 21 Apr 2004 (UTC)
 * For my taste, this "takes the power out of the people", and goes against equality. What if "the people" want to desysop or desteward someone but no other sysop/steward will start the poll? Just asking for an action to be taken ought not to  be prevented for "lack of power". Freedom of speech includes freedom to complain.Pfortuny 10:54, 21 Apr 2004 (UTC)
 * I agree, this could also ensure that the introduction of each quickpoll were properly made, which has been a weakness during this 30-days trial. --Ruhrjung


 * Remove the ability to revert from non-admins.
 * And, how would I use my sandbox with out the power to do a revert? -- John Gohde 22:06, 12 May 2004 (UTC)
 * Create a goodly number of "mediators with power" who are accessable, can do all the things a quickpoll can, and whose descision is subject to public review and is able to be overturned by varying degrees of majority (depending on the number of people voting)
 * Create a system for anonymously alerting afor mentioned empowered mediators of conduct violations
 * Create a body of clear rules and guidelines, short and sweet, vague and flexable, which are to be enforced.
 * These are my suggestions. Sam Spade 13:29, 21 Apr 2004 (UTC)


 * As before I would support considerably longer bans than 24 hours,
 * and I would support defining un-cooperative behavior (as demonstrative and systematic disregard of edit summaries and talk pages) as a justified reason beside intense reverting habits for banning (or other remedies). --Ruhrjung


 * Create new remedy - partial ban - prevent specific user editing specific pages, but for longer (e.g. week). Obvious diasadvantage - it would be more attractive to abuse Quickpolls against opponents in edit wars - so only together with some change preventing Quickpoll abuse.
 * Only an admin should have the right to start Quickpoll, but admins should be bold.Wikimol 20:46, 22 Apr 2004 (UTC)
 * 24 hours is too short a time period. I would make it one week.  We need to be slowing editing down in this place and one day voting is for people with a short attention span. -- John Gohde 22:02, 12 May 2004 (UTC)

Other comments
I'd like to see some form of automation for condorcett voting --this alllows ranking of several alternitives. Hottly disputed subjects might also benifit from a split page line by line "rebuttal" or alternate opinion format wiki deuling? But I am interested in trolproofing the concept. perhaps with "voting and forking" of "sections and versions". It seems appealing to let memes rise on their own. -ws


 * We need to think of automatizing the whole thing. It takes too long to solve a "quick"poll (for my taste) and so they lose their utility. I think this is not the place (now) for more thinking, so I stop here. Pfortuny 18:04, 20 Apr 2004 (UTC)
 * I don't remember to have seen Quickpolls used against misusing sysops or users confessing trolling. If this is really so, maybe it's premature to "evaluate" functions we haven't tested yet. --Ruhrjung 20:02, 20 Apr 2004 (UTC)
 * The Ed Poor quickpoll was for misusing sysop capabilities. The scenario of confessed trolling was a late addition to the policy. --Michael Snow 20:28, 20 Apr 2004 (UTC)
 * And don't forget: I was exonerated by a huge margin. You know, a note on my talk page would have worked better. I still have no idea what you all want me to do, next time a similar situation comes up... --Uncle Ed 17:20, 21 Apr 2004 (UTC)
 * Which is precisely why SysOps are unable to handle the problem of disruptive editors. They are clueless and are just as likely to side with the wrong editor and to take the wrong corrective action. -- John Gohde, aka Mr-Natural-Health 02:34, 25 Apr 2004 (UTC)
 * Quickpolls seem to work only to prosecute users who have no following. If a user is part of a faction which is pushing a point of view that faction (with the addition of innocents who join in) support them resulting in a ambiguous result. In some cases one of the two equally guilty parties to a dispute has been banned while the other was not. Fred Bauder 10:39, Apr 21, 2004 (UTC)
 * In your opinion, have the votes from sysops been more about the facts than the arguments? Perhaps a system where only sysops can vote, but non-sysops can explain things as far as necessary, might have merit as a 30 day trial to see if it works better? Jamesday 21:11, 21 Apr 2004 (UTC)

Quickpolls are polls among Wikipedia regulars on issues that need to be quickly resolved. Can someone explain to me why review of sysop actions should be done in a Quickpoll anyway? We have Requests for review of admin actions AND Possible misuses of sysop rights for reviewing sysop abuses. How often are we going to need emergency desysoping? The only cases I can think of involve a sysop going on some sort of rampage, and in that situation you could Quickpoll and ban, just like you'd do for any other user. Isomorphic 19:32, 21 Apr 2004 (UTC)


 * Sysops can unblock themselves, so you'd probably need to quickpoll, unsysop, and block. At which point, why not just unsysop, if it's the sysop powers that are the problem (eg, a sysop starts blocking every editor of Wikipedia). Martin 23:27, 23 Apr 2004 (UTC)

Three reverts rule - if that will be enforced diligently, uncooperative users will just do revert-edits instead of normal reverts, e.g. change word order slightly. Wikimol 20:46, 22 Apr 2004 (UTC)

Moratorium on three revert quickpolls
Based on the above votes and discussion, I suggest a moratorium on three revert quickpolls for the next four weeks at least. During this time we may think about other ways to enforce the three revert guideline. I do agree that some of the past QPs on 3RR violations effectively turned into popularity contests, which is unfortunate.

One route I think is promising to address edit wars is to give sysops more power in enforcing a cooldown period - right now, they are already allowed to protect pages, we may simple add the authority to ban users for 24 hours if they have been warned, and if the enforcement is consistent. If the edit warrior is a sysop, we may ask them to voluntarily stop editing for 24 hours and call in a steward or developer if they refuse.

I suggest that the other quickpoll conditions stay in place.--Eloquence* 20:34, Apr 22, 2004 (UTC)


 * I agree that we have to call off enforcement of the 3-revert limit through quickpolls. Really, the only situation for which I see a strong need for a quickpoll-type mechanism is the rampage, ill-defined though that term is. I think all of the quickpolls that led to a ban could qualify as rampages. A rampage could be vandalism, inflammatory/offensive posts, excessive reverts (probably significantly more than 3), misuse of sysop capabilities, and perhaps more that I can't foresee.


 * Anything that's not a rampage, including individual instances of exceeding 3 reverts, sysop misbehavior, and confessed trolling (whatever trolling means), can I believe be handled through the normal dispute resolution process. They don't require the urgency of a quickpoll. However, I have an alternative idea to float as a trial balloon, which I will add momentarily. --Michael Snow 22:58, 22 Apr 2004 (UTC)


 * Moratorium should be for all use of Quickpolls. They were introduced/marketed as a means to enforce the three-revert rule. To keep it for the hard-to-define "rampage" and "confessed trolling" is ridiculous and will only lead to furtherly worsened reputation. See the current call for quickpoll on TDC.
 * --Ruhrjung 17:15, 23 Apr 2004 (UTC)
 * Actually as far as I know, Quickpolls were introduced so that admins could get community backing in blocking disruptive users and not have to make judgement calls. We had a period where admins were acting unilaterally because they had no other choice except to wait weeks for the Arbitration committee to make a decision.  In the case of Plautus satire a couple months ago, we had a poll, and a strong consensus to ban a very obnoxious user, but no means of enforcing this quickly because the Quickpoll concept didn't exist yet.  The fact that such judgement calls are hard is precisely why you have a Quickpoll.  If it were easy, you could trust an admin to make a decision unilaterally. Isomorphic 21:51, 23 Apr 2004 (UTC)

(William M. Connolley 20:49, 2004 Apr 23 (UTC)) I can't see the sense in this moratorium. What, in the absence of some enforcement, enforces the 3 revert rule?


 * There's still arbitration, protection, peer pressure, and the ability for the protecting sysop to choose to revert to the version disliked by the most frequent reverter (see protection policy). Not much, but it is something. Martin 23:29, 23 Apr 2004 (UTC)


 * (William M. Connolley 21:27, 2004 Apr 24 (UTC)) Arbitration is too slow and heavy. Protection is bad. Peer pressure is weak without enforcement.

(William M. Connolley 14:28, 2004 May 1 (UTC)) The current state of voting is 15-15-21 so support for a moratorium isn't obvious.

Alternative proposal: Slowpolls
In the case of Lord Kenneth, Tannin blocked Lord Kenneth to stop a rampage, then called a quickpoll to confirm the blocking as a 24-hour ban. This action was supported by the quickpoll. I propose that we start from this as a model.

I suggest that we allow admins the authority to impose 24-hour blocks of signed-in users for rampages, based on their individual judgment. However, this must be subject to community review, and I also feel strongly that the admin who takes such an action should have something on the line. My idea is that an admin who blocks a user for a rampage should put admin status at stake when doing so. This would be done through what I'll call a slowpoll, in contrast with quickpolls.

The procedure would be something like this:
 * The admin must, obviously, warn the rampaging user before imposing the block. If the rampage continues, the block can be imposed.
 * After imposing the block, the admin must promptly start the slowpoll (say, within an hour).
 * Failure to warn the rampaging user, or failure to start the slowpoll, would be grounds for immediate de-sysoping.
 * The slowpoll is a vote on whether the admin's decision was justified.
 * The slowpoll is open for 7 days to allow wider community input. Some kind of advertising is probably called for.
 * The decision is based on a simple majority vote - not a consensus. If the admin cannot get majority support for the action, it is obvious that it was not supported by a consensus of the community, and the admin should give up admin status. (tie goes to the admin, presumably)
 * If the initial results of the slowpoll indicate that the decision was controversial, the block should probably be reversed.
 * Anyone who loses admin status through this process may re-apply in the normal fashion.

This should cause admins to consider carefully and make sure that the block is justified. We won't have problems with repeat offenders, because any admin who imposes an unjustified block will lose the ability to do so in the future. Potentially, admins will have only one bullet in their gun (sorry for the violent imagery). Meanwhile, we can have the benefits of quick action where it is needed, while also allowing the community to engage in more careful deliberation.

As I said, this is a trial balloon, so feel free to comment, or just shoot it full of holes. --Michael Snow 23:32, 22 Apr 2004 (UTC)

I STRONGLY oppose having the admin lose his admin status if the slowpoll doesn't go in his/her favor. The chance of this happening would put a serious damper on anybody wanting to block somebody who went on a rampage. RickK


 * I would agree with RickK, and suggest as an alternative -- the admin who loses this slowpoll will be blocked for 24 hours (a block not to be relieved by any means). I can think of trolls I'd risk a 24 hour ban to stop, but it's hard to envision putting admin status on the line when one never knows who's paying attention to polls and how they'll perceive a situation.  The threat of a day's ban (and the implied lost respect, given that the community disapproved of my decision) would be adequate penalty for unbacked action, I think.  After all, all I'd be threatening someone with if the slowpoll goes my way is a day's ban and implied loss of some respect.  Makes sense to me. :-) Jwrosenzweig 23:40, 22 Apr 2004 (UTC)


 * I object to this, too. Why any admin who's trying to solve a crisis should risk any retribution is beyond me.  RickK


 * I don't know if I like this principle in general yet, but I don't like the black/white "somebody loses" nature. How about instead of 51% means someone gets blocked and 49% means someone gets their sysop privileges removed, put the edges at like 80% and 20%, or 67% and 33% or something.  That way when things are borderline we (as a community) can say "We disagree with your judgement, but not so much as to desysop you."  moink 00:04, 23 Apr 2004 (UTC)


 * Would Jwrosenzweig's 24-hour ban be better? I think it's important that admins have good judgment about consensus, and a majority against the admin is clear enough to me that the consensus wasn't there. Frankly, I suggested admin status because it's "not supposed to be a big deal". You can keep editing as you did before. I'm concerned that being overly attached to adminship only goes to support the arguments of those who say that adminship is a status symbol and that Wikipedia is run by a cabal. "Sysops are not imbued with any special authority", or so we say. --Michael Snow 00:21, 23 Apr 2004 (UTC)


 * You make a good point, and I could certainly live without my custodial priveleges for 24 hours (or the rest of my life if I had to). But desysopping, even temp-desysopping, is often seen as a pretty severe punishment, just by virtue of its infrequency and the fact that Jimbo has been loathe to use it as a disciplinary tool.  I'm afraid that this might make us lose some good contributors (sysops on average are excellent contributors) by upsetting someone who made a judgement call without malice.  moink 00:26, 23 Apr 2004 (UTC)


 * I object because putting any punishment on the admin who was trying to solve a crisis in good faith just shows that the community has no faith in that admin and it's a slap in the face to those of us trying to keep the place clean despite thosewho laugh off vandalism as no big deal. RickK

Well, I for one would support the 24-hour ban; desysopping goes too far, though. There needs to be something at stake for the instigator of the poll, just to avoid handing all power to sysops. Meelar 00:38, 23 Apr 2004 (UTC)

I'd support the warning > block > poll > ban timeline, but not punishment of the admin. Why on earth would we want to punish sysops? We already assume that they won't block people arbitrarily or go on rampages. Exploding Boy 08:37, Apr 23, 2004 (UTC)

The current system works just fine in the cases it was intended for. There's no reason to go looking for new means of enforcement. Right now, the bigger and more obvious the problem, the faster a QuickPoll can get you consensus. The problem is that the threshold for starting a QuickPoll has been too low, so we get long drawn-out arguments on cases that really aren't urgent at all. Even worse was when we felt obligated by the system to start a quickpoll on a user who was acting in good faith (Stevenj) just because of the 3RR. The only thing we need to worry about is how to have quickpolls only when we actually need them. Isomorphic 16:39, 23 Apr 2004 (UTC)

I support this proposal, as long as possible cases for blocking/slowpoll action are narrowly defined by specific policies, to avoid lynch-mob rule.

Also if a sysop is personally involved in a dispute with the user, he cannot be assumed to be neutral. Every police or court system has rules to prevent such conflict of interest.

About desysopping: The point is not to punish the sysop; it is to make pretty much all sysops spend some time off-duty - so they realise the uncomfortable situation of a non-sysop; so they will maybe behave more nicely next time they are sysop. This is how judges are judged - by how likely their verdicts are to be reversed on appeal. But likely, sysops will fight against this proposition which remove them their godpower JRR Trollkien (see warning) 22:06, 23 Apr 2004 (UTC)


 * May I ask what goodies or godpowers you actually think of, and how being a non-sysop is that very "uncomfortable"? /Tuomas 23:26, 23 Apr 2004 (UTC)


 * You should ask that to people like RickK, Jwrosenzweig or Meelar. As Michael Snow said, 'I'm concerned that being overly attached to adminship only goes to support the arguments of those who say that adminship is a status symbol and that Wikipedia is run by a cabal. "Sysops are not imbued with any special authority", or so we say.'. JRR Trollkien (see warning) 23:32, 23 Apr 2004 (UTC)


 * I ask you, since it's you who uses it as an argument. /Tuomas 23:46, 23 Apr 2004 (UTC)


 * My objections have nothing to do with any supposed godlike powers. They have to do with the slap in the face that desysoping a sysop gives.  And as Jimbo has said, he opposes desysoping in general.  RickK 23:36, 23 Apr 2004 (UTC)


 * In addition, sysopship is given pretty easily to people who deserve it. Unlike trolls.  RickK 23:37, 23 Apr 2004 (UTC)


 * Desysoping a sysop should be no severe punishment or "slap in the face" -- EVERY sysop should spend some time off-duty, based on how likely their actions are to be reversed. If, as you say, "sysopship is given pretty easily to people who deserve it", they should have no problem re-applying in the normal fashion. JRR Trollkien (see warning) 23:42, 23 Apr 2004 (UTC)

This seems like both a terrific waste of time and a good way to ensure that no one ever gets listed for a "slowpoll." Sysops have no godlike powers. Most of them spend their time quietly editing and watching the articles they care about. Probably the "power" they use most often is the rollback. They're not immune from punishment as it is; few sysops abuse their so-called power, and if they do they're called on it pretty fast. I think this proposal is ridiculous. Exploding Boy 00:09, Apr 24, 2004 (UTC)

I have no sympathy for the quirks of this proposal, but I can't agree with your rosy picture. If sysops do wrong, they are considered human or well-meaning and defended by other sysops. This is natural. It's how social beings are wired. But keep it in mind, please! /Tuomas 00:40, 24 Apr 2004 (UTC)

A quickpoll was not needed when I went on my rampage. I should have been instantly blocked according to the rules, quickpoll or no. I do not apologize for my behavior, as I'm sick of how the corrupt system works. It's ridiculous. Quickpolls do little more than hurt users with good-intentions on certain topics. I propose greater arbitration for the quacks and goons who pollute articles, not silly "let's get someone if they revert > 3 times IN A 24 HOUR PERIOD EVEN THOUGH IT ONLY SAYS "PER DAY" WHICH WOULD MEAN AFTER 12:00... gah, so corrupt...

Fix the system. I'm sick of being punished for trying to keep an article NPOV and accurate. I've been working on [scientific_skepticism] for months now, trying to keep Reddi's crazy edits from misleading people, and few other people seemed to care-- their policy is "let the wikiprocess sort is out", which is basically saying "Let someone else do it." What happens when everyone else decides to just let the "wikiprocess" sort it out? NOTHING GETS GODDAMN DONE.- Lord Kenneth 02:24, Apr 24, 2004 (UTC)

Re: Tuomas's comments: there are at least two admins whose behaviour I strongly disagree with, but they confine themselves to being a nuisance on certain articles rather than abusing their admin powers. So far I've not seen them user their admin powers (blocking, for example) against other users -- but if they did I've no doubt they'd be called on it pdq. Exploding Boy 13:22, Apr 24, 2004 (UTC)

Compare with the Quickpoll(s) against User:172, or earlier discussion on him scaring users away. --Ruhrjung 19:24, 24 Apr 2004 (UTC)

Limitations of democracy
It's not very useful to rely on voting, but we can't seem to live without it either. How can we solve this conundrum?

Let's make the guidelines so clear that ALL admins will feel empowered to use their judgment. If someone gets out of line, we can try:
 * teaching
 * warning
 * blocking

I would say that in case of doubt, an admin should consult other contributors (regardless of "rank").

And if any action (like blocking) is taken, definitely report this to the community for review.

We've tried all sorts of systems. When the only way to block a user was to run a SQL query to mangle their password, that pretty much left it up to developers like Tim Starling, Brion Vibber and Eloquence. (By courtesy, I was called a developer even though I never did anything but run SQL queries and make annoying comments about speeding up the server! :-)

Then sysop powers were broadened to include the "Block this user" button. But we haven't come to a stable consensus on where and how this power should be used. Be too nice, and trolls over-run the dungeon (so to speak). Be too aggressive, and you run the risk of being called on the carpet yourself for "abuse of power". What's an admin to do? --Uncle Ed 12:39, 26 Apr 2004 (UTC)

Rampage
What constitutes a "rampage" anyway"? I think I may have just been a victim of such by 172, who has been following me around rolling back even simple housekeeping edits on all sorts of articles. -- VV 05:43, 28 Apr 2004 (UTC)

Let he who is without sin...while I can't defend what would appear to be 172's decision to revert as many of your recent edits as possible, he seems to have been triggered by your decision to revert his edits to Muammar al-Qaddafi (which appears to be utterly unjustified, to me, and certainly not actually defended) and George W. Bush (which, I get the idea why you didn't like his version, but the changes he made were such as to at least require some remarks from you on the talk page). Until you do that, your reverts of those articles are equally unjustified as 172's reverts of your articles, which I imagine was his point. Sigh, this is just childish on both your parts. john 05:57, 28 Apr 2004 (UTC)


 * 172 sticks controverisal assertions into articles - one could call it trolling - as that whole mess at Pinochet proved. He starts an edit war whether I try to explain or not, so I removed POV/misinfo without immediate explanation in these two cases, much as I feel he has done to me.  I noted on the talk page my objections re Bush, and as for Qaddafi, besides the slanted phrasing of the new paragraphs, I don't believe it's proven that Qaddafi's daughter was killed; at best, it was his adopted daughter, and I don't think we're even sure of that.  Anyway, these objections were valid ones and cannot be equated to his rampage of reverts against me (nor violation of protection policy). -- VV 06:14, 28 Apr 2004 (UTC)


 * Slanted phrasing does not justify a wholesale revert, and I don't see as the phrasing was especially slanted. I have no idea about the daughter - the version you reverted to said adopted daughter.  Encarta says it was his daughter, says my quick google search.  At any rate, I don't see why an adopted daughter does not count as a daughter.  As to George W. Bush, your comments seem to have more merit, but you ought have added them at the time, and, to be honest, you ought have, as you say, picked through and made changes rather than reverting.  Given the intensity of editing on the George W. Bush article, and given your own history with 172, it was unwise of you to move into that one anyway.  I'm sure Cecropia, for instance, would have been all over any left-wing bias 172 may have introduced, without the need for this kind of crap. john 06:24, 28 Apr 2004 (UTC)


 * John, how would you handle it if he reverted your edits to articles over and over again arbitrarily just to prove the point that he can keep you from editing articles? It's clear that he didn't even read either of the edits that I'd made. After a while you'd say to hell with civility and 3-revert rules. I might've made that string of changes just to prove a point, but you know that I never insist on making changes to articles without a reason. 172 06:05, 28 Apr 2004 (UTC)


 * Keep you from editing articles? You've been editing dozens of articles lately, and just now I took exception to two.  You don't see me coming down on Zimbabwe and imperialism (although for all I know you're POV-pushing there too). -- VV 06:37, 28 Apr 2004 (UTC)

Well, some of the edits you reverted of his looked pretty ridiculous. Caldera and Susning.nu, for instance - on the latter you reverted to a version containing a sentence fragment. At any rate, his revert on Muammar al-Qaddafi looks utterly unjustified to me. The George W. Bush one I can see where somebody might think it's POV, but there was absolutely no cause for a revert. And his responses suggest that, at least now, he has some reason for his actions. But seriously, this is ridiculous. Any revert of anything but pure vandalism should be accompanied by an explanation on the talk page. If there's anything salvageable, one shouldn't revert at all. I think that, whatever POV problems there may have been in 172's edits, they did not call for a wholesale revert. On the other hand, retaliation reverts are ridiculous too. What good does it do you to revert things wholesale? They don't stay that way, and it just leads to nonsensical edit wars. john 06:24, 28 Apr 2004 (UTC)


 * (John's comment is addressed to 172, but....) I definitely agree that wholesale reverting is rarely appropriate.  Unfortunately, it's how 172 has always done business (remember FOX News? state terrorism?), and the sort of careful salvaging you suggest I do has been too often rewarded with a rollback. -- VV 06:37, 28 Apr 2004 (UTC)

That's why you discuss problems on talk, and so on and so forth. Another person's poor behavior doesn't justify bad behavior on one's own part. And that applies to both of you. john 06:48, 28 Apr 2004 (UTC)


 * I acknowledge I've been increasingly impelled by frustration. But I still find that handicapping oneself with behavior restrictions when locking horns with 172 just means capitulation. -- VV 07:19, 28 Apr 2004 (UTC)

You do realize that he says exactly the same things about you that you do about him? At any rate, I'm going to sleep. john 07:44, 28 Apr 2004 (UTC)

Sleep well, john. If it's a "rampage", it's a rampage on the part of both. I thought you two (172 and VeryVerily, that is) had found an accommodation to stay away from each other, but I should have anticipated that it wouldn't last long. By the way, I didn't get any response to my follow-up about mediation from either of you. The longer this goes on, the more it looks like your personality conflicts will end up in arbitration. --Michael Snow 20:41, 28 Apr 2004 (UTC)


 * I battled 172 on two controversial articles; 172 wildly mass reverted my edits on a slew of articles, most wholly uncontroversial. Surely there is a difference?  I don't know what the status of said accommodation is; I'm still not touching a few articles, for now, but that does not mean the rest of Wikipedia is fair game.  No, sorry, I haven't put in the time to look deeper into the mediation option. -- VV 08:33, 30 Apr 2004 (UTC)

VV & 172, if it's really true that you two have been reverting each other's edits for reasons other than disagreeing with their accuracy or neutrality -- like just to prove you can, or to "harass" or "punish" -- then I think you both ought to take some time off from Wikipedia.

How about it guys? Take a vacation and chill out for a week! :-) --Uncle Ed 12:47, 30 Apr 2004 (UTC)


 * I haven't been doing that. Who says I have? -- VV 18:20, 30 Apr 2004 (UTC)

Could someone do something about Gene Poole? He reverted Sealand six times in 24 hours. --Wik 12:04, May 4, 2004 (UTC)

Strongly suggest you learn how to count.--Gene_poole 10:37, 5 May 2004 (UTC)

I counted 6 they are at 11.21, 11.12 on 5th May 1.34, 15.41, 12.54 and 12.49 4th May. Perhaps you should apologize for calling Wik a liar? theresa knott 10:56, 5 May 2004 (UTC)

5th May: 0.34, 10.12, 10.21. That's three in 1 day. It's a *daily* limit - not a rolling 24-hour limit. Latest example of Wik mendacity stands.--Gene_poole 08:40, 6 May 2004 (UTC)


 * No, it's a rolling 24-hour limit; remember, Wikipedia is time-zone-less. Wik is correct, although this is not the first such misunderstanding. -- VV 09:05, 6 May 2004 (UTC)


 * You might want to check your facts (as I've just done): http://en.wikipedia.org/wiki/Wikipedia:How_to_revert_a_page_to_an_earlier_version#Revert_wars_considered_harmful ; "Never revert the same article more than three times in the same day". "Day" is not defined - in fact there's whole paragraph describing just how undefined it is.--Gene_poole 09:25, 6 May 2004 (UTC)


 * You really take the biscuit Gene. Look at the reverts I mentioned above. Even by your definition (which by precedent is not the correct one, Wik has been banned under the 24 hour period interpretiation) you reverted 4 times in one day on the 4th May. An apology at this point is required don't you think? theresa knott 11:09, 6 May 2004 (UTC)

Cantus likewise keeps violating the rule, e.g. on Template:Europe. Is anyone going to do anything about this, or am I the only one the rule applies to? --Wik 18:51, May 5, 2004 (UTC)


 * Well, right now I think we're generally not holding quickpolls for simple violations of the 3-revert limit. See the discussion above. The rule that applies to you, Wik, is a separate ruling by the arbitration committee. I note that they said it was to be "in addition to any wider policy that the community may decide to apply." --Michael Snow 21:45, 5 May 2004 (UTC)

Might I suggest that Cantus ought to be put under the same rule as Wik - he flagrantly and repeatedly violates the three revert rule, and barely discusses it on talk. The only difference between him and Wik in this regard is that Wik is usually right. john 23:55, 5 May 2004 (UTC)

I agree, if cantus is aware of the three revert rule, and chooses to ignore it none the less, then he should be blocked for 24 hours. However at the moment we have no authority to actually block him, and until the quickpoll process is sorted out (whick will probably take ages) we are unlikely to have the power to do it in the near future. However a review of Revert tells me that I can revert to cantus's disliked version and protect. This is what I plan do every time I see him exceed 3 reverts in one 24 hour period. theresa knott 13:18, 6 May 2004 (UTC)

Alternative proposal: Symmetric 3-revert rule
It seems to me that the key objections to the quickpoll process is that some people get a pass for violations of the 3-revert rule because their position has more support or they are more popular than their opponents. (Certainly that's my view. :-)

I have a proposal to fix the asymmetric application of the 3-revert rule. Rather than nominating individual editors for violations, a quickpoll proposer can only specify the article in which the 3-revert rule has been violated. The proposer and others may then check the edit history for all violations of the 3-revert rule within a certain time period (more details below), listing them under the quickpoll header with links for the reverts of each violator. If the vote goes through, all violators listed receive a 24 hour block.

As with the previous implementation, I propose that this system undergo a 30-day trial period.

Time period
Simply specifying a time period like the 24 hours previous to the quickpoll posting leaves the proposal vulnerable to an asymmetry, as the last person to revert could be caught after other violations have passed the time limit horizon. Therefore, I propose that the listing must occur within 24 hours of the most recent violation (counting from the last violating edit), but violations that are within 24 hours of each other can (must!) be listed under the same quickpoll header. This means a listing could potentially extend forward and backwards in time without bound (more or less), allowing it to cover a complete edit war among multiple parties that extends over days or weeks.

Gaming the system
There is a certain ambiguity in what constitutes a revert. For example, a clever individual might try a number of mild variations on a theme and then claim that these do not constitute a violation of the 3-revert rule. In order to address this problem, I propose that if there is a legitimate dispute over whether a particular edit violates the 3-revert rule (i.e., disinterested parties have different interpretations), or some other ambiguity of the quickpoll process, then the quickpoll becomes invalid. This proposal addresses the problem in that a clever individual could expend time and effort gaming the system, but if he or she does so, his or her opponent will get a free ride even for outright reversions. The key is to preserve the symmetry of the application of the rule.

Discussion
I leave it to others to determine if this proposal has enough support to warrant an adoption poll. -- Cyan 14:34, 6 May 2004 (UTC)
 * At first reading the proposal looks good, and i will support it, although I'm not sure I understand the timing part. Why can't we just specify a longer time period of say I week from the poll starting. Anyone who makes three reverts in any 24 hour period in that week must be listed? theresa knott 15:42, 6 May 2004 (UTC)

Any hard time limit is subject to "horizon" effects. Consider: you and I go temporarily insane and get into an edit war. You revert last, then we come to our senses and let others work out a compromise while we log off and cool down; and so no one sees the need for a block. But a week later, another edit war breaks out between other people, and the the article is nominated for a quickpoll. In the time period under consideration, I have three reverts, and you have four, even though we both have actually accumulated a larger number of reverts.

Of course, it wouldn't be fair for you to be blocked in this situation. It may be extremely rare, but if does happen, it's bound to be upsetting, so I wrote the rule to avoid this kind of situation entirely. Maybe it's unnecessary, though. -- Cyan 16:06, 6 May 2004 (UTC)

I thought about making a similar proposal, where if two users both reverted, the quickpoll would only be valid if it included both users and applied equally to both. I'm willing to consider this, but there's still a practical problem. However you define a revert, in a two-person revert war one user has to revert first, and one user has to exceed the 3-revert limit first. Once one user exceeds the limit, the other user can stop reverting and get the first user quickpolled. A scenario:
 * X's 1st revert
 * Y's 1st revert
 * X's 2nd revert
 * Y's 2nd revert
 * X's 3rd revert
 * Y's 3rd revert
 * X's 4th revert
 * Y does not revert, quickpoll starts and X gets a 24-hour ban

In this situation, there is little to differentiate between X and Y, except that Y has technically avoided violating the rule. Basically, we would be letting two people walk all the way up to the edge of the cliff together, then watching one push the other over the edge. --Michael Snow 16:54, 6 May 2004 (UTC)

That isn't a feature of this scheme to enforce the 3-revert rule; that's a feature of the 3-revert rule itself. But I regard it as a good feature: the one who takes the first step up the cliff is the first one over the edge. If that person had instead tried to achieve a compromise, the entire problem could have been avoided. Note that this will only occur if there are only two opponents, and not during a pile-on edit war (for pile-ons, majority rules under the 3-revert rule, provided a sysop doesn't step in and protect). If there are only two edit warriors, then neither can really legitimately claim to represent community consensus, so I don't really have a problem if the one who escalates first is the one who is punished. -- Cyan 17:17, 6 May 2004 (UTC)


 * I agree that it's ultimately a product of the rule, but that may be why some people remain uncomfortable with converting what was a "guideline" into a "rule", and why people have concerns about our enforcement mechanisms. On the other hand, you can also argue that this possibility might deter edit warriors from being the one to make the first revert. That could go a long way towards improving the general atmosphere. --Michael Snow 17:40, 6 May 2004 (UTC)
 * I am one of the people concerned about our enforcemment mechanisms. I'm concerned that we don't have any! I've never seen the situation you describe above actually happen. Useually what happens in my experience is X rv Y rv X rv Y rv X rv Y rv (that's my lot} X rv Z rv (saw Y's edit summary in RC) but even if it did happen so what? A 24 hour ban is a very minor punishment. theresa knott 18:43, 6 May 2004 (UTC)

YAAP (Yet Another Alternative Proposal)
Cyan's version might fix the problem of preferential treatment when two editors are working on the same article. It will not fix that same problem when two editors are working on separate articles. Determining whether or not an editor broke the 3 revert rule is really a trivial task (providing that the rule is clear, and with the exception of cases where reverts are obfuscated by additional edits - the 'clever people' Cyan mentioned.) The policy should therefore be designed to ensure that *every* editor who breakes the 3 revert rule gets the penalty. Quickpolls are not necessary here. These matters should only go up one level in the dispute resolution process if an editor objects, and additional penalty can be imposed when such objections are found to have no basis (whithout any cost to raising an objection even in relatively straightforward cases too many people would do so.)

So, if this issue can't have a software solution (which I still think it should,) then I propose something like the following system:

Chose a group of people whose task it would be to automatically ban whoever reverts 4 times in a UTC day (this could simply be all sysops.) Create a special page where bans will be announced, much like the current Protection log.

For simple reverts the procedure would be the following:


 * 1) Anyone can post a notice about someone exceeding the 3 revert limit (on a special log page.)
 * 2) The next sysop who happens to see the log page will ban all the names on the list in the order in which they are listed (this ensures equal treatment.) If the 3 revert rule hasn't been broken he or she puts either "less than 4 reverts" or "not an obvious revert" next to the name listed.
 * 3) Another sysop checks the list and if there are no problems, puts 'confirmed' next to all unconfirmed entries in the list, before banning any new names on the list.
 * 4) The person who was banned can object and the matter is then passed on to arbitration. If the arbitration comittee decides that he was right, the two sysops involved will be prohibited from banning for 30 days (or a week.) If he or she was wrong, the ban will be extended for 24 hours.
 * 5) Anyone can object if someone wasn't banned even though the 3 revert rule was broken. The matter goes to arbitration and for sysops the panalty is the same as above. For the person raising the objection the panalty is either the same as above, or alteratively the person could be placed on a list of wikipedians with restricted rights for 30 days (for example no right to raise objections.)

In cases when reverts are obfuscated by additional edits:


 * 1) same as above
 * 2) same as above
 * 3) If there is at least one 'not an obvious revert' message next to a name, a poll starts under the name, and lasts 48 hours. All sysops can vote, and are required to follow strict guidelines when determining whether the edits in question are reverts.
 * 4) same as above
 * 5) Since the issue was not straightforward, and a decision was based on a poll result no panalties are issued to sysops.

Also, instead of banning someone for 24 hours a person breaking the 3 revert rule could simply be limited to 3 edits per article for 30 days, which could be enforced by software.

The above system is meant to fix the 3 things that didn't work in the quickpolls version.


 * 1) Interested parties could not start quickpolls. In any legal system the ability of interested parties to bring the issue before the justice system is the best guarantee that whenever a law is broken penalties are imposed.
 * 2) Anyone could vote, including people known for making biased decisions (or just stupid ones.) In a well functioning system such people are never promoted in the hierarchy.
 * 3) Those who voted had nothing to lose. Thus they were only concerned with following their agenda, and voted against banning if they agreed with the person making the edits, or if they felt the system wasn't working etc. In a well functioning system people always know that there is a personal cost to making the wrong decision, and generally are not willing to pay the price to support someone they like.

The 3 revert rule needs to be changed to 3 reverts in a utc day, (instead of the current 'there is a speed limit of 3 rods per something or other, and if that's confusing you should just slow down' version.)

The guidelines for determining whether an edit is a revert should be something like: "Any edit that includes either the removal or addition of a controversial phrase that was previously removed or added on the same day or the day before will be counted as a revert." To further clarify you could require the other party to identify the phrase in the summary before edits of that phrase are counted as a revert. --Voodoo 02:07, 10 May 2004 (UTC)

Any system that has a reasonable chance of avoiding asymmetric application of the 3-revert rule has my support. -- Cyan 04:03, 10 May 2004 (UTC)


 * Mine too theresa knott 08:29, 10 May 2004 (UTC)

The simplest solution
Why don't we go with the simplest proposal of all: the same edit being reverted three times triggers automatic page protection by the first admin that notices it occurring (one not involved in the dispute, that is.) Then keep it protected until consensus is hacked out on the talk page, with a temp version created if needed. This forces those involved to work out their differences, since you simply don't unprotect until they have. If someone violates the consensus solution after it's been achieved, as far as I'm concerned that's vandalism, and it can be dealt with as such.

Protection is simple, it ends the revert war, and it doesn't have all the finger-pointing and recriminations. Quickpolls shouldn't be for content disputes, and that's what almost all reverts are. All this legalistic nonsense about the 3RR just breeds disputes. Isomorphic 04:27, 11 May 2004 (UTC)

We already do that but it doesn't work. Pages get protected, talk stops, it looks like everything is a calm now so an admin unprotects. The war restarts. Some users manitain lists of pages that they intend to revert as soon as protection ends. The page has to be protected on one of the versions, the "winner/s" have no incentive to discuss alternatives. In the meantime "innocents" are unable to edit the page. I'm not interested in legislating content disputes, I am interested however in forcing certain users to stop revert wars of if that can't be done slowing them down to three reverts each per day. theresa knott 07:41, 11 May 2004 (UTC)


 * (William M. Connolley 08:51, 2004 May 11 (UTC)) TK has just proposed a solution I haven't seen explicitly mentioned before (have I not been paying attention? am I re-interpreting her words?) which is a limit of 3 reverts per user, rather than per-user-per-page. This sounds like a good idea. ps I agree with her re protection.


 * Yes, that's been mentioned before, and the problem is that a vandal who wrecks ten pages at once could not be undone with fewer than four people. That is, I couldn't note their rampage and clean it up.  Also, I suppose people might try to win edit wars by provoking a fight on some other page the other user cares about. -- VV 09:51, 11 May 2004 (UTC)


 * Yes you [WMC] were misinterpreting me (my fault though; I worded it badly) I often revert more than 3 times a day myself when dealing with vandalism. It has to be three reverts per person per article. theresa knott 10:30, 11 May 2004 (UTC)


 * I was referring to the comment by William M. Connolley, which I think means what I thought it meant :). -- VV 10:37, 11 May 2004 (UTC)


 * This can be solved by allowing people to revert clear vandalism without it counting towards their 3. I'm not proposing this to be done in software anyway. What would happen is that *after* someone complains you've hit 3+ reverts in 24h, then you have to stop.

I have a proposal of my own, somewhat related to one noted above. Its flaw is it presents software implementation issues. Earlier, it was suggested a user would not be allowed to re-edit a page they just edited until after some time had elapsed. I think this is a bit extreme, although I like the pressure to use "Show preview", but it could be adapted. I suggest the first three edits are "free", but for the fourth edit a wait of an hour is necessary, a fifth three hours, a sixth twelve hours, a seventh twenty-four. After 24 hours all is forgotten. Or some such cascading pattern. Just an idea. -- VV 09:57, 11 May 2004 (UTC)


 * Hmm but sometimes pages get very heavy editing, especially when they are new. I'm concerned about using software here because software is stupid. I have no problem for example a newbie making lot's of edits to a page. I would rather teach them about previewing, than force them to be unable to edit. I don't want anything complex that will discourage legit users. theresa knott 10:30, 11 May 2004 (UTC)


 * Well, my thought (for the newbie) was that there would be adequate warning on the first and second edits so that by the third they'd know about the consequences. That is, the edit page would note that there was a limit on the number of edits and encouraging previewing.  Then at least it wouldn't be a rude shock.  As for the new pages issue, I really don't have a good sense as to whether that would be a problem. -- VV 10:37, 11 May 2004 (UTC)


 * (William M. Connolley 12:37, 2004 May 11 (UTC)) I rather like VV's idea. It would take someone to do the coding, but has the great advantage that after that has been done its all automatic. To cope with new pages, it would be easy enough for the feature not to kick in until its had 10 (or 20, or whatever) edits.
 * The thing I'm slightly uncomfortable about is the whole idea of restricting editing for everyone becuase of the behaviour of a few. However I'm not that much against it I suppose. It would have to be for articles only, not talk pages, user pages and the like. It would be nice to have it apply only to certain users though, so the AC can appply it as a punishment, and the rest of us can get on with editting unhindered. theresa knott 15:22, 11 May 2004 (UTC)


 * I think this is a bad idea. There are many situations where one might want to make several edits to a page.  I myself am generally pretty lax about using preview, so I often have to go back and modify typos, but beyond that, when one is contributing something long, one wants to save a couple of times to make sure it doesn't get lost.  And all kinds of other reasonable things.  I think this is a bad idea.  I'ld also say that I think Isomorphic is right - three reverts should trigger protection, and then the people should have to work it out on the talk page.  Perhaps some kind of automatic mediation is in order?  I just think that all the finger pointing is essentially worthless, and that's what using quickpolls to "punish" people who violate the three revert rule ultimately leads to. john 16:06, 11 May 2004 (UTC)


 * I'm seeing other flaws with the idea; e.g., sockpuppets would be richly rewarded. Also, perhaps this could be imposed on a per-page basis?  Well, anyway one purpose was to have a fair number of "free edits" before this kicked in (which the earlier proposal did not).  I thought three would be enough for legitimate editors, but we could make it five or something.  (Case in point, I seem to have accidentally hit "Save page" before finishing this comment.)  At any rate, I'm becoming less excited about this idea now.  -- VV 22:56, 11 May 2004 (UTC)


 * By the way, if protection and discussion doesn't work as a way to end edit wars, can we really say that 24 hour bans work? They're even less effective. john 16:16, 11 May 2004 (UTC)


 * (William M. Connolley 21:16, 2004 May 11 (UTC)) 24 h bans are surprisingly effective. They do work. Given how small a penalty this really is (tiny in proportion to the agonising that goes into them) it is odd how effective they are. I presume its because people don't want to get "form". Page protection, by contrast, is a pain for all legitimate users, and inevitably locks in one side or another.


 * I agree with John that while a page protection may not guarantee a resolution, a 24-hour ban never addresses the underlying dispute. Concerning the automatic mediation suggestion, I suggest that if you choose not to participate in the talk discussion on a protected article, then you forfeit your say in the consensus version.  Whatever is decided in talk is the consensus, and your edits after protection is lifted may be treated as vandalism if they go against this.  That way, both sides are forced to participate or have their opinions ignored outright.  Of course "participate" is subjective, but this policy should prevent one side from obviously stonewalling or ignoring the discussion entirely. Isomorphic 17:30, 11 May 2004 (UTC)

Norwich NR4 7HU

"Then keep it protected until consensus is hacked out on the talk page" This won't work in many situations, because there are a number of people who would rather have a page stay protected indefinitely than work out consensus on the talk page. This is especially true when one or more of the reverters don't think the page should exist at all, such as Wik with Atlantium. anthony (see warning)

Isn't the idea that whatever is worked out on the talk page counts as consensus, even if one "side" has refused to discuss it? Although I agree this can be problematic. Mr. Connolly, can you give examples of effective 24 hour bans? john 04:52, 12 May 2004 (UTC)

I don't know about the rest of you, but my preference is to have a community-endorsed and -enforced penalty for edit warring. The 24 block is a minor punishment whose purpose is to symbolize the community's disapproval (and also give us some respite). -- Cyan 21:34, 12 May 2004 (UTC)

But what good is that when it just leads to bitterness? And why have the quickpoll at all? It is easy to determine violations of the three revert rule. Why not just let any admin ban anybody who violates the rule? The process of quickpolls just leads to escalating the conflict. john 22:04, 12 May 2004 (UTC)

Your question presumes that a 24-hour block will lead to bitterness. I don't think it will in all cases: if all parties who violate the 3-revert rule in an edit war are blocked, no one of them can complain of unfairness. If I think I'm right, but 20 people think I'm being a dick and agree that I need a time-out, maybe, if I have some capacity for self-reflection, I'll change my behavior patterns.

My object is to set a standard that edit wars just aren't done, and to enforce it with community condemnation of edit wars. Letting sysops pick and choose defeats this purpose.

You wrote, "Isn't the idea that whatever is worked out on the talk page counts as consensus, even if one "side" has refused to discuss it?" No, consensus means that all interested parties come to some kind of agreement. Operationally, that means that no one feels it is necessary to revert. If, for whatever reason, I refuse to discuss the matter with others, that doesn't stop me from being an "interested party" in practical terms &mdash; I can still revert, forcing the page to remain protected for long, long periods of time, to the detriment of the wiki editing process on that article. -- Cyan 04:21, 13 May 2004 (UTC)
 * You contradict yourself. If all parties are required to be banned, then that's not "20 people think I'm a dick." It implies nothing at all, because what you got was knee-jerk required response, not a reasoned judgement that you were acting inappropriately.  There's no more moral weight there.  Isomorphic 05:20, 13 May 2004 (UTC)

What leads to bitterness is not so much the being banned for twenty-four hours. What causes bitterness is the process of having a quickpoll, having twenty people say they think you're a jerk (some of them frequently voting to ban you because they don't like you), and so on and so forth. If people were automatically banned, with no quickpolls, there would not be the same level of bitterness. It might not have any moral weight, but it would certainly make people think twice about breaking the three revert rule, wouldn't it? If there's going to be a three revert rule, it would make sense to just ban them outright, rather than forcing through a quickpoll. john 05:38, 13 May 2004 (UTC)
 * I was only pointing out that if there's no room for reasoned judgement then it isn't going to cause any self-reflection or resolve any disputes. I further agree that quickpolls cause unnecessary animosity.  That's why I favor protection.  It's the least messy.  I've been through a page protection/temp version experience at terrorism, and found it a generally useful process. Isomorphic 05:50, 13 May 2004 (UTC)

I agree with you. But, I mean, if you're going to ban people for violating the three revert rule, you should just ban people for violating the three revert rule. All of the pro-people have been all "oh, this would work, if only people weren't biased." But having a quickpoll is an invitation to bias, and to hurt feelings, and to unfair results, and so on and so forth. And violations of the three revert rule are a largely objective standard - certainly outright violations, but also the kind of clever, almost reverts, can be detected without much difficulty. So if you're going to do it, why have the quickpoll at all. It's just an intermediary that makes things less efficient. john 07:22, 13 May 2004 (UTC)

Inactivity Notice
The Quickpoll process isn't being used, and hasn't been in quite a while. There is no strong consensus here that Quickpolls are useful. Because of this and of the long inactivity, it's not clear to me what would happen if a new Quickpoll were started now, or whether its result would still be considered binding. I have put a note on the page in an attempt to reflect this, so that newer users who come across it will understand the context. Isomorphic 15:21, 3 Jun 2004 (UTC)


 * Along as the wikipolice are silent on this issue. The first editor to create a quickpoll would receive the answer.  I see no reason why you cannot create a quickpoll since is has been over 30 days. -- John Gohde 16:10, 3 Jun 2004 (UTC)
 * Been over 30 days since what? I think you must be confused about something.  Quickpolls were tried for a period of 30 days.  At the end of that 30 days, we started the review above to determine whether we would continue using Quickpolls at all, and if so whether any changes to policy were required.  The reactions in the 30 day review were very much mixed, so there wasn't a clear conclusion reached.  People have stopped using the page, so it seems that there is a de facto consensus not to use quickpolls anymore.  You could still start one, but it's not clear that they still carry any weight given the number of people above who said we should stop using them. Isomorphic 16:44, 3 Jun 2004 (UTC)


 * I think a few people may have been confused by the fact that some participants, perhaps including myself, occasionally referred to a 30-day review, and therefore concluded that the review period would last for 30 days. I agree with Isomorphic's conclusion that the quickpoll procedure lacks consensus support for continued use, and therefore should be considered inactive. --Michael Snow 22:52, 4 Jun 2004 (UTC)


 * If quickpolls really are supposed to have been discontinued, it might be helpful to redirect the page to WP:RFC to prevent people not realising this and trying to use it. Angela. 02:06, 17 Jul 2004 (UTC)
 * Quickpolls as a formal process are clearly dead at this point. Still, I don't see much harm in the sort of occasional and informal use we've seen since. I don't personally frequent RfC, so I don't know how similar/dissimilar they are. Isomorphic 02:24, 17 Jul 2004 (UTC)

Other uses for quickpolls
I'm wondering what people think of using quickpolls for content disputes as an alternative (or compliment) to RfC's. IE a quickpoll on which version of an article is preferred, on which side is right in a dispute etc? AndyL 14:38, 18 Jul 2004 (UTC)
 * I'm a bet leery of settling edit wars with quickpolls. I think this could be abused, one side calling for a quick poll at a time when the other side is unlikely to be online, for example. I think the protect and discuss method we have now will be more productive than a quickpoll to enforce a particular version of an article. Gentgeen 15:04, 18 Jul 2004 (UTC)

Perhaps we should have a "modified" quickpoll that must be up for a minimum of 72 hours and must be referenced on the articles Talk page? AndyL 15:15, 18 Jul 2004 (UTC)
 * The problem with using "quick"polls to solve content disputes is that content disputes can't (and shouldn't) be solved hastilly. They need discussion and consideration, not just a random poll of whoever wanders through.  Besides, I can easilly foresee constant arguments here if we did that. Isomorphic 20:43, 18 Jul 2004 (UTC)