Wikipedia talk:Administrator recall/Archive 2

Proposal notes

Intent

My intent in proposing this policy is simple--to introduce an extremely simple (I tried to make the process as "stupidly basic" as possible) and hard to abuse system of peer review and oversight of people with administrative access. As it presently is, barring extraordinary and public abuse of their priviledges as an admin, it is extremely unlikely for anyone to be de-sysoped--for all intents and purposes, the simple act of passing an RfA is a lifetime appointment with little practical oversight and/or possible repercussions for misbehavior under current systems.

Notice how the certification process works: it's basically just RfC certification, but with an extra requirement on who can certify. The RfA recall vote is literally a forced RfA, where a support vote is an endorsement of desysoping. I have no ill will towards any admin, and this process and tool is for administrators as well as editors. Think about it: you need 1 filer, and 25 certifiers to justify that the community needs to reassess an administrator's standing, and the filer also has to state his reasons, justification, and evidence (to which the admin can reply). Then it goes to general community review a second time, and early closingare specifically forbidden and disallowed from this proposed process--to at all times give the admin being recalled the benefit of the doubt.

Reapplication as an admin

Any previous admin who was removed by the community is free to reapply as an administrator again at their own discretion, per normal application for adminship guidelines and the will of the community.

Sample Timeline, explanation

Just to clarify/expand on how it would work.

• Day 1: A user determines he should file an Admin Recall. He posts the above notice of his complaint with full basis/evidence sections on the admin's personal Talk page, to give the admin 24 hours' notice.
• Day 2: The user files the Admin Recall for certification.
• Days 2-9: Certification window
• Day 9: certification closes. If not met, Recall request closes. If met, within 24 hours the recall vote must be initiated.
• Days 9-30: Recall vote.

The following is a closed discussion of the proposal. Please do not modify it. Subsequent comments should be made in a new section on the talk page. No further edits should be made to this section.

The result of the proposal was rejected as formulated. -- nae'blis 18:33, 17 August 2006 (UTC)

Straw poll

In general, details of sufferage etc. still to be worked out. Please indicate if you are an admin or not.

Users who support adoption of this policy/procedure

1. rootology (T) 14:34, 14 August 2006 (UTC) Non-admin.
2. Herostratus 14:38, 14 August 2006 (UTC) Admin.
3. badlydrawnjeff talk 14:45, 14 August 2006 (UTC), non-admin.
4. --Robdurbar 17:58, 14 August 2006 (UTC) I don't see why a policy such as this cannot work alongside existing processes such as the admins for recall; its just another method of accountability. Admin
5. Tyrenius 19:46, 14 August 2006 (UTC) The community is trusted with the responsibility of appointing admins, so they should be equally entitled to recall. I see no reason not to trust the community with acting responsibly over it, and this proposal is designed to avoid abuse of such a process. Admin

   After the discussions, particularly regarding the existing role of ArbCom, I don't feel this proposal is the answer to possible dissatisfactions with admins. I'm leaving my name here as moral support as I do feel something needs to be instituted, possibly a halfway house to ArbCom to deal with less serious grievances. Tyrenius 17:55, 17 August 2006 (UTC)

6. BigDT 21:04, 14 August 2006 (UTC) - 31 people with 2000 edits hardly qualifies as a lynch mob. I support this idea, but honestly, I think it errs on the side of being TOO careful. If you find an 80% majority wanting to desysop someone, then there is a problem somewhere along the line. I'd make it 20 to indict and 66% to convict.

   I agree. The question is what percentage of users have to have faith in an admin to make their work viable. If 66% of the community is that unhappy with someone, their continued adminship surely has to be counter-productive. ArbCom is a majority vote, so one might even consider that precedent — if more than 50% of editors no longer wish someone to have those powers, it indicates there has not been an appropriate use of them. Another way of looking at that would be that it needs 75%+ support to be sysopped, but only 51% support to confirm continuation. Tyrenius 21:43, 14 August 2006 (UTC)

7. Karwynn (talk) 15:03, 14 August 2006 (UTC) Accountability is never a bad thing. We can work out the details, but eh baseline concept is gold. non-admin

   It might not be a bad concept, but this application of it is terrible, and you're currently supporting the application, not the concept itself. --Rory096 21:45, 14 August 2006 (UTC)

   Obviously I don't think so, or I wouldn't be supporting it :-) Karwynn (talk) 21:47, 14 August 2006 (UTC)

8. HResearcher Filed requests must be based on documented abuse (differences) --HResearcher 09:30, 15 August 2006 (UTC)
9. Johntex\^talk 20:30, 15 August 2006 (UTC) Moral support for imperfect, but needed, proposal. I don't yet like all the specifics of this proposal, but I am more and more convinced that some form of community recall is needed. It may be that we should adopt an imperfect proposal and improve it over time through usage. Johntex\^talk 20:30, 15 August 2006 (UTC)

Users opposed to adoption of this policy/procedure

1. Fan-1967 14:42, 14 August 2006 (UTC) Non-admin.
2. kingboyk 14:44, 14 August 2006 (UTC) Admin.
3. Too complex and number-driven. -- nae'blis 14:52, 14 August 2006 (UTC) (non-admin)
4. Fails the KISS test. Aren't I Obscure? 17:23, 14 August 2006 (UTC)
5. I much prefer desysopping by ArbCom over desysopping by lynch mob. --Cyde Weys 17:39, 14 August 2006 (UTC)

   Hrrmmm ... do I even need to indicate that I'm an admin? Doesn't everyone already know that, for good or for bad? --Cyde Weys 17:40, 14 August 2006 (UTC)

   The usual term is community consensus, rather than lynch mob.Tyrenius 19:51, 14 August 2006 (UTC)

6. I support accountability, but prefer the much softer more flexible, and community centric Category:Administrators open to recall which is grass roots driven, doesn't have the weight of policy, and has as an option, a discussion of what the issues are after far fewer people speak up. This proposal is too complex and number driven and is too close to ArbCom in heaviness, as ewll as feelig like a lynch mob. what is missing is the idea that reasonable people can ask for discussion and that doesn't necessarily mean a repeat of RfA or a forwarding to ArbCom. ++Lar: t/c 17:47, 14 August 2006 (UTC) (Admin.)
7. I prefer desysopping by ArbCom. --Siva1979^{Talk to me} 18:24, 14 August 2006 (UTC) (Non-admin)
8. opposed. Cyde is right. pschemp | talk 20:05, 14 August 2006 (UTC) (Admin.)
9. I trust the ArbCom more than RfA; generally if an admin is doing something right, some people will be pissed off. Anyway, the requirements to file and certify a recall are too stringent, which could encourage cabals and people thinking that some people's opinions mean more than others'. By the way, I refuse to specify whether I am an admin or not, as that too makes it seem that some people's opinions are worth more. Admin opinions mean no more than any other editor's; they just have some extra buttons in case policy is being broken. --Rory096 21:27, 14 August 2006 (UTC)

   THe admin/non-admin thing is just for statistical purposes, this isn't even a real vote. Karwynn (talk) 21:35, 14 August 2006 (UTC)

   Even so, it's not relevant to any discussion, so no statistics of it should be kept. It's like keeping statistics of the genders or races of the "voters," it's just not necessary. --Rory096 01:50, 15 August 2006 (UTC)

   Of course it's useful. Admins have a potential conflict of interest here, wouldn't you suppose. Herostratus 03:10, 15 August 2006 (UTC)

   I would hope that our admins would act in the best interests of the encyclopaedia, not themselves. --Rory096 03:03, 17 August 2006 (UTC)

10. With all these precautions to prevent a desysopping from being a 'lynch mob' as Cyde put it, only make arbcom look simpler to desysop an admin. I think we already have a system in place that can be used for such matters, and this one is unnecessary. Cowman109^Talk 22:29, 14 August 2006 (UTC) (Admin.)
11. Per below. --Sam Blanning^(talk) 22:33, 14 August 2006 (UTC) (Admin.)
12. Per Siva1979. Non-admin Mike Christie 23:38, 14 August 2006 (UTC)
13. This reminds me of the song "hole in the bucket". --SB_Johnny | ^talk 23:46, 14 August 2006 (UTC) (Non-admin, I guess.)
14. Firstly: what exactly is a user in good standing? (I don't know if I'm in good standing or not). Secondly, what is preventing a more than capable, but unpopular, admin being recalled for something spurious; like not making enough portal talk edits, or contributing a featured article etc. Thirdly: I think there might be a struggle to get enough qualified people to support a request apart from in the most serious cases (the criteria are quite high, and there is evidence to suggest that vote pages can have a low readership). Fourth, why should a recall RfA last for three weeks? (most RfAs tend to spike in the first couple of days, take a look). Fifth, I won't support a policy with consensus misspelt throughout. I am not necessarily against the idea of a recall policy, but I cannot support until it is both better though out and at least the equall of an ArbCom ruling in terms of fairness. Rje 01:19, 15 August 2006 (UTC) (Admin.)
15. Most definitely. Horrible proposal full of bureaucratic impossibilities and troll implications. User:Zoe|(talk) 02:42, 15 August 2006 (UTC) (Admin.)
16. Oppose, too much process, with 30+ long term users and admins, might as well just bring it to arbcom. xaosflux (talk · contribs · blocks · protections · deletions · page moves · rights · RfA) — xaosflux ^Talk 04:15, 15 August 2006 (UTC) ^admin
17. Opppose - too much process for so little result -- Tawker 14:55, 16 August 2006 (UTC)`
18. Oppose - If 30+ active editors are hacked off at someone they will definitely have a case for ArbCom. --StuffOfInterest 17:56, 16 August 2006 (UTC)
19. Oppose. More numbers-driven bureaucray, more chances for wikilawyering, another pseudo-legalistic cudgel for the tireless and tiresome to wield with assumed authority. First-class troll-bait. No. --Calton | Talk 02:06, 17 August 2006 (UTC)
20. Oppose: Edit counts ARE NOT the end all-be all for good standing. Time served, fine. Edit counts, no. --Signed and Sealed, JJJJust (T C) 09:11, 17 August 2006 (UTC)

Users neutral to adoption of this policy/procedure

1. TheronJ 20:34, 14 August 2006 (UTC) I don't have a strong opinion on the underlying policy, but I do note that it will be almost impossible to get 30 votes on anything without vote spamming/advertising, and that seems to be on its way out.

   Yeah, the standards are too high now, thanks to "suggestions" from overly suspicious admins. Honestly, with their mindset, RfAs, AfDs and RfCs would all be scrapped as completely worthless. THe real issue is an aversion to accountability. Karwynn (talk) 20:39, 14 August 2006 (UTC)

   Yes, that's absolutely how we feel. We're getting close to scrapping the edit function on the encyclopaedia because too many people use it to complain about us. In fact I favour deleting the entire encyclopaedia and replacing Main_Page with a web browser version of Grue's Rouge Admin - The Game. --Sam Blanning^(talk) 23:18, 14 August 2006 (UTC)

2. I really want to support this, but the requirements are too stringent. I'd rather see >1000 edits over >1 year, and >20 voters in agreement. My current problems [1] with an admin and the processes to get a decent resolution made me look for this, by the way. --Kickstart70-T-C 23:32, 14 August 2006 (UTC)
3. I would like to see a community-based admin recall process (it seems to work in the German Wikipedia, which has no ArbCom). As the process here does not present a way in which anybody will ever be desysopped (ArbCom will be faster and easier in all cases I could imagine), the proposal is useless. Kusma (討論) 09:36, 15 August 2006 (UTC)

   Oh, and I am an admin. Kusma (討論) 09:37, 15 August 2006 (UTC)

Voting is evil

• Tony Sidaway 19:50, 14 August 2006 (UTC) This isn't going to become policy any time soon, but I thought I'd just point out the obvious.

  Funny, RfA is a vote. rootology (T) 19:52, 14 August 2006 (UTC)

  No, it's not. It may indicate why you're approaching this from such a numbers-based focus, however. -- nae'blis 20:21, 14 August 2006 (UTC)

  Sheesh, it's just a straw poll, you see them all the time, lighten up. Herostratus 20:31, 14 August 2006 (UTC)

  A typical RfA is formatted as "(24/4/4)", right? And based on an approximate consensus of what? 65%? 75%? The person becomes admin. We can call a spade a spade... and it's just a straw pole as Hero said to gauge support. rootology (T) 20:34, 14 August 2006 (UTC)

  Don't assume that I approve of what RFA has become. --Tony Sidaway 20:39, 14 August 2006 (UTC)

  Out of curiosity, what has it become? Your own RfA was roughly those percentages and yours looks fairly standard compared to today's. rootology (T) 20:43, 14 August 2006 (UTC)

  I do not approve of a voting process for admin bits at all. I would rather they were appointed after discussion of the application, at the discretion of a bureaucrat. This would, I hope, burn off the RFA groupies, Esperanza members and me-too voters. --Tony Sidaway 14:15, 15 August 2006 (UTC)

  Well, if something like that happened than in fairness every existing adminship would have to be reviewed in such a fashion. rootology (T) 14:59, 15 August 2006 (UTC)

  ArbCom, as the existing method of de-sysopping, is a vote. This just extends the franchise, but keeps the same principle. Tyrenius 20:56, 14 August 2006 (UTC)

  That is a severe misrepresentation of the arbitration process. It's a deliberative process. --Tony Sidaway 14:17, 15 August 2006 (UTC)

  It is a deliberative process, which ends with a vote on each deliberation. It would be good to try to incorporate some of the deliberation you have highlighted into the current proposal, using the model of ArbCom. Tyrenius 14:29, 15 August 2006 (UTC)

  Why go to the effort of duplication? We already have an arbitration committee and it's quite capable of desysopping a bad administrator in a much shorter period than that envisioned in this proposal. --Tony Sidaway 14:37, 15 August 2006 (UTC)

  Because if 31 users brought a laundry list of things to ArbCom that weren't flaming, violent abuses of admin buttons, and 200+ people said, "You know, this guy really isn't a good guy, and shouldn't be trusted..." ArbCom would clearly remove them? The point isn't duplication and that is a severe misrepresentation of the proposed Recall process. The point is empower the community and the admins' peers to oversee their own end-to-end. rootology (T) 15:02, 15 August 2006 (UTC)

• Holy crap, Tony and I agree on something. -- nae'blis 19:52, 14 August 2006 (UTC)
• IMO straw polls should only start once the discussion goes stale. Note that Template:Cent displays 'Discussion' and 'Straw polls' as separate stages, one after the other. I'm not sure whether I should move the entry there or not. --Sam Blanning^(talk) 22:34, 14 August 2006 (UTC)
• Agreeing against voting. I think debating is better than voting. Anomo 00:13, 15 August 2006 (UTC)

The above discussion is preserved as an archive of the proposal. Please do not modify it. Subsequent comments should be made in a new section on this talk page. No further edits should be made to this section.

Discussion

What do you think? rootology (T) 22:46, 10 August 2006 (UTC)

How is this different or better than existing de-adminship procedures? Deco 22:54, 10 August 2006 (UTC)

We have an existing de-adminship procedure? (Other than ArbCom and Jimbo, that is?) Kirill Lokshin 22:57, 10 August 2006 (UTC)

This is different, a simplified, transparent, and centralized approach. There is Wikipedia:Requests for de-adminship, but that requires going through ArbCom and Dispute Resolution, and is at best ill-defined. This is a simple community oversight approach that is intended to be crystal clear, and very hard to abuse. rootology (T) 23:01, 10 August 2006 (UTC)

(This is probably not the kind of reply you're soliciting, but I think the page should be moved to Wikipedia:Admin recall (i.e., without CamelCase).) — mark ✎ 22:56, 10 August 2006 (UTC)

All set, good idea, and I added WP:RECALL as it's shortcut. rootology (T) 23:27, 10 August 2006 (UTC)

• I think their should be some discussion first before asking people to draw lines. If I thought there was a 90% chance that consensus would immediately be for or against this, then I might support taking a poll first to find that out right away. I don't think that is likely in this case, so I urge the "polling" section be removed for now, pending some discussion first. Johntex\^talk 22:57, 10 August 2006 (UTC)

  Good idea, I removed the poll. rootology (T) 23:01, 10 August 2006 (UTC)

It's a troll's charter. --Tony Sidaway 22:59, 10 August 2006 (UTC)

It's a proposed process for good faith oversight that has a many steps built into it as designed to prevent and limit abuse. rootology (T) 23:02, 10 August 2006 (UTC)

There are procedures by which an admin can be removed from admin status. A policy along these lines could be manipulated too easily by malicious intent and external factors. Bastique▼^parler _voir 23:04, 10 August 2006 (UTC)

Which parts of the first draft of the proposal from reading it do you think are most abusable? rootology (T) 23:09, 10 August 2006 (UTC)

I don't particularly feel the need to say anything at this time other than "This is going to be fun". --Cyde Weys 23:14, 10 August 2006 (UTC)

• This proposal has existing for less than an hour and I'll want to take more time than that to review a serious, good faith effort at fixing an important problem.
  

So, I can't comment on the specifics of the proposal yet, but I can say that I applaud doing *something* to keep administrators more accountable to the project. Currently, admins are effectively appointed for life, despite whatever conduct they may exhibit after being given the mop.

Just in the past 48 hours I've seen two different, prominent admins scoff at the idea than an admin would ever be desysopped. That is unfortunate.

My hope is that people will come together to seriously discuss this proposal and how it can be improved, rather than being immediately dismissive of this attempt. Johntex\^talk 23:16, 10 August 2006 (UTC)

It should be noted (not casting anything, just saying) that the dismissive ones have been admins. rootology (T) 23:17, 10 August 2006 (UTC)

Really? Well, I'm also an admin, and I think it should be given careful consideration. Hopefully some more admins (as well as non-admins) will come along and help discuss the merits of this. Johntex\^talk 23:29, 10 August 2006 (UTC)

I think we should have a community based admin recall process (and I am an admin). I am not sure that this one is the right thing, as it is extremely protective of admins (while it should be possible to start a recall in a typical RfC case with this procedure, the actual desysopping will only happen in extremely rare cases if it needs a 2/3 majority). Ex-admins have often had more than 50% support in re-RfAs, and none of those would have been desysopped by this procedure here. However, anybody who gets recalled by a procedure with a pro-admin bias such as this definitely does not have the community's trust, and should really not have admin tools. Kusma (討論) 09:19, 11 August 2006 (UTC)

I think if this was just a place where people could report lots of abuse by an admin, it would be better. No voting stuff, just evidence. And it does not get people de-admined until the evidence piles up enough. Just make sure evidence can't be erased and it will be better than this. Anomo 09:33, 11 August 2006 (UTC)

The problem with this is that it's rather extreme ... it goes directly to the worst punishment available, desysopping. In most situations I think placing some restrictions on the administrator would be more appropriate. --Cyde Weys 17:48, 14 August 2006 (UTC)

Something like First Recall: lose admin rights for 1-2 months, Second Recall: desysopped, and reduce the 9 month period somewhat? I.e., the first community action is a "You're on temp probation, don't do that again," as a message, and the second would be to remove them as admin (with the right to reapply as desired)? rootology (T) 17:51, 14 August 2006 (UTC)

Why it doesn't work ...

Bureaucrats don't have the ability to remove sysop or other faculties from users. --Cyde Weys 23:39, 10 August 2006 (UTC)

Who does? rootology (T) 23:44, 10 August 2006 (UTC)

Please see Wikipedia:Bureaucrats and [2]. Oh, and Wikipedia:User access levels. Johntex\^talk 23:48, 10 August 2006 (UTC)

Thanks, I'll clarify the wording more. rootology (T) 23:55, 10 August 2006 (UTC)

I changed the exclusions to read as

Jimbo Wales, Danny Wool, Brad Patrick, Stewards, and any active member of the Board of Directors, are all excluded from this policy. Arbitrators or any current or future "higher than administrator" users (arbitrators, beaurocrats currently) may be called in this fashion for administrative recall, but their removal should it reach completion must be approved by the Board of Directors, Jimbo, or an official WP:OFFICE action as determined by and seen fit by them. Recall of other administrators will be as per this policy. Normal "admins" for this policy shall encompass anyone from "basic" admin level up to and including those who are regular admins with additional priviledges and duties, such as Check User and Oversight."

Much better. rootology (T) 00:00, 11 August 2006 (UTC)

Wording changed to "In the event of a successful recall, the admin in question after the conclusion of the second seven day time frame is to be immediately de-sysoped, and removed of any "admin type" or "admin level" additional access priviledges per the community's concensus will and decision." And that's a pretty small semantic nit to pick. :) rootology (T) 23:46, 10 August 2006 (UTC)

Anomo questions

Some questions. First: What is to stop the nominated admin or one of their buddies from deleting one of these, banning everyone who supported it, and erasing their votes? Second question: I thought on outside of wikipedia, like in wiki farms and such (e.g. wikia, editthis.info, elwiki, etc.) that those with bureaucrat access could remove sysop access from others. Is it just a wikipedia thing that bureaucrats can't remove sysop access? Anomo 00:07, 11 August 2006 (UTC)

What is to stop the nominated admin or one of their buddies from deleting one of these, banning everyone who supported it, and erasing their votes?

There are over 1,000 admins (or thereabout). Abusive actions like that could be posted at WP:AN/I, and it would obviously be reversed. The beauty of the system is that any improper action can be undone. I've added the "Prevention of potential admin abuse in process" section in response to this. Basically, anything blatantly done as an admin action in regards to a recall filing/vote would be a bannable offense.

I thought on outside of wikipedia, like in wiki farms and such (e.g. wikia, editthis.info, elwiki, etc.) that those with bureaucrat access could remove sysop access from others. Is it just a wikipedia thing that bureaucrats can't remove sysop access?

Not sure, unfortunately. But there are people who can remove access to any user, and I adjusted the previous version to simply say "they'll be desysopped" rather than by whom. rootology (T) 00:19, 11 August 2006 (UTC)

For, "Prevention of potential admin abuse in process," you left out the common, "Oh this person hasn't edited enough so I am banning them. Okay now that they are a banned user and what banned users say doesn't matter. And if someone complains about my actions to WP:AN/I, I will call them a sockpuppet of the banned user and ban them and erase their comments, too." Anomo 08:09, 11 August 2006 (UTC)

Take a look at what I added--that should discourage anyone from doing anything stupid. Any frivilous blocking or harassment by "friends" would go through normal channels, and likely reverted as well by other admins. The purpose and intent of this isn't editors vs. admins--it's to create as transparent as possible method of fair community oversight up and down Wikipedia. rootology (T) 08:34, 11 August 2006 (UTC)

WP:CREEP

A solution in search of a problem. User:Zoe|(talk) 01:57, 11 August 2006 (UTC)

Please describe to me what oversight the community has with admins at this time that would not be served positively by having a transparent and fair process for removal of an admin whom the community collectively judges as unfit to serve? rootology (T) 02:03, 11 August 2006 (UTC)

WP:RFAr. User:Zoe|(talk) 01:51, 12 August 2006 (UTC)

Which is fine, but subject to acceptance by a small group of individuals. This would grant control of the process to the community collectivly. rootology (T) 01:57, 12 August 2006 (UTC)

A small group of people who have been put in their positions because of the support of the community. Unlike this, which is a small group of self-appointed vigilantes. User:Zoe|(talk) 02:45, 15 August 2006 (UTC)

So as an admin you consider the will of the collective community vigilantism, rather than oversight? rootology (T) 03:19, 15 August 2006 (UTC)

Long term vandals can game the system too easily

Wikipedia has attracted a number of long term vandals that have periodically caused problems and returning to their vandalism after occasional sabbaticals. Many of these vandals are known to create large numbers of sockpuppets and sleeper accounts to aid in their endeavors. There are also semi-organized groups of editors who are more interested in advancing certain POVs than abiding by Wikipedia's policies.

This proposal's basis of performing a recall based strictly on numbers and without any other considerations of Wikipedia policy and practices means that any one of these vandals can remove an admin with the audacity to block blatant vandals by simply creating a stable of puppets, making a series of minor edits with each of them, and then waiting till the puppets have aged before beginning a campaign to have any vandal fighting admin de-sysopped. POV pushers will have similar abilities against admins enforcing Wikipedia policies on content or user behavior. Unless some form of reasonableness check is added to the procedures to filter out complaints about admins doing unpopular but needed actions all this proposal will do is reduce the number of admins willing to fight vandals, POV pushers, trolls, and other parties trying to advance a personal agenda over the good of Wikipedia as a whole. --Allen3 ^talk 03:28, 11 August 2006 (UTC)

As mentioned below, the numbers required for someone to pull this off would be just insane. However, the numbers for the threshold for certification could always be raised a bit. What do you think? The RfA portion keep in mind is no more gameable than an actual RfA is. rootology (T) 06:17, 11 August 2006 (UTC)

• Allen brings a great point, and why the reviewing process should take a longer period of time, so that we can investigate who are making these claims. Also who would review these filings, who would be incharge of the proceedings. Who would hold meetings on these filings. This policy is far from being ratifyed. Seems to me this will take a while to fully implement the right policy for this issue.

--Zonerocks 04:10, 11 August 2006 (UTC)

Who currently actually processes RfAs after completion? rootology (T) 06:17, 11 August 2006 (UTC)

• Well, I think you're giving too much credit to the forces of darkness. I mean, if trolls etc. can do all that, they could just as easily prevent the person from becoming an admin in the first place. Presumably the usual investigation and removal can be made if puppetry is suspected, just as with other RfA and other discussions. Besides, if someone wants to make 500 edits from (say) ten accounts, that's 5,000 edits. Is anyone really willing to make 5,000 edits just to troll? And if they do, hey, that's 5,000 good edits; that's worth an admin. Admin status is supposed to be no big deal, getting desysoped should be no big deal, just a chance to concentrate on article work for awhile. Any good organization weeds out the bottom 5% or so periodically. It's ludicrous that admins have, effectively, lifetime tenure from day one. And if the odd admin does get unfairly quashed, that'd be too bad, but no system is perfect. I guess in that case we could dust off that old saying, "That's the way it is. Get used to it." Herostratus 04:09, 11 August 2006 (UTC)

• It would be quite hard, though. Initiating a recall would require 25 sockpuppets and 12,500 edits. The subsequent recall would easily gain two hundred votes in favor of keeping the guy adminned -- attempts to stuff the ballot box are the best way I know of to draw participation -- so a vandal would need another 600 sockpuppets. And finally, the vandal would need to convince a steward to ignore clear bad-faith voting. --Carnildo 04:07, 11 August 2006 (UTC)

  The filing editor has to meet the 5 month/500 edit requirement as well, for 26 users in "good standing" with 13,000+ edits between them for certification to occur. rootology (T) 07:54, 11 August 2006 (UTC)

These reasons are exactly why I put in the given numbers (initially 300 edits, then 500 on second thought)--no one is going to spend that much time doing quality edits, and if some moron actually goes to the trouble of using like a bot or something to build up the accounts, seconds of inspection would show them as fakes during certification. Any editor that gets to 500 edits and 5 months is probably going to have something on their user and talk pages, some interaction, and be somewhat known. The 1 filer/25 certifiers by this method makes it so that you have *real* people certifying it. The numbers could also be theoretically raised, but 500 seems a good value to use as it allows people who aren't particularly "hard core" editors to have access to the process. I'd be leery of setting it at say something like "1,000 main space edits". rootology (T) 06:05, 11 August 2006 (UTC)

Any number you come up with can be overcome by a group or individual with determination, or else will be so large that there will be an insufficiently large pool of qualified individuals to make this a viable policy. It is not the bored school kiddies that present a problem but the organized POV pushers and long-term trolls. Accumulating a couple hundred edits takes a little effort, but not as much as you imply. I have had days were I accumulated close to 500 edits doing simple maintenance and there are tools such as Wikipedia:AutoWikiBrowser make it very easy for anyone wishing to make a large number of main space edits to perform close to an edit per minute. Your proposed suffrage requirements would remove the drive-by trouble makers but simply creates a minimum bar for a hardcore vandal, political campaign, business interest, or fringe group to pass before they gain an effective veto over Wikipedia's ability to police itself. --Allen3 ^talk 11:25, 11 August 2006 (UTC)

The problem is that there needs to a method by which the community and admins can police fellow admins without having to go through ArbCom in a completely ill-defined process as it exists now. Like I mentioned below to Amono, frivilous filings will be apparent, as their evidence will have no basis, and the RfA in all likelihood would reflect that. This bit--"organized POV pushers and long-term trolls"--is actually a troubling comment. No editor (admin or regular) is infallible. If an admin did something (action, course/trend of action, etc.) that made someone feel that they were due for recall, and took the time to file and certify, and the community at large agreed with them in the RfA recall vote that the admin should be recalled, wouldn't that be indicative of a good recall filing in the end? The filing itself is no different than an RfC: "I think this is a problem, here's why. Who agrees?" The issue to me is that--on the flipside--is 26 people went to ArbCom to file a complaint all at once about an admin due to perceived admin abuse, or if 26 people actually filed an RfC together against an admin due to perceived admin abuse, it would be a massive deal as so many people would be involved. That's why the threshold and values are there--26 users is a lot of people, but it's still just certification--the community at large in RfA actually makes the decision to de-admin someone, not that group. rootology (T) 16:44, 11 August 2006 (UTC)

As described below by Sam Blanning, the problem with this proposal is not trying to determine the correct number of certifiers but instead showing that some number exists that is a correct number. Tweaking numbers will not fix an intrinsically flawed system. --Allen3 ^talk 03:23, 12 August 2006 (UTC)

Question about Admin rights during Procedure

Here's my thing, Shouldn't the admin be able to appeal or be able to give a reason why he should not be removed as admin? Also, kudos, this sounds like a good plan which if we can al work together to make it even better. --Zonerocks 03:57, 11 August 2006 (UTC)

• Well, the recall RfA-like procedure would give plenty of opportunity for discussion, including from the affected admin, wouldn't it? Herostratus 04:13, 11 August 2006 (UTC)
• I would hope so. I mean this policy should state that. or at least reference to something it says in the RfA policy. --Zonerocks 04:24, 11 August 2006 (UTC)

  The policy does state it, and gives the admin two chances to explain his actions and why he should continue on as admin. During the Certification, his reply is there for all to see, and he's free to participate him/herself in any discussion on the talk page during that week. During the actual recall vote (should it get to that) he also has of course full access to participate. As for appeal, that's inherently built in as well--I tried to form the policy idea in such a way that would discourage a lynch mobbing type of scenario. Simply put;

Certification: admin gets a week to explain his actions if he so chooses, whilst basically a petition runs via the certification to see if 25 tenured users agree or not whether they should be recalled based on the case, their actions, and their history.

Recall vote/RfA: Normal RfA type scenario--ask questions, the whole nine yards. The only difference is it can't be closed early; it will run seven days to give the admin a fair shake and benefit of the doubt. And of course that the votes are inversed, for a "support" vote being a supporting vote for desysoping.

Appeal process: Simply put, a full certification plus a vote to desysop is a vote of no confidence in the given admin from his peers and community at the time of the incident. The appeal is that the admin is free to rerun for RfA anytime--60 seconds after the Recall closes, a month, or a year later--totally up to him. rootology (T) 06:11, 11 August 2006 (UTC)

Nothing to do with the Board

"their removal should it reach completion must be approved by the Board of Directors.."

The Board has, and should have, no role in decisions about adminship. This should be determined by the community or through community-determined procedures, such as arbitration. Wikimedia's Board is irrelevant to this sort of policy. Angela. 04:40, 11 August 2006 (UTC)

That exclusion portion is specifically for arbitrators and beaurocrats only, as to get into those particular positions you have to be (I presume) extremely experienced, and at least generally accepted by the community. The reason a successful recall vote for one of them would get kicked up to the next level would be for review, and to not causes utter chaos--also, as ArbCom decisions can seriously piss losing ArbCom users off I'd imagine, it's to provide a modicum of insulation to the legitimate higher end people that run Wikipedia. They're still subject to the process, but given their position I think it should be a litter more involved. What do you think could should be done for this if anything? rootology (T) 06:15, 11 August 2006 (UTC)

By the way, I just realized who you are--I was a bit slow tonight, just home from a show. Thanks for taking the time you did! rootology (T) 06:41, 11 August 2006 (UTC)

Whoa! Angela Beasley. I am honored. You co-founded wikia, so do you know... I thought on outside of wikipedia, like in wiki farms and such (e.g. wikia, editthis.info, elwiki, etc.) that those with bureaucrat access could remove sysop access from others. Is it just a wikipedia thing that bureaucrats can't remove sysop access? Anomo 06:22, 11 August 2006 (UTC)

The problem here is the perception of there being a "hierarchy" (user, administrator, bureaucrat, arbitrator, steward, board). This isn't how it works. Attempts to deal with managing Wikipedia from a hierarchical standpoint are misguided and will inevitably fail. Kelly Martin (talk) 20:03, 14 August 2006 (UTC)

This policy isn't intended to govern/oversee positions whose oversight is better left to people that are more fit to judge those positions--I wouldn't even consider wanting this policy to be applicable to people on the Steward/Board level; it's purpose to offer community recall/oversight over admins in general. If the board, or Office in however their bylaws were setup decied to oust someone from the board, I would think it would be best to do so on that level. This is strictly a counter-balance/other side to the RfA policy and process--no more, no less. rootology (T) 20:07, 14 August 2006 (UTC)

I agree. THat's one component of admin abuse is the thought that admins have more of a say in content disputes than "normal" editors. Karwynn (talk) 20:05, 14 August 2006 (UTC)

not vandals that can game the system, but cliques

Someone earlier said long term vandals can game it by making 25 accounts and doing this. To me that's laughable. The accounts would be easily found out by checkuser as coming from proxies. Also 25 users x 300 edits = 7500 edits.

The problem I see is that it is 25 people and not 25 instances of admin abuse. Of course an administrator could probably hide the abuse, making it tougher.

What I think would happen is two administrators dislike each other. And there's millions of people on wikipedia. Well one administrator has tons of friends and the other doesn't have many friends on wikipedia. The one with friends gets them all together to de-sysop their enemy.

If there were 25 instances of admin abuse and not 25 people, this would be a good system. Anomo 06:16, 11 August 2006 (UTC)

Theoretically a vindictive person could get together 25 people who meet the criteria, to get them to certify a given Recall. The thing is that this is only the first step--the larger community still has to actually decide to desysop the admin afterwards via an RfA vote. While yes, it's feasible that someone vindictive could somehow muster up 25 people--their doing it on-wikipedia would be pretty blatant from their contributions, and they'd still be hardly assured of actually succeeding in anything via the recall vote. The other thing is, they have to list an actual basis for the recall, plus evidence. Something frivilous if it actually got the RfA vote going would be shouted down by the larger community. I mean, even this infamous Cabal--how big is it, if someone were to take all the potential members? Anyone reading this--think about who you think is part of a given clique--is that clique really big enough and cohesive enough to sway an entire RfA without being completely obvious and called out for it? Keep in mind as well that the RfAs to actually make someone an admin would be a lot different in tone than if Recall RfA.

Worst case: there will be frivilous filings for Recall, the same as there as frivilous DRs, RfCs, RfARs, etc., but the fact that real users have to agree to put forward the basis of the dispute to have any value will keep it in check. In fact if someone *REALLY* wanted an admin stripped of power, the *worst* thing you can do is file frivilously and without cause: if you the vote is successful they're protected from more shenanigans for nine months' time. If you disliked someone a frivilous Recall is the worst thing you can actually do, as you're protecting in all likelihood the person you don't like; I tried to make the system self correcting in this way. rootology (T) 06:50, 11 August 2006 (UTC)

The post I made above [3] relates to this. Anomo 09:34, 11 August 2006 (UTC)

See what I posted here, just don't want to repost all the info again on the same page, but that should address this. rootology (T) 16:37, 11 August 2006 (UTC)

Okay I just want to reassert that votes is a bad idea

Basically I think this should be redone where it purely lists evidence of abuse by an administrator and that is it. No voting. No popularity contests. No admins de-sysoped because of enemies whether vandals, regular editors, political groups, etc. Just like a big complaint pile. The admin only gets de-sysopped if there is sufficient real evidence, not 25 people who hate them. I personally think it would be better to pile up evidence because on Wikipedia:Requests for de-adminship I see admins being de-admined for single instances, which to me that's out of balance and de-adminning should be taken over the course of long term actions. Also it would be important to require that evidence here remain viewable in histories and not deleted.

I also think this would save much time on arbitration and such. If it's a mere voting thing, any admin that does something controversial will get 25 real editors voting to de-sysop them because they're unpopular (such as if they have to remove a Brian Peppers-like article for privacy reasons). But if it's an evidence thing, then if people think an admin did wrong, instead of making a case, they ad it to a list of evidence under an admin's name. If the admin is being abusive, then evidence will pile up. Furthermore, if it's just evidence, then admins also have to be nice to newcomers as well as long-time editors. Anomo 12:29, 11 August 2006 (UTC)

If this were the case, however, it would be easy to de-admin problem admins and admins who have lost the trust of the community. It isn't, and this is a response to that, and one of the better ones that have come along so far. --badlydrawnjeff talk 12:43, 11 August 2006 (UTC)

I think you're missing part of the structure of the Recall policy. 25/26 people cannot get someone desysopped. Period. 26 people are needed to validate, justify, and certify the merit of the complaint, at which point the motion is basically opened to a community census of the admin's performance--another round of WP:RFA for that admin. If the admin is good and has done nothing wrong, he's got nothing to worry about at all, and will breeze through as the basis of the dispute and the evidence will be right there for all to see. Having a "crap drawer" for admins is a bad idea in the way you describe, as it would be just that... a gargantuan record of complaints that eventually will be literally people flinging garbage whenever there is a disagreement. The filing process here and the time limitations will bar that right out, and as mentioned before any frivilous, without merit filings would be Very Stupid if you wanted to desysop an admin--the community is smart, and you'd just be giving the guy you dislike a 9 month vacation from your nonsense (plus tipping off the community to your bias). rootology (T) 16:36, 11 August 2006 (UTC)

Besides, how many fine admins are there that are not popular? THe only popularity problem I see is that it will schew votes in favor of admins possibly. Karwynn (talk) 17:34, 11 August 2006 (UTC)

Exactly. rootology (T) 17:48, 11 August 2006 (UTC)

Different Format for Filing

This is the current format

               Filing user: (name/signature)
               Administrator being recalled: (name/link to their User space)
               Basis of Recall: (description of why the recall is being filed)
               Evidence: (evidence--diffs, citations, etc.)
               Administrator response: (response by the admin)
               Those supporting creation of the recall: (signatures of 25 users--see criteria below)

Am I the only that thinks this is a bad format? First what is the point of having an administrator response? Wouldn't The admin give his response or appeal during the proceedings of each step. We also need to implement an appeal process during the first stage. Again Having an Administrator response in the filing, will delay the filing process. This should be the new format for filing.

               Filing user: (name/signature)
               Administrator being recalled: (name/link to their User space)
               Basis of Recall: (description of why the recall is being filed)
               Evidence: (evidence--diffs, citations, etc.)
               Those supporting creation of the recall: (signatures of 25 users--see criteria below)

That should be the new process.

The response does not hold up the filing in any way. The procedure would be--lets say I'm filing vs. AdminX:

• 12am Jan 1st: I leave AdminX the above notice on their talk page.
• Anytime after 12am Jan 2nd: I can officially file to begin recall--I don't need to wait. The 24 hours is a courtesy to the admin to let them know I will be doing it; diligence is upon them to keep an eye out for it then.
• 7am Jan 2nd: Let's say I file here--the certification process begins, and shall run to 7am Jan 9th.
• The admin can enter their response anytime they want after I have filed--no different than an RfC. If you file RfC for something, you don't need to wait for them or even give them a heads up, beyond a courtesy message that the RfC has been filed. The 24 hour talk page notification for the admin is because Recall is a Very Big Deal, and to give them the oppurtunity to clear their heads and begin to prepare a response. The ability for an admin to have a response field is very important; Recall isn't intended to be a soapbox for me to go and say "Haha! AdminX sux!!" rootology (T) 16:31, 11 August 2006 (UTC)

First of all, We don't even have a proposed full procedure plan. Now im going to propose a full Procedure plan, and wouldn't that give the admin a time to maybe......hmmmmmmm..... Ban your screen name or IP --Zonerocks 22:05, 11 August 2006 (UTC)

Which would be an abuse of the sysem based on the policy and a bannable offense for the admin--based on the policy as I wrote it. What do you think? Whats a better way? rootology (T) 22:07, 11 August 2006 (UTC)

Number of edits

COuld you clarify for me and in the policy too: Is 500 edits to "any" namespace edits 500 total across all spaces (like 300 main, 200 talk) or 500 to any one space? Oh, and please email me if this proposition gets voted on so I can support it. (I'm on break, but heard about this... elsewhere) Karwynn (talk) 15:15, 11 August 2006 (UTC)

It is 500 edits "total". So, if you have 300 main space edits, 100 Talk edits, and 100 Wikipedia space edits, you qualify, or if you have 400 Main, 50 talk, 25 Wikipedia, 25 Image. Any combination plus 5 months' experience. rootology (T) 16:24, 11 August 2006 (UTC)

Possible Abuse

Whats stops an admin from having a filing against them fail on purprose to protect them for the next 9 months? For instance admin A has person b and C file against them leading to a failure by lack of supprot protecting them for 9months after. Am I reading something wrong? What happens if they get into problems 2 months after defeating a recall? --zero faults ^{|sockpuppets|} 17:01, 11 August 2006 (UTC)

Let's say Admin1 did this--gets another admin or editor to file a bogus complaint vs. him for "protection". First, a good admin won't need to do this. By your assumption of this, it would be done by a hypothetical "bad" admin, or one which was extremely concerned about their standing in the community. If Admin1 did this, and he was really perceived as that bad of an admin or unpopular enough to be concerned about something like this because of their tone, actions, whatever you'd like to list, a "protection" filing would be a tremendously bad idea as it would open them to the community review process--there is no guarantee the certification or RfA would "fail" to their benefit. rootology (T) 17:10, 11 August 2006 (UTC)

How about all certifiers at the end of 7 days are counted as support votes. Or the first 25. That makes it a little risky. Additionally, if it's apparent it can be taken to RfC or RfAR for WP:DISRUPT. Karwynn (talk) 17:11, 11 August 2006 (UTC)

I'd really rather the certification process be kept isolated in this way from the vote process. When I wrote this I saw certification as community agreement that yes, this guy is doing something wrong, and it needs to be looked into further--just that. The actual RfA recall vote is the community decision based on the admins actions, evidence, history, actions during the Recall process, etc., of whether they are meritous to continue serving as an admin. rootology (T) 17:15, 11 August 2006 (UTC)

If one of these fails to be certiied, does it reset the 9-month cycle? Karwynn (talk) 17:17, 11 August 2006 (UTC)

The filing does not, a "dead" filing/vote stays archived just as old RfCs/ArbCom/RfAs do. However, a 'new' good faith filing can be brought after nine months' time if a situation warrants it. The nine months is to stop idiots from flooding the system with a zillion bogus Recalls. rootology (T) 17:20, 11 August 2006 (UTC)

Vote confusion

FIrst of all, major applause to this proposal now that I've read it in full. Very, airtight, good benefit of the doubt, and increases accountability.

This confusion of voting "support" or "oppose" - perhaps it would be in order to add that at NO time may ANY user change the vote of another user, even in cases of apparent confusion. For example, if I said -

Support, So-and-so is a fine admin

CLearly I don't want yo desysop this person, but my vote would count as a vote for de-sysopping, and someone might change it because I was confused on the support/oppose thing. So we should say that you shouldn't change the vote for them, you should message them on their talk page and tell them to do it.

The reason I bring this up is not for the above case, but for possible confusing votes. Like,

Support - poor administrator and incivil WP:BITE violator - not sure desysopping is really in order though

That could be hard to judge. I may very well be "unsure" about desysopping, but vote to desysop anyway. SOmeone could say "Oh they were confused" and change it though. The policy should take a step to prevent that. Karwynn (talk) 17:17, 11 August 2006 (UTC)

How about if votes were--instead of support/oppose, labeled as Recall or Remove and Keep. Recall being desysop, Keep being keep as an admin. It can't be more clear than that. rootology (T) 17:24, 11 August 2006 (UTC)

THat's actually a better idea than mine. Just like MfD votes sort of. Ooh, speeking of which you may want to add that they can't be subjected for consideration for speedy deletion, unless you already did. Karwynn (talk) 17:26, 11 August 2006 (UTC)

My wording as basically "...or anything related." Basically, none of these can be deleted per the policy's wording unless Jimbo, Danny, or Brad say so. rootology (T) 17:31, 11 August 2006 (UTC)

SPecify for wheel warring

A thought more than a suggestion: what about saying no admin who is voting to support de-sysopping can place a block on said admin ( the 2-day, 5-day after block)? The only problem is, you'd have to say that if THAT happens, no admin who OPPOSES the desysopping can UNDO that block, but then no admin who SUPPORTS it could undo THAT, etc.

Basically, this policy has some room for wheel-warring, maybe no involved admin should be able to take any administrative action on the "recall subject", and cannot take any action to undo any block placed ON him or BY him on another user. I don't know... confusing. Karwynn (talk) 17:22, 11 August 2006 (UTC)

Since blocks are undoable by anyone, and I'd (hope) everyone had faith in admins in general to do things right, I wasn't even worried about this aspect--I block is just a block, there's no reason another person can't undo one if it's in good faith and per the rules. The time based block for abuse of recall process--I'll add a note to this effect--should not be reduced by anyone. The rules as they are written are pretty explicit: the admin under nomination is not to game the system or use his administrative tools in regards to the filer or certifiers. It takes just a moment for someone else to be brought it, or a ten second post on WP:AN or WP:ANI; doing so is *not* a great burden on the admin under nomination or the admin community. I'll add a modification now. rootology (T) 17:30, 11 August 2006 (UTC)

Wheel warring in this sense is now covered. Say Karwynn the admin is nominated. You mess twice with the process, using two admin actions. 48+120 = a 168 hour block for flagrant abuse. I see this, and unblock you or shorten the block. Another admin undoes my action and reblocks you for the remaining duration and warns me. I do it again. I'm now blocked for 48 hours. I come back, do it again. Now I'm blocked for 120 hours. Basically, the same remedy applies now in the wording to prevent this. Basically, the system should not be gamed by use of admin tools. rootology (T) 17:38, 11 August 2006 (UTC)

If it isn't broken, don't fix it

This is becoming an extremely complex proposal. We already have a way of dealing with abusive administrators--arbitration. The Arbitration Committee can settle such cases expeditiously (in as little as six days where speed is desirable). More urgent cases can be dealt with summarily by steward powers (for instance, Jimbo's occasional summary desysoppings). --Tony Sidaway 18:43, 11 August 2006 (UTC)

Hi, it's actually a very simple proposal; I simply broke down the level of detail for maximum clarity and discussion. Do you feel that the community should not have the ability to undo it's own previous decisions? rootology (T) 18:52, 11 August 2006 (UTC)

The ArbCom route means the issue is settled by only a small group of editors who -- let's face it -- pretty much know each other and constitute a sort of "insider" group, mostly. Human nature being what it is, what guarantee do we have that personal feeling might not intrude. Herostratus 14:04, 13 August 2006 (UTC)

The reason that it's becoming so complicated is that there are an incredible number of "what if" objections. I piss people off all the time, will step down if six people ask me, and have yet to have even _one_ person suggest I do so. The "trolls" must be able to smell fear, because I'm not even slightly worried about them and that is keeping them away. - brenneman ^{L} 14:25, 14 August 2006 (UTC)

Are there any "what ifs" that you feel are not already answered in the policy? Additionally, this isn't just some casual "let's get this guy desysopped" thing, it's not something a lot of admins will have to worry about. If no one has ever adked you to step down, it probably means you're a good administrator, thus it would be unlikely this policy would ever affect you. However, not all admins are as well trusted by the community. All the "what ifs" are designed to root out trolling and gaming, and even though this isn't my proposal, I think I'm right when I say that bringing up any specific "what ifs" that you think the policy doesn't satisfy would be very helpful. Karwynn (talk) 14:29, 14 August 2006 (UTC)

You're not even on Brandt's list of admins so you're clearly a good admin. I don't judge admins, though, I just see what people say. I think this policy is more for the same ten admins everyone complains about on wikipedia review and places like that. Anomo 14:48, 14 August 2006 (UTC)

This policy has nothing to do with any one admin, it has to do with the hole that exists in current policy. Admins and editors, if they find the person they previously RfA'd in is not the same person they supported, and no longer feels he is right to serve them, have no means without this new policy adoption to change that. rootology (T) 14:53, 14 August 2006 (UTC)

Small suggestions

This is a space for simple, small suggestions, just outright yes or nos basically. I only have one now, but that's only because I forgot the others.

I think the notification 24 prior to filing should be required to link to WP:RECALL. Maybe a wholly standardized message even, but that's just a thought, not a suggestion yet. Maybe even a notice on their userpage? Although that could help or hurt their chances, that last one might be going a bit far. Karwynn (talk) 17:39, 11 August 2006 (UTC)

Removes the human factor

Is ArbCom really overloaded with de-sysop requests that we need to create a policy such as this one? My main issue with it is that there's nothing to stop an admin from being stripped of privileges simply because a majority of users don't like the admin. Other than that I think there's a lot of thought in here. Fagstein 18:40, 11 August 2006 (UTC)

ArbCom is not overloaded as such, but the process for ArbCom will (from reviewing previous issues) only remove privs based on massive or extremely flagrant issues. The problem with the system as it exists now is that it's--in all honesty--extremely easy relatively to get admin privs. But, once appointed, the same community that made the decision to give those powers has no ability to reverse that same decision if the community has decided they no longer support the ability of that admin to serve. It's a one-way street at present; this makes it a true function of the community collectively deciding who is fit to serve as an administrator. It's fine that it's basically easy to get admin privs--it should be, as if the community has faith to give someone the buttons, it should give it to them easily. This process makes people jump through some pretty intensive community support hoops to take away those same buttons, twice, in a very streamlined fashion. Right now, if 50-70+ people decide you should be an admin, odds are you will be. This requires a minimum of 26 people to approve the request that you should be reviewed. If a clear concensus majority then decides that you shouldn't be an admin afterwards, based on your evidence and career, that is exceptionally fair. rootology (T) 18:47, 11 August 2006 (UTC)

First stone

• What happens today if an admin makes enough mistakes to get blocked ? Does (s)he stay admin a lifelong ?
• Do the 25 admins proposing a recall have to be chastily pure ? What if someone finds, after a recall, that one of the justice admins was once blocked, or made 499 edits in his own page and one poor stub or spelling correction in a main article ?
• The Rfa process is too poor and this recall process is too rich.
• Here comes a related subject when it comes to cleaning. "Hey, I'm a sci-fi or disco or TV series (or sports) fan, filling plenty of non encyclopedic pages with non notable stuff that ones like me like. I did no damage and finally got my Rfa thanks to all, I'm glad helping the community." I do not want to go on making fun of it, because the question is serious : what do we want to build ? Noxious people must be sacked. But fandom and trivia could be gentily pushed toward their own wiki. This could discourage some help but, and this is important, discourage also plenty of abuse and trolling. How many random pages did you peruse to find (sigh) that the contents are not relevant here and not even interesting to wikify ? What do you think ? -- DLL ^{.. T} 18:51, 11 August 2006 (UTC)

What happens today if an admin makes enough mistakes to get blocked ? Does (s)he stay admin a lifelong ?

At this time, they stay lifetime admins unless they do something utterly severe enough to get ArbCom or higher to take action. The community at large (admins and editors) have no ability to directly undo their previous RfA decision which is the problem--that simple vote is a lifetime spot with no real appeal.

Do the 25 admins proposing a recall have to be chastily pure ? What if someone finds, after a recall, that one of the justice admins was once blocked, or made 499 edits in his own page and one poor stub or spelling correction in a main article ?

No, previous blocking shouldn't matter. Many admins have been blocked for trolling, 3rr, etc., the same as editors. The 500 threshold in arbitrary one to clear out the obvious puppets. Keep in mind--the certification is not removal, it's certification. The removal would be if it happened from the community's decision based on the evidence in the recall RfA.

The Rfa process is too poor and this recall process is too rich.

Actually, this process is really pretty simple. I just broke down all the details for possibilities, to cover any possible concerns up front. Most policy proposals I've seen come through are so stripped down that they are just the germ of an idea. I wanted to bring up an actual policy/process idea that was developed.

You lost me on your last paragraph. Are you saying that people who work on things that aren't academic in nature are trolls? Huh? rootology (T) 19:00, 11 August 2006 (UTC)

Thank you, Rootology, for your proposal, it is necessary. For my last part : I'm not calling for elitist or academic work here. Did not Diderot begin his encyclopedia when (and maybe because) there was already the French Academy ? Non notable and non encyclopedic subjects are not trolling, they are uninteresting to my taste, this is why I ask questions, and , worse, they may bring trolling. They already (and the examples I choose are not the worse) fill up plenty of space and use plenty of our energies in the Afd pages. -- DLL ^{.. T} 19:32, 11 August 2006 (UTC)

Thanks for clearing it up--and that's an excellent point. rootology (T) 19:36, 11 August 2006 (UTC)

Assume bad faith?

Is it my imagination, or is a huge portion of this proposal devoted to specifying penalties for the admins who are, it would seem, expected to try to dishonestly interfere with the process? Fan-1967 19:21, 11 August 2006 (UTC)

Those were additions today to address concerns raised by comments here on the talk page. Those need to be stripped down. rootology (T) 19:23, 11 August 2006 (UTC)

Look now, specifically at Wikipedia:Admin_recall#Prevention_of_potential_admin_abuse_in_process rootology (T) 19:29, 11 August 2006 (UTC)

Fan, there's no harm in addressing the possibility. If they don't interfere with the process, those provisions won't apply to them. NO reason not to hae a safeguard. Karwynn (talk) 20:01, 11 August 2006 (UTC)

Prevention of potential admin abuse in process

Based on comments above, to just ensure that nothing untowards may happen, I added this with the intent of having a very simple, prescribed remedy for the unlikely event that anything like this might happen (retribution, etc.). What do you think of this section? What might be changed? rootology (T) 19:44, 11 August 2006 (UTC)

Instruction creep

This proposal tries to define in minute detail the precise way it should work, with a high amount of arbitrary numbers and precise rules. IIRC, the only place on Wikipedia where the amount of detail gets near what can be found on this proposal is on a board vote. This amount of instruction creep will only lead to wikilawyering, which is not a good thing. Wikipedia is not a nomic. --cesarb 19:53, 11 August 2006 (UTC)

Would you be willing to let me know what specifically you think can be removed? I'd be happy to edit this down to something more concise. rootology (T) 19:58, 11 August 2006 (UTC)

IMO this is way too detailed to begin with. Rather than starting with something slightly less complex than the US tax code, a better process might be to start with a very geneal outline of a proposal and then fill in the details. Starting with an explanation of what the problem is, and why existing procedures have been ineffective in handling such issues. Fan-1967 20:38, 11 August 2006 (UTC)

Also, the factoring in of a way for it to run is actually for the protection of the admins--it should be feasible for the community to remove an admin for abuse, loss of faith in their abilities, etc., if a sufficient segment of the community agrees, just as for RfA. The "precise" nature of it was specifically to protect admins, the same with the numbers mentioned. Plus as Karwynn mentioned below, it was (I hope) meant to kill any wikilawyering, on any side, so that any removal an admin the community deemed necessary was done 100% fairly and "above board". rootology (T) 20:05, 11 August 2006 (UTC)

To me, the amount of instruction actually lends itself to a much stricter interpretation, leading to less wikilawyering. Have you actually read the proposal, or did you just notice it was kind of long? Karwynn (talk) 20:03, 11 August 2006 (UTC)

From Wikipedia:Wikilawyering: "[...] Typically wikilawyering raises procedural or evidentiary points in a manner analogous to formal legal proceedings, [...]". Having excessively detailed rules which are supposed to be followed to the letter will only help wikilawyers. The more strict the rules, the more the wikilawyers are happy; the one rule they most dislike would be Wikipedia:Ignore all rules (how can you argue the rules if you are supposed to ignore them?).

An example of one instance of instruction creep is the wait time of six or nine months before a new review request can be filled. We actually already have something similar to that on AfD, the main difference being that there's no set period (a renomination can be removed if it's filled "too soon", but sometimes you can legitimately have a renomination the following day). You could change it to "no admin should be submitted to recall if he was submitted before recently, unless there's a good reason", but even that is not needed; repeatedly resubmitting an admin to recall, after he was already submitted five times in the last month, would be considered disruption (itself a rule with rather fuzzy boundaries), even if you wasn't the one who submitted the previous five times. Almost the whole proposal is littered with unneeded (and potentially harmful) detail like that (to see why this particular one is harmful: it creates perverse incentives, since the admin is immune for nine full months if he passed a review, and if he thinks his next review and RfA will fail he might go rogue while he can).

If you remove all the redundant rules, and expect people to use common sense, what you would end up with is simply something like "any user which is not a new user can nominate an admin for recall; if it's endorsed in seven days by 25 other users (also not new users), the admin must go through RfA again; if it's not endorsed, it gets desysopped". The rest can be dealt with by the other policies and common sense. If the policy is accepted, simply copy the procedures from RfC. And now that it's been reduced to the core, we can discuss the essentials (including how it relates to the several other proposals which failed in the past, and which kinds of effects it would create), instead of getting bogged down in detail. If needed, the detail can be added back later (but I doubt much would be needed, even if it were to pass). --cesarb 20:37, 11 August 2006 (UTC)

If you're that broad, there will be MORE wikilawyering. People will spend all the time arguing about who is and isn't a new user, who isn't established enough to vote, etc. This policy makes sure the process can be done, rather than disrupted or deleted entirely by an admin who doesn't like it. Frankly, common sense doesn't always apply to policies like these - policies that specifically affect admins more often than not. The point of this is to stop potential abuse right? ell, it's useless if the "provisions" are loose enough as to manipulated by those admins by deleting/protecting/whatevering they can do to disrupt the process. THe detail prevents that by not allowing a lot of room for interpretation. It's hard for me to believe that loosening the syntax of the proposition will actually decrease wikilawyering. This way, there is no interpretation, not techniicalities - it's spelled out, very clearly, so as to not be taken advantage of. If you still disagree, could you maybe walk me through a scenario where this process would be weakened or taken advantage of by wikilawyering? maybe I can see your viewpoint better with an example. --Karwynn (talk) 20:57, 11 August 2006 (UTC)

Signifcantly simplified

Please take a look now. rootology (T) 21:07, 11 August 2006 (UTC)

Specifically, the entirety of the "policy" is now contained here. rootology (T) 21:36, 11 August 2006 (UTC)

Time delay/exemption window

Looking back over this, the current version does not have the nine month exclusion. This worries me, as it might invite trolling and/or abusive usage of this system. Granted, if someone filed repeatedly, it would be clear disruption, but it could also be a problem if multiple users starting chaining these vs. a given admin. On the other hand, if an admin were to draw that level of apparent level community ire--that it was certified each time--it might be a sign that they are not fit to admin. I don't know. I am inclined to include the nine month window again, or least a six month one. Admin abuse during the window--should it occur hypothetically--could still be addressed via AN/I, RfC, and ArbCom. Thoughts? rootology (T) 22:35, 11 August 2006 (UTC)

A system of this kind is biased against particular kinds of admin work

I don't think any system of this kind, no matter how much effort is spent on creating safeguards, is going to get enough support to come into force. One problem I see that hasn't been covered so far, and can't be remedied by more safeguards, is that admins who do certain types of work will be more vulnerable than others. I don't mean that a certain type of attitude (rouge vs. non-rouge) would make an admin more vulnerable - that brings up the argument over whether self-declared rouge admins are to be commended for their boldness or whether they're just making excuses to be impetuous and wrong-headed, and I don't think it's necessary to drag that up.

What I mean is that, for example, admins who monitor the 3RR alert noticeboard will be much more vulnerable than admins who monitor the vandalism alert board. WP:AIV generally involves IPs and new accounts, who won't have suffrage under this proposal and rarely have any reason to claim that they were hard done by, whereas WP:AN/3RR involves long-established users who often feel intensely aggrieved at being blocked and think that they have good defences. "I was reverting vandalism", "If I didn't revert the page would be left at The Wrong Version, "My 4th revert was 25 hours after the 1st", etc. Blocking someone for vandalism gets your userpage defaced at worst, whereas if this proposal was active, blocking someone for 3RR would put you one step closer to losing adminship. It doesn't just apply to 3RR - those who close AfDs, which may involve long-standing pages worked on by users with more than 500 edits, are going to be more vulnerable than those who speedy delete new pages, generally created by those who just got here. MfDs tend to be even more personal than AfDs. And so on.

Also, any administrator who does get recalled will need at least 104 votes to be safe with the RfA. Remember that all the users that signed the recall petition will be certain to oppose the RfA, whereas no matter how many hundreds of users support the admin's work, they may not notice or support the RfA. Any attempt to remedy this would amount to vote-stacking. To get over 100 support votes at RfA is quite exceptional - that's why we have a special page for those that manage it.

Oh, and one more thing. 500 edits can be done standing on one's head if you know the right places (stub sorting, RC patrol, categorising uncategorised articles, and so on). The only real criterion is the 5 months.

So bearing that in mind, if this proposal went live, all you need to do is annoy 26 users whose account is more than 5 months old, and you've got a very high chance, if not close to a certainty, of being desysopped. It would certainly create a chilling effect on the vulnerable areas I mentioned above, and we haven't got enough admins to discourage people from taking on any of our areas of work. As Zoe says, this is a problem in search of a solution, and we'd need a really, really big problem to justify this solution. We haven't got one. Arbcom will do. If you're seething at something an admin has said or done and thinking "He/she should be desysopped" - take a deep breath, go outside, get a pint with some friends, go to sleep and in the morning you'll probably need your watchlist to remember what it even was. --Sam Blanning^(talk) 23:12, 11 August 2006 (UTC)

Thank you for the extremely thoughtful comments. It's got nothing to do with anything done by any admin, really, it just looked like a big hole in the current system. I've made a couple of adjustments to the Policy proposal specifically to address your concerns. Specifically:

1. Requires 30 users, with 10 months of experience each, and 2000 edits each to certify or file (31 total).

2. 3 of the certifiers must be fellow Administrators.

3. Re-added the nine-month limitation.

4. Recall RfA concensus must be 80% or higher.

5. Recall RfA remains open for 21 days. Certification remains open for 7 days.

I will admit that some areas of the admin work will draw more ire by their nature, but if 31 total users, 3 or more of which are administrators, agree that a Recall of the admin is in order... wouldn't that be a sign that perhaps something was actually an issue with the given administrator? I'd just like to reiterate that my goal in this is to develop a means for the community (admins and editors) to actually control who has administrative access on an ongoing basis. rootology (T) 23:51, 11 August 2006 (UTC)

Now the requirements are getting so sky-high I think it would be easier to get an Arbcom case filed and accepted before those conditions were met. Certainly I'd much rather do that than hunt around for people licking wounds from the administrator in question, not to mention finding 3 administrators (you can't possibly expect this to work without allowing filers to solicit certifiers - no way would 30 potential ceritifiers randomly stumble over this page). You'd almost certainly piss off every admin and probably every non-admin you asked to certify who didn't think it justified, and there would be far more bruised feelings when the recall failed than after Arbcom rejects a case.

I don't think the requirements are the problem - in my mind there is no particular set which would be lenient enough to make recall possible but stringent enough to make it only possible when justified. I think the process is intrinsically flawed, sorry. --Sam Blanning^(talk) 00:06, 12 August 2006 (UTC)

I figured that any kind of action like this would result in bruised feeling and hurt egos, unfortunately--the same with RfC or if someone actually filed an ArbCom against you. But if the certification and the Recall RfA were both listed both on the Recall and RfA pages, you don't think it would generate enough views by the community to make it potentially worthwhile? Keeping in mind that the first would be there for 7 days, and the latter for 21? rootology (T) 00:14, 12 August 2006 (UTC)

Not 30 users with more than 10 months of editing and 2000 edits who would publicly support a recall. Bear in mind that I only passed that threshhold myself last month. I think at any point where that became possible, one of the 30 would have filed an Arbcom case (or anyone for that matter, as anyone can file Arbcom cases). I also think that the way Arbcom deals with evidence for and against, and suggesting what remedies should be taken, is far better than anything that could be set up here. --Sam Blanning^(talk) 00:29, 12 August 2006 (UTC)

Sam, what do you think of my idea of just a place to post evidence of abuse? If the evidence of abuse has to stay there and can't be deleted, then it would be some progress. If it is evidence, then someone could submit it anonymously and not get retribution. It's like if your neighbor's blasting their music loud and you confont them all the time, they hate you, but if you call the cops, the noisy neighbor won't know whose tires to slash. An evidence page may simply shame an admin who is doing the wrong thing into doing the right thing. Badlydrawnjeff said that is too little, but right now evidence pages would probably get deleted and the person who made them banned.

I also have an idea that would solve the editing thing. You mentioned that someone smart could do lots of edits really fast. This also is hurtful to newcomers who get picked on well about everywhere as "newbies." What of instead of so many edits, we require say 300 different days of one edit each to an article or a talk page. That way nobody could just do a bunch of editing over a few days, but they'd have to come back every single day. I think 300 different days with each an edit to an article or article's talk page is harder than 10 months and 2000 edits. I think article talk pages should count because it would be harder to do--no spelling fixes but actually writing something that doesn't look like spam.Anomo 03:42, 12 August 2006 (UTC)

For the first, we have RfC for that. If there's a genuine problem, you'll find two certifiers, and then you can collect as much evidence as you like.

For the second - well, I'm inclined to think that any criterion that excludes me as 'too new' isn't a good one, and I don't think that's just self-obsession, I think I have enough length and breadth of experience with Wikipedia's policy and people that any criterion that excludes me from participating as 'too new' is rather bizarre. 300 days is 10 months, and I've been here for 11, and I'm sure there must have been 30 days where I didn't make an edit. I mean, I really hope there was or I'm more of a Wikiholic than I feared :-). Anyway, I don't think we have a system to even measure that at the moment. And as I said - the requirements aren't the problem for me. --Sam Blanning^(talk) 12:58, 12 August 2006 (UTC)

Admin Recall committee

I think the procedure for this policy needs to be fully planned. First of, Who will review this claim? Who will plan the vote, who will plan meetings. This is why I believe there needs to be a recall committee. Here is my proposed plan:

A committee called the Admin Recall Committee (ARC) This committee will be a 7 person committee that will review filings on admins. There will be a chairman of the committee who will be appointed by the maker of this policy. The Chairman will be incharge of... Setting mettings, Having hearings, Be in Charge of documenting what goes on. This committee Will also be incharge of the proceedings for the reviewing process and voting process. --Zonerocks 01:49, 12 August 2006 (UTC)

Good God, just how many levels of bureaucracy are you planning on creating in all of this? User:Zoe|(talk) 03:49, 12 August 2006 (UTC)

Hopefully enough to sink the proposal. --Carnildo 03:55, 12 August 2006 (UTC)

I said much the same thing below ;-) Kirill Lokshin 03:56, 12 August 2006 (UTC)

Which was probably Zonerocks' intent; "who will plan meetings"??? -- nae'blis 02:36, 15 August 2006 (UTC)

Then we come back to: why not just let ArbCom handle this? Fagstein 16:36, 12 August 2006 (UTC)

Because they don't do it well, or when necessary? --badlydrawnjeff talk 16:39, 12 August 2006 (UTC)

We already have the Admin Recall Committee, it's called the Arbitration Committee. --Cyde Weys 16:41, 12 August 2006 (UTC)

Quite. This is becoming absurd. Why create this byzantine and untried mechanism when we already have the ability to revoke someone's administrator status and have done so? --Tony Sidaway 17:14, 12 August 2006 (UTC)

According to Wikipedia:Requests_for_de-adminship, "As of June 2006, there are twelve cases, involving fifteen users, where adminship has been removed involuntarily for more than a trivial length of time." Why create a system? Because it's simply not easy enoguh to get poor admins removed. --badlydrawnjeff talk 17:20, 12 August 2006 (UTC)

There have only been twelve cases because most calls for desysopping are over trivial matters. Incorrect administrator actions that lead to calls for blood can almost always be reversed with little harm done (assuming they were wrong in the first place), and the only case for further action is 'hurt feelings', which are usually solely the problem of the person feeling them. --Sam Blanning^(talk) 17:25, 12 August 2006 (UTC)

I'm sure many are. Certainly, however, there have been many, many non-trivial situations that have been ignored or discouraged from pursuing due to Arbcom's lack of action on the matter. --badlydrawnjeff talk 17:30, 12 August 2006 (UTC)

So, jeff, what's the procedure for getting admins unsysopped over at ED? User:Zoe|(talk) 03:15, 13 August 2006 (UTC)

I honestly have no clue. --badlydrawnjeff talk 12:27, 13 August 2006 (UTC)

The problem is that the current system empowers an extreme minority of users to make decisions on a process that on the other end of the same phase gives absolute 100% control to the end users of the project. Currently, if 30 admins collectively determined that one of their peers should not be a fellow administrator any long due to his actions, and ArbCom does not agree, that is the end of the story. This is inherently broken. This better way basically opens the door to allow those 30 admins to create a venue for public discussion, where the majority then gets to decide if they are right or wrong, and whether the "bad" admin in question should go. I fail to see any logical reason (no offense) for this to not be policy, except that some admins would be nervous about being Recalled themselves from their lifetime appointment. This system by it's nature will empower the communit to control itself fully, and no admin in good repute with his peers (see the 3 admin requirement) and the general community will have nothing to worry about. rootology (T) 17:42, 12 August 2006 (UTC)

It's only inherently broken if your example - that "30 admins collectively determined that one of their peers should not be a fellow administrator any long due to his actions, and ArbCom does not agree" - ever actually happened. It's never happened and I consider it so unlikely that we're back to the fact that this is a nigh-on unworkable solution looking in vain for a massive problem. The problem only exists in the swamps of "might"s and "theoretically"s and "yes, but what if?"s where process creep is born. Of course some admins would be nervous about being recalled; as I said, the trouble is that only the minority of them would actually deserve desysopping, and that minority can be dealt with by Arbcom. --Sam Blanning^(talk) 18:29, 12 August 2006 (UTC)

OK. Reasons for this not to become policy:

1. It's instruction creep.
2. It assumes a problem that hasn't existed before.
3. It gives power to users to punish (or threaten to punish -- even if it fails it still has a negative impact) administrators without giving a valid reason or showing evidence of abuse.
4. It contains a number of arbitrary thresholds which are at the same time too strict and not strict enough.
5. There is no prior requirement for RFC or anything other than a 24-hour warning on the admin's Talk page (which will no doubt become a favourite threat of trolls).
6. It doesn't clearly explain why ArbCom can't handle this using its existing procedures.

I'm sorry. It's a great idea in theory, but would be a horrible disappointment in practice. Fagstein 19:23, 12 August 2006 (UTC)

Some rebuttals:

1. Not all instruction creep is bad.
2. Admins are not really accountable by the same editors who promote them, so it assumes a problem that has continued to exist, not one that is being conjured up.
3. Wrong, this has those failsafes in place. If you feel said failsafes are not enough, perhaps you can expand this.
4. That can be fixed as well.
5. Again, this could be fixed.
6. ArbCom has not shown a history of removing administrative power.

Many of these reasons can be dealt with. Why not? --badlydrawnjeff talk 19:28, 12 August 2006 (UTC)

Arbcom committee would probably be biased and stick up for the admin. not allowing it to go to another rpoceeding. --Zonerocks 18:18, 12 August 2006 (UTC)

Please keep it civil. rootology (T) 18:20, 12 August 2006 (UTC)

Zonerocks's long suggestion

Appeal Review Committee

This committee will be in charge of reviewing appeals made by the admin, during all procedings. )Read Section appeal to understand this next part.) Say an admin gets voted off as an admin by the community. He may file an appeal as any defendant can after a criminal trial to suggest that his proceeding was either an abuse of power by the members on the Admin Recall Committee. That the members of the committee were unfair, seemed to side with on side of the issue, or violate any wikipedia policies. You may also file appeals against the filers or supporters of the persons on the admin, for these reasons.... If your rights, policies, abuse of power, misrepersentaion of the facts, or ruining the admin's reputation (delibritly), can be on the grounds of an appeal. So back on the example, The removal of the admin won't intiate until the Appeal Review Committee makes a decision on the appeal. If they pass the appeal, the committee will take in consider that there was a community vote, and they will move to have a vote on wither the admin should be removed or not knowing that there was an abuse of power or any of the other grounds for an appeal. --Zonerocks 01:50, 12 August 2006 (UTC)

Appeal

The grounds for an appeal are: f your rights, policies, abuse of power, misrepersentaion of the facts, or ruining the admin's reputation (delibritly), can be on the grounds of an appeal. Of course there has to be documentation and evidence and can't be off a n opinon. --Zonerocks 01:50, 12 August 2006 (UTC)

Procedrual Process: Review Process

The First step of the procedrual process will be as the timeline of 7 day period of meeting. The review procedure will endure The First step of the procedrual process will be as the timeline of 7 day period of meeting. The review procedure will endure interviewing the filier and the admin. Discussing evidence etc. Then on the 7th day period the committee will take a vote on the filing. If it passes by a majority, it goes to the next step. --Zonerocks 01:51, 12 August 2006 (UTC)

Procedrual Process: Voting Process

The Administrator will be able to testify and give his aguments to claim. The Chairman of the ARC will be in charge of the procedrual schedule during that time. the evidence will be discussed by the filier and Admin. Then after three days of this, a poll will be takin during a seven day period. After that the votes will be tallied and revelead. --Zonerocks 01:51, 12 August 2006 (UTC)

Bureaucracy

It looks like this proposal is quickly being strangled with red tape. Filings, certifications, votes, reviews, appeals, committees... this is quite ridiculous; in my opinion, it hasn't a snowball's chance in hell of achieving consensus—or of actually working if it does. Kirill Lokshin 01:59, 12 August 2006 (UTC)

This is a proposal to be added to the policy at hand. This policy ( Admin Recall ) should be taken seriously, especially when your dealing with someone losing there adminship and respect. So there needs to be added proposals to this policy. Look at the admin recall policy as a DOD bill, and this propsal I made is like an amendment to be added to the bill. --Zonerocks 03:57, 12 August 2006 (UTC)

Exactly. Someone losing their adminship is a big deal. That's why ArbCom handles it. They're not overloaded with de-adminship cases, and while having community control over adminship would theoretically mean a policy like this one to take that away, in practice this will be used almost exclusively by users with gripes against administrators, while people with real, serious complaints will still go to ArbCom to have it dealt with more quickly. Fagstein 16:43, 12 August 2006 (UTC)

It allows the community to have a vote. Arbcom is going to be more favorable of the admin's. This should be seriously taken --Zonerocks 18:20, 12 August 2006 (UTC)

It would perhaps be taken more seriously if you stopped engaging in monumental assumptions of bad faith like that one. Kirill Lokshin 18:23, 12 August 2006 (UTC)

If we don't trust Arbcom to deal with bad admins, then I think we have a serious problem with Arbcom and need to deal with that instead. Fagstein 19:12, 12 August 2006 (UTC)

And that's the heart of this proposal. I'd like to see a detailed explanation of why the existing system is flawed, and why a new one is needed. Does the author of this proposal feel there's an admin who needs to be removed that Arbcom hasn't addressed properly? Fan-1967 21:19, 12 August 2006 (UTC)

How about cut all the bureaucracy? If a certification is made (with the 31 users, including 3 admins), the admin must re-RFA. If they fail that, they lose their adminship. They would, of course, be welcome to re-apply for adminship (no need for 'appeals' BS) - just like most people desysopped by the arbcom - assuming of course that they feel able to pass an RFA having just failed one. Cynical 23:39, 13 August 2006 (UTC)

Where is all this bureacracy? There are two steps: getting 30 or however many people to certify, and then putting it to a vote. THere are applicaple time periods - one week for certifying, one week for voting, nine months for re-submission. And that's it. Karwynn (talk) 14:05, 14 August 2006 (UTC)

About the only chance I can see to get this close to achieving consensus

Can those supporting this proposal point out any Arbcom cases where they felt that an administrator who deserved desysopping wasn't? This proposal is clearly not going to pass the way the discussion is going, so there's not much point in politely refraining from naming names at this point. --Sam Blanning^(talk) 18:31, 12 August 2006 (UTC)

It may help here if editors take a look at the page Wikipedia:Requests for arbitration/Completed requests. Each completed case is listed, and there is a summary of the parties and remedies, and a link to the case page with full details. --Tony Sidaway 21:16, 12 August 2006 (UTC)

One sysop who was de-sysoped for using sock puppets combined with their own sysop account to win edit wars. That sysop did a lot of damage before finally being caught. I'm sure a lot was even now not found out. Anomo 12:10, 14 August 2006 (UTC)

No-one knew what he was doing until he was caught, and when that happened he was summarily desysopped. Admin recall wouldn't have been able to do anything. --Sam Blanning^(talk) 13:34, 14 August 2006 (UTC)

Googling his name turned up people complaining about him off-wiki that went back many months before he was de-sysoped. It was message boards. Anomo 14:23, 14 August 2006 (UTC) By the way, something really weird happened. I began writing this before they other two started talking and it was left in show preview. When I finished it was after they wrote this but magically no edit conflict happened. I don't know how. Maybe wikimedia made something new in their software? Anomo 14:27, 14 August 2006 (UTC)

Naming cases would be an insult to the admin in question, so that's not a good idea. Plus, another perceived problem is that there are times when an admin needs desysopping, but doesn't even GO to ArbCom in the first place. Karwynn (talk) 14:06, 14 August 2006 (UTC)

Sam, what if you have a situation where a good sized group of admins and editors all come to the conclusion that "Admin X" is not fit to be an admin anymore, but the scale of evidence isn't something that ArbCom would act on--but they see a trend of let's say continued behavior for a hypothetical situation, based on which they would not have passed this person on the original RfA. What then? It's not the sort of thing that ArbCom would ever remove someone for, but based on their behavior they would in all likelihood have difficulty passing an RfA. What if these admins and editors feel this person based on their actions shouldn't have the tools? They just have to sit on it? Live with it, because they made the decision once to empower this person and now their behavior is completely different than previously? rootology (T) 14:21, 14 August 2006 (UTC)

You mean small abuses over the long term instead of a single large instance? Anomo 14:27, 14 August 2006 (UTC)

Question on policy vs belief for admins reading this

For those that have commented: are your objections based on the fact that you feel the policy proposal is flawed, and if so, what specific points? Or, is your disagreement with it based on a belief that admins should not be de-powered by the community, and if so, why? rootology (T) 20:20, 12 August 2006 (UTC)

Why do you feel that the proposal is necessary. How have existing processes failed? Fan-1967 03:14, 13 August 2006 (UTC)

My opposition is based on the idea that there is no need for this proposal, as no one has shown that the current process is flawed. User:Zoe|(talk) 01:20, 14 August 2006 (UTC)

Not an admin, but any system that uses a hard-and-fast number of certifiers while we're experiencing the level of growth of Wikipedia currently ongoing is unable to get my support. Sorry. I'm glad you've trimmed a lot of the bureaucracy, but this still isn't scalable or ungameable. -- nae'blis 04:30, 14 August 2006 (UTC)

Not an admin either, but I think the problem is that it will create more problems than it will solve. The points are expressed numerous times on this page, but here's an analogy for you: Should police officers be elected by the people they fine for speeding? Fagstein 07:10, 14 August 2006 (UTC)

It will be everyone who is allowed to vote here, not just those upset with the admin. This is like saying, "should local judges be elected by supporters and non-supporters alike?" The answer, of course, is yes.

I also notice that almost no one has answered this question, other than to raise other questions. Could we maybe put new questions in a different header and actually answer this one? I think answering this one will be a big step towards productive discussion, rather than the all-around iciness and suspicion that seems to be going on here. Karwynn (talk) 14:24, 14 August 2006 (UTC)

Total waste of time as currently framed

This is far too strict, biased towards inertia and admin power, and so unlikely to be used a worthwhile amount that it isn't worth wasting time on. My suggestion:

All administratorships run for six months. Three months after a period as an admin has expired a former admin may apply for a futher stint.

That should put an end to administrator arrogance and the development of an invidious oligarchy. This proposal is just windowdressing. Sumahoy 21:34, 12 August 2006 (UTC)

Please keep it civil. Expiring duration adminiships would be a bad idea--can you imagine having to rerun a thousand RfAs every six months? Or 2000+ in a couple of years? This process is only to enable the community to remove adminship from one of their peers that the collective community no longer feels should be an admin. rootology (T) 23:21, 12 August 2006 (UTC)

• Been there, done that, didn't work. I tried pushing for that myself! - Mailer Diablo 10:10, 13 August 2006 (UTC)

I have to agree with the original poster, the 2000 edit limit severely hinders the proposal's effectiveness. There are times when a harmful admin simply wouldn't affect that many people, but would still deserve desysopping. I think it ought to go back to 5 months, 500 edits. Karwynn (talk) 14:10, 14 August 2006 (UTC)

Recall by peers

This should instead be a process of recalling administrators by their peers. If you can't even pass RFA yourself it doesn't make sense that you should be able to get admins kicked out. Generally fellow admins are much better at knowing which admins are acting badly than the larger editor population does. Admins face lots of problems and situations that the general editor population just doesn't understand, and they might take it out on an admin who didn't do a thing wrong; it was just poorly perceived by the people who were affected by the actions. Other admins will understand this, however. --Cyde Weys 03:18, 13 August 2006 (UTC)

That sounds dangerously like the argument where I tell a music snob that I don't like (insert shite prog rock band here) and he asks me if I can play as good as them, then says "you shouldn't criticise if you can't do better". Or "how can you consider yourself qualified to support this article's deletion when you know nothing about the subject". I don't think the judgement of a given admin is necessarily more reliable than a given user with the same edit count and length of experience. There are plenty of sensible users who understand Wikipedia policy that, for various reasons, don't become admins, including lack of time, being unable (or unwilling) to maintain the necessary standard of civility or simply not wanting to be an admin, and taking or appearing to take an elitist attitude towards adminship should be avoided. It should be repeated at this point that I think any admin recall process is unworkable and unnecessary. By the time we actually did want to recall a fellow admin, one of us would file a request for arbitration. --Sam Blanning^(talk) 13:00, 13 August 2006 (UTC)

What an inviation to back-scratching and logrolling, Cyde. Non-admins are not necessarily dolts. Herostratus 16:46, 13 August 2006 (UTC)

Cyde, there is already such a provision within the proposal - three of the certifiers must be active administrators. Therefore no admin who has the support of his (or her) peers will be subject to a recall. Cynical 23:34, 13 August 2006 (UTC)

I have to echo Mr. Blanning's statements: I'm not really taken with this particular proposal, but it bears repeating, again, that adminship should be "no big deal." It's primarily a janitorial assignment, and while I have a great deal of personal respect for the people who take it upon themselves to carry out those tasks, I firmly oppose any attempt to reinterpret it beyond those boundaries. Additionally, I worry that this kind of rhetoric is profoundly counterproductive, leading to the perception that administrators are elitist and "out of touch" with non-administrative editors. Whether or not you personally believe this to be the case, that perception does exist. The fact that we're having this discussion now is testament to that. I see no reason to fan those particular flames. – Sean Daugherty (talk) 04:31, 15 August 2006 (UTC)

Policy to hard to file

How can you go from 5 months 500 edits, to 10 months 2000 edits. It makes no sense. It is a big jump. Also i have a probelm with 31 signatures. I wouldn't think one case could be filed with a needed 31 signatures. In fact I would object to 26. I say 15. --Zonerocks 04:39, 13 August 2006 (UTC)

suggested a better idea of 300 unique days with edits. That's real hard to fake. Anomo 07:11, 13 August 2006 (UTC)

Too hard to track, Anomo. --Zonerocks 07:36, 13 August 2006 (UTC)

Actually, it would be easy if you use an automated system to do it. Already that's what people do to count edits. Anomo 12:04, 14 August 2006 (UTC)

• Eek, that is a huge jump. Even a Grognard would not qualify. That is too high a bar in my opinion. Really, is the concern puppets, or what? If it's puppets, is it not so that the admin in question and/or his supporters will be able to root out puppets, just as is done for current RfA's (and AfD's, etc.). If the question is standing, I don't think you need that high of a bar to be considered to be in good standing as a contributing member of the community, at all. If the previous level was considered too low, how about 6 months, 1,000 edits. That still is awfully high but an improvement at least. Herostratus 16:29, 13 August 2006 (UTC)

15 signatures herostratus. I say 3 months 700 edits, but 16 signatures. --Zonerocks 17:37, 13 August 2006 (UTC)

Why does everyone think 2,000 edits is that hard...? I did 2k+ in about eight months' time, between state tag sorting (about 40%ish) and the rest between main space and mostly Wikipedia side... rootology (T) 06:48, 14 August 2006 (UTC)

A telling comment on the current procedure

I think this diff speaks pretty loudly about the current recall procedure. Here we see an administrator laughing in the face of the community over the very idea that an admin could actually be desysopped. What a way to run an organization Herostratus 14:23, 13 August 2006 (UTC)

Way to misinterpret. I was laughing at the suggestion that a particular administrator could be desysopped over that particular action. People who are constantly calling for desysopping when it really doesn't make sense in that circumstance at all serve only to discredit themselves. --Cyde Weys 16:38, 13 August 2006 (UTC)

Uh OK, I gave you two chances earlier to explain that diff, Cyde, at the location of the diff and on your talk page. It doesn't scan the way you explain to me, but if you say. Herostratus 19:13, 13 August 2006 (UTC)

Adminship can be revoked. Just ask Karmafist or HOTR or Guanace (twice). --Cyde Weys 19:22, 13 August 2006 (UTC)

Yes but the point is that it generally can't be revoke just (sic) because an admin has lost the confidence of the community - he or she has to lose the confidence of the Arbcom. Cynical 23:36, 13 August 2006 (UTC)

The community chooses the members of the Arbcom from the community. Arbcom is part of the community, just with better judgement and experience than most of it.

Anyway, confidence of the community is pretty nebulous. People call for desysopping all the time, they just get ignored, I'm not sure where you could draw the line between "lost confidence of the community" and "storm in a teacup that will be forgotten tomorrow". Indeed, one of the problems that occurs to me with this process is that when someone feels they've been dealt with unjustly by an admin, rather than going to sleep, waking up with some perspective and quietly forgetting it, they'll be encouraged to start stalking the admin, digging through their past edits looking for other previously wronged editors, and going around trying to organise a recall - in effect scratching their own itch until it becomes a gaping sore. That's not good for anyone. --Sam Blanning^(talk) 23:49, 13 August 2006 (UTC)

Well, 23 other editors with 500 edits and 4 months service is kinda a way to draw the line between "lost confidence of the community" and "storm in a teacup", n'ese-ce pas? There are situations where "go sleep it off" is not sufficient. Herostratus 06:19, 14 August 2006 (UTC)

{{subst:Wikipedia talk:Admin recall/And}} --Sam Blanning^(talk) 13:32, 14 August 2006 (UTC)

Which is slow and bureaucratic. Haukur 17:31, 14 August 2006 (UTC)

So replacing it with a month-long recall vote will be faster? Fagstein 18:07, 14 August 2006 (UTC)

It would be in my ArbCom experience. --badlydrawnjeff talk 18:09, 14 August 2006 (UTC)

Not meant to be fast--the idea isn't to steamroll an admin, it's to give the community and the admin's peers to make the determination if he's due to be an admin based on his history after his initial RfA. Certification: is the filing for Recall with merit (one week)? Recall RfA: is he fit to perform as an admin (21 days--to at all times give the admin benefit of the doubt, and to allow people to change their opinions/votes). Plus as mentioned above, with the 80% concensus requirement, it amounts to a supermajority of Wikipedians agreeing that the admin shouldn't be an admin--totally fair. It's meant to be a fair and simple process, with ample time for people to change their minds. rootology (T) 18:13, 14 August 2006 (UTC)

I've had people, even administrators, angrily demanding my desysopping over something that seemed like a big deal to them at the time, but that in the long run, ended up being, at most, a minor problem. For instance, I may have close a certain "rationales for impeachment" AFD in a controversial fashion ... does anyone even remember that? It got sorted out in the long run, and I don't think (or at least I hope) that people still aren't holding a grudge against me for that, but at the time, tempers were very inflamed. I would thus propose some sort of waiting period for this Admin Recall period so that people can have the benefit of 20/20 hindsight and determine if something was truly worth losing adminship over. --Cyde Weys 17:33, 14 August 2006 (UTC)

That's already built in, inherently. It takes an entire month for the community process to run, and the admin can't be caught flat footed with a "SURPRISE YOU GOT RECALLED!!!" action due to the 24 hour waiting period. Even then, it requires a substantial number of people to agree. Say you closed an AfD on me that I was POSITIVE had to be a keep, but it was borderline--concensus wasn't happening, so you judge delete. I think to myself, "Oh man, that Cyde is a son of a bitch, I'll desysop him!" (hypothetical, not saying your a SOB). So I file (which I can't since I don't even meet my own criteria for another month or so, but I digress). People *can* file a stupid Recall; the same as I can file a bogus RfC that says "Cyde was mean to me," or "Cyde hates ferrets," or "CydeBot totally is like lame". But the difference is that my evidence would consist of "Cyde's dumb machine made fun of my pet ferret Whiskers," and no one would certify it aside from a few stupid trolls, and you'd have a good laugh over it. And then no one could Recall you for nine months. Even if by some insane fate 3+ admins and 27+ others agreed, it'd never vote you out on Recall for being mean to my ferret. And even IF somehow you got railroaded out and desysopped--see how it's getting more and more implausible for a bogus Recall?--your rerunning 5 minutes later for a new RfA would almost certainly pass given your history. rootology (T) 17:42, 14 August 2006 (UTC)

THe process takes 4 weeks. By then, steam will have blown off long enough for people to be rational and possible reconsider. Karwynn (talk) 17:38, 14 August 2006 (UTC)

Thanks for being more succint than my small animal-based explanation. rootology (T) 17:43, 14 August 2006 (UTC)

? Must've missed your reply when I posted mine, whoops. FOr the record, I'd be in support of a "SURPRISE YOU GOT RECALLED!!!" system just for the laughs. But in all seriousness, I don't think short-term grudges will be a problem, given the length of time and the numerous safeguards against recaller abuse. No one takes that stuff seriously, it'll never go through. Frankly, after seeing in a couple of RfCs how quickly fellow admins leap to one another's defense unconditionally even when presented with at least well-thought out, not emotion-driven evidence, the probability that this will be a problem is hovering around zero. 3 admins, 27 users, and a supermajority vote is what they'd have to go through. And how many grudge-bearers and trolls even have the patience for that, now that I think of it? Karwynn (talk) 17:54, 14 August 2006 (UTC)

A lot. And the lamest grudges are usually the longest. Fagstein 18:11, 14 August 2006 (UTC)

The lamest grudges won't have community support. This isn't some closed, individual incident process, people will flock to them like RfAs and AfDs. Probably moreso, since these I'm guessing will be rare. Karwynn (talk) 18:17, 14 August 2006 (UTC)

Posted to Sam up-page, but for everyone

What if you have a situation where a good sized group of admins and editors all come to the conclusion that "Admin X" is not fit to be an admin anymore, but the scale of evidence isn't something that ArbCom would act on--but they see a trend of let's say continued behavior for a hypothetical situation, based on which they would not have passed this person on the original RfA. What then? It's not the sort of thing that ArbCom would ever remove someone for, but based on their behavior they would in all likelihood have difficulty passing an RfA. What if these admins and editors feel this person based on their actions shouldn't have the tools? They just have to sit on it? Live with it, because they made the decision once to empower this person and now their behavior is completely different than previously? That is what it boils down to. ArbCom will nuke me as a rougue admin if I directly abuse the tools, but that is not the only reason for removing someone. As for the police comment above--does the community decide who is a police officer in their community? Actually, yes, they do. They fund them, they finance them, and there are grievance processes. In any community which was not corrupt in it's support of it's officers, if 31 people were to sign an affadavit expressing extreme concern about the actions of a given officer, WITH recorded evidence of the officer's actions, you can bet your butt that same officer would be hauled in front of a review committee of some sort. That's the problem. The community gets to install admins but then that's it. The man or woman they install a year later, two years later, might not be anything of what they DID approve, and the community and their admin peers need a way to rectify that.

Nothing in this policy requires evidence of abuse, nor is any provision made for dismissing frivolous petitions. And I would certainly hope that police officers where I live would be subject to a vote if people sign a petition against them. Fagstein 18:22, 14 August 2006 (UTC)

Actually, the filing clearly has an EVIDENCE section in it. And based on the merits of that, the certification either happens or it doesn't. If it doesn't, Recall closed, see you later. The community and the admin's peers (3+) decides if it's frivilous or not.rootology (T) 18:28, 14 August 2006 (UTC)

I'd guess that 99.9% of admins would have NOTHING to worry about with this, and those that work well with their peers (admin or editor) have nothing to worry about. I strongly suspect any disagreement with this proposed process by admins in this talk page--with the exception of Sam's erudite opinions which are based on possible problems for admins who handle one specific subset or two of work--are unfounded and based on a belief that admins will be targetted for removal for being unpopular. THIS IS NOT A PERSONAL attack, but some of the most vociferous anti-recall voices here, Sam excluded, from my tenure here on Wikipedia I see generally getting into the nastiest scraps with their peers over a variety of issues. I think anyone viewing this except objectively does the project a disservice. Good admins that work well with others and respect others would have nothing to fear. rootology (T) 14:32, 14 August 2006 (UTC)

Isn't the whole point of this policy to remove admins who are unpopular? Fagstein 18:22, 14 August 2006 (UTC)

No, the point is to give admins and their peers a form of oversight over each other without having to wait for ArbCom to decide whether or not to take action. ArbCom covers in theory 'anything', this is specifically "part 2" of the RfA. Basically, the point is that if someone presents evidence that someone is not a good admin/fit to admin, the community decides (certification) if that evidence and cause is worthwhile. If the community and the admin's peers agree, then it moves to a long RfA to allow the community to weigh all the evidence and evaluate it. rootology (T) 18:30, 14 August 2006 (UTC)

Karwynn's comment

This proposal is pretty analogous to the part of the blocking policy that deals with editors who have "exhausted the community's patience" and can be banned indefinitely without going to ArbCom. If someone expressed opposition to that, think about it: you don't have to answer here, but what would you honestly think? Probably that the person who did so was a troll who wanted less accountability for himself. So you'll forgive me if I too suspect a little baseless worry. I am NOT saying admins who oppose this are beligerent admins or anything of the sort, just that you seem to be, like rootology said, the ones most often in conflict (nothing inherently wrong with that), and it would seem as though you are pushing against administrator accountability. Wikipedia is built up on accountability to the community. The contributions of all editors and the history of all modifications to any page, even in wiki-central namespaces like Wikpedia: and Template:, etc., are available for anyone to view. Proposal is written, actively edited, and enforced by editors, mostly non-admin editors (who "enforce" policy by reporting vandalism, trolling etc.). COntent disputes in articles, the central units of WIkipedia as an encyclopedia, are decided by a group consensus. Normal editors are subject to a community agreement to remove their editor status (by banning them). Is there any reason why administrators should not be subjected to a similar process? The policy seems to me very thorough in guarding against trolling, puppetting, and pettiness. Look at the requirements in the number and status of editors for certification - no way a sockpuppeteer is going to achieve that, nor will a group of 25+ like-minded trolls get that many edits without being banned AND manage to convince 3 administrators to support them. Admins are notified 24 hours in advance, and so cannot be "ambushed" with this. I can see no room for gaming or manipulation here - if you see ways this can be taken advantage of, please name them. 'll ask this again because I am really eager for an answer: Is there any reason why administrators should not be subjected to a community evaluation process? Karwynn (talk) 14:53, 14 August 2006 (UTC)

I think Karwynn means a binding community review process, as opposed to RfC which is as toothless as you can get (beside it's ability to expose misdeeds). rootology (T) 14:56, 14 August 2006 (UTC)

Yeah, RfCs are worthless in cases of admins or admin favorites - my experience is that it just turns into an unproductive troll-scare. A community evaluation with actual potential results is needed. Karwynn (talk) 14:59, 14 August 2006 (UTC)

Ad hominems aside, 30 users can't sign a petition to have me banned. Yes, Wikipedia is a consensus-oriented organization, but administrators by their very nature will do things to make them unpopular. Closing AfDs, banning users, protecting pages. Admins who are prolific in these things will be punished by recall petitions every nine months, and threats to file them on their Talk pages every day. I think they need extra protection from potential retaliation. Fagstein 18:31, 14 August 2006 (UTC)

A) Which ad hominems?

B) YOu will not get three admins, 27 editors with 2000 edits, and a supermajority vote against you if the recall is based on something invalid like that. Karwynn (talk) 18:34, 14 August 2006 (UTC)

What do you think of an addendum that if someone files the complaint they must go forward with certification process after the 24 hour period? That would push it directly into the process and offer the protection window. And at the same time, would stop people from filing frivilous complaints for two reasons: 1) stops stupid filings, by limiting their frequency; 2) stops potential reverse abuse by people getting someone to complain/file vs. them to get protection. All sides covered? rootology (T) 18:38, 14 August 2006 (UTC)

It sounds kind of silly to force people to begin a process. It would also mean that the notice would become part of the process itself, bringing back the "SURPRISE YOU GOT RECALLED" problem. It would probably make it better, but I don't think it would solve all the problems. Fagstein 19:31, 15 August 2006 (UTC)

A new requirement idea

I hear people saying "oh what if willy on wheels does this" and I think the danger is actually of false ones is more of if people don't like each other. SeeWikipedia_talk:Admin_recall#not_vandals_that_can_game_the_system.2C_but_cliques. I don't think 25 admins or long term editors should be able to use this to gang up on someone. So I'm just proposing that a good idea is a requirement on evidence. A lot of the criticism on message boards found with a simple googling isn't a big incident, but small stuff. Just as an example, it would be always assuming bad faith. If an admin always assumes bad faith about people, well it's bad. So why not put into the requirements that a page of evidence needs to be up. It would be a certain required number of evidence of breaking policy, even WP:BITE before this can start. The important thing is, even newbies should be able to provide on one of these, even if they don't meet the criteria to vote. The evidence needs to be able to hang around--staying up forever--until enough gets gathered and someone should review it as enough solid proof before the 25 can vote. Let's say maybe there's 100 evidence of things like assuming bad faith, personal attacks, WP:BITE, etc. before anyone can do that vote. What do you think? Anomo 15:12, 14 August 2006 (UTC)

Do you mean a numerical value? I.e., admin has to have x number of policy violations as a condition of filing? There's already an evidence section by the way that has to be completed for filing (really, this is just a VERY streamlined combination of RfC, RFA, and an ArbCom type filing). rootology (T) 15:15, 14 August 2006 (UTC)

This seems like an opening of the floodgates of wikilawyering. The current proposal leaves little or nothing to interpretation. With this addition, there will be little discussion on the merits of the admin and much bickering about what counts as evidence and what doesn't, what is a legitimate policy violation, and it will all spiral, as all things critical of admins seem to do, into a contest to see who can find the meanest way of calling the other side a pack of trolls. I think this is a bad idea. If there is no credible evidence of abuse, it will be very difficult to find people to certify, and even harder to find enough people to get a 60-70% consensus to desysop. In short, meritless recalls will speak for themselves and fail, protecting the "targetted" admin for 9 months in the process. Am I overlooking any of your concerns, Anomo? I just want to make sure I'm not missing anything you're saying. Karwynn (talk) 15:29, 14 August 2006 (UTC)

What Karwynn said. I really tried hard to make this idea of a system 1) as simple as possible; 2) as user-friendly as possible--it's literally just a few steps from end to end, but each step is increasingly difficult to get passage through; 3) as hard to abuse as possible. The initial version was any 25 users plus certifiers, with 3 months/300 edits each. Now its 1 file, 30 users (3 of them admins), with 10 months/2000 edits each. Basically, if you're here 10 months and have 2000 edits odds are you know a thing or ten about wikipedia, and the requirements as laid out make ambush recalls impossible--admin HAS to be notified. He doesn't have to accept or confirm, or he'd just never do that. Then he gets a WEEK to address whether or not there should be a recall RfA. Then he gets THREE weeks in the current revision to display and explain why he should remain, during RfA. Fairness across the board. rootology (T) 15:33, 14 August 2006 (UTC)

Complexity concerns--why? Please read.

People keeping saying the process is too complex as written for a reason to not support it, and I'd like to address your concerns. What is too complex? Short version:

• 31 people (3 of them admins) with at least 10 months experience and 2000 edits (i.e. veterans with assumed knowledge of WP) agree by signing on that the admin should be recalled, over one week.
• Recall RfA--same as an RfA, but to desysop.
• If not Recalled, admin is immune to this process for nine months to discourage frivilous filings.

Thats basically "it". What is too complex? rootology (T) 18:15, 14 August 2006 (UTC)

Well, maybe some of the users are perplexed by the rules of this proposed policy. Anyway, I don't find anything complex in this. --Siva1979^{Talk to me} 18:28, 14 August 2006 (UTC)

My concerns are not over complexity, they're over the chill this would put on administrators who make controversial decisions. Fagstein 18:38, 14 August 2006 (UTC)

Perhaps some of them need a chill? --badlydrawnjeff talk 18:39, 14 August 2006 (UTC)

Why? Why should admins worry more about being popular than doing their jobs? Fagstein 19:35, 15 August 2006 (UTC)

Thats the thing--admins with the support of their peers even in something like this would have really nothing to worry about it. Especially with the addition of the 3-admin support requirement, it's pretty much you need to be removed by a vast, vast majority of your peers for it to ever come about. The point basically is just to empower the community to fully oversee itself. rootology (T) 19:50, 14 August 2006 (UTC)

I don't see where this "vast, vast majority" comes from. You need 31 users. After that happens, unless at least eight users come to your rescue, you lose your privileges. Fagstein 19:35, 15 August 2006 (UTC)

Admin certifiers

Three admin certifiers isn't nearly enough. It should be something like fifteen. Off the top of my head I can easily think of three whacko admins who I trust less than most regular users who would be incredibly happy to leap onto any such call for admin recall. --Cyde Weys 19:58, 14 August 2006 (UTC)

Would you be able to comfortably support such a process and vouch for it with additional certification criteria similar to this? Remember, my goal really is to introduce a fair other side to the one-way street that RfA is for the community to be able to exercise. rootology (T) 20:00, 14 August 2006 (UTC)

Cyde, would 7-10 admins make you comfortable? rootology (T) 20:10, 14 August 2006 (UTC)

Keep in mind that iff you assume bad faith on the part of the certifiers like this, then it could just as easily be 15 "whacko" admins that could sign. There has to be a reasonable limitation. This policy shouldn't be formulated based on the fact that admins are outright not to be trusted. Also keep in mind that any bogus recall would get the smackdown when it came down to the vote. Karwynn (talk) 20:02, 14 August 2006 (UTC)

...let's keep it on topic a bit moreso if we can. rootology (T) 20:10, 14 August 2006 (UTC)

Also, would regular users' "votes" get the same weight in the re-RFA? That doesn't particularly make sense to me. If sixty trolls show up wanting an admin recalled and sixty admins stand up for him saying he's doing the right thing, that's technically only 50%, but we'd be insane to take away the sysop bit at that point. --Cyde Weys 20:04, 14 August 2006 (UTC)

Do we discount trolls, cliques, or cabals from current RfA? The certification process is the measure designed to weed out the crap from the good filings. rootology (T) 20:10, 14 August 2006 (UTC)

Of course they would, just like on normal RfAs. THe "danger" is the same. If Wikipedia policy were formed based on your idea of the threat posed by trolls, we'd never get anything done. Karwynn (talk) 20:07, 14 August 2006 (UTC)

Actually if we gave trolls no attention just like they deserve we'd get a lot more done. --Cyde Weys 20:13, 14 August 2006 (UTC)

Ok, so how do you distinguish between processes that would be taken advantage of by trolls (such as this proposal) and existing processes that seem to do fine (AfD, RfA, RfC, etc.)? Karwynn (talk) 20:14, 14 August 2006 (UTC)

RFC is extensively used and taken advantage of by trolls. Stop kidding around. --Cyde Weys 20:16, 14 August 2006 (UTC)

Cyde is right, RfC as it's formed is a joke for anything beyond to document legitimate issues historically. But he's right: thats abused. rootology (T) 20:18, 14 August 2006 (UTC)

No one's kidding around, Cyde. FInd a more civil way to express yourself. And what about RfA and AfD then? Karwynn (talk) 20:20, 14 August 2006 (UTC)

Voting is evil. Kelly Martin (talk) 20:06, 14 August 2006 (UTC)

RfA *is* a vote. :) rootology (T) 20:18, 14 August 2006 (UTC)

If voting is evil, what alternative system do you propose for ArbCom to make decisions? Tyrenius 01:08, 15 August 2006 (UTC)

Statute of limitations

Another proposed amendment: all of the bad stuff brought up about the admin in the recall reasoning would have to be from the past month. This is to prevent people from trolling through years worth of admin contributions and cherry-picking out all of the "bad". Also, this means that if someone sees a serious admin infringement they will need to get on it right away; if they actually bother to wait more than a month they're either trying to manipulate something or they didn't think it was that important, it which case it shouldn't be eligible for a recall. --Cyde Weys 20:15, 14 August 2006 (UTC)

What about in a case where an admin (or several) wish to demonstrate that one of their peers has a history of x, y, or z? The thing that might be bad about this--what if some admin were to take an unsuccessful Recall as a license to do x, y, or z ongoing? The principle of this is Good, and I like this idea. How would you word the change? rootology (T) 20:18, 14 August 2006 (UTC)

Huh, I had thought that in a previous draft it was supposed to be about a single issue or type of behavior, not just trolling for anything worth noting (unless it fits in a category of behavior, like issuing bad blocks). -- nae'blis 20:19, 14 August 2006 (UTC)

A previous version might have had the wording like, "Evidence/basis is for xyz issue, but you can cite additional supplemental evidence". Which is not a bad thing--people do that in RfA now all the time, as in "Such and such did This, which is Bad, and has a history of blah blah..." rootology (T) 20:22, 14 August 2006 (UTC)

That seems rather counterintuitive, as I would think that we would want to avoid making spur-of-the-moment decisions based on isolated incidents, and instead reserve something like this for cases where an administrator might have an extensive history of unacceptable behavior. Kirill Lokshin 20:19, 14 August 2006 (UTC)

Spur of the moment is bad--people can and will say stupid things on talk pages like "I'LL RECALL YOU OVER LOL blah blah," but unless they file its no more meaningful than saying "UR BLOCKZ SUCK". rootology (T) 20:21, 14 August 2006 (UTC)

Exactly, RfAR cases about admins and non-admins are about patterns of behavior. Desysopping is a serious step, and should not be based on a month's worth of transgressions. Because, sometimes concerns ARE brought up immediately, but are ignored until a later time. Karwynn (talk) 20:22, 14 August 2006 (UTC)

Stupid processlawyering comment, then: If an admin confines their behavior to small, low-impact "badness" that is spaced out to every 32 days, would they be immune to this process under Cyde's amendment? -- nae'blis 20:23, 14 August 2006 (UTC)

Yes, but that's not very likely. Karwynn (talk) 20:25, 14 August 2006 (UTC)

I think the idea that the basis of the dispute has to be for something recent has a kernel of a very good idea there, that can be hammered out to make this whole process fly and be implementable. What would be a bad thing is any sort of "immunity" per se, because that's not the point--people can and are historically judged by their actions. No one should be Recalled for a crap action say in 2005, since they may have been ace since then. Perhaps a modification of: 1 filer, 30 certifiers, 10 of which are admins. Basis of dispute/reason for recall is based on actions in past 90 days (based on diffs/actions). Evidence section may include previous diffs/history to establish patterns, behavior, etc. Keep the nine month time limitation. If someone really, really pushes the admin envelope beyond that, especially after being perhaps unsuccessfully recalled, you can bet then that someone WILL haul them up before ArbCom. rootology (T) 20:28, 14 August 2006 (UTC)

The less conditions the better. Let an admin's whole record be available. The real problem is not an uncharacteristic moment of bad decision-making, but a pattern of long-term abuse which many users have been at the receiving end of (something I trust which does not apply to any of our present admins). Obviously if something from a year previously is cited and nothing of the kind has been repeated, then this is likely to be discounted, just as it would be in a conventional RfA. Let's base things on the existing model, and have some faith in the community's common sense. Tyrenius 21:15, 14 August 2006 (UTC)

After the Sean Black re-RFA I am seriously lacking faith in the communal common sense. --Cyde Weys 21:21, 14 August 2006 (UTC)

Why, because he was re-sysopped? Tyrenius 05:54, 15 August 2006 (UTC)

For better or worse, anything below Foundation level is the will of the people though. If his RfA was a bad idea, and there is cause/grounds to show it, this would be a good avenue for that to Recall. rootology (T) 21:24, 14 August 2006 (UTC)

For what it's worth, Sean Black would have gone all the way through the recall process (as the proposal is currently formulated), but wouldn't have been recalled in the end. Something like 14 admins voted oppose, and more than 31 total users voted oppose who had 2000/10months. I think that admin accountability is rather important, and I think that admins should be respectful, but I don't think that recall is the right way to encourage respect/civility by admins. If recall is formulated to address anything beyond specific actions that obviously hurt the encyclopedia, then recall would encourage admins to think about politics on a day-to-day basis, and that wouldn't really be productive for the encyclopedia (janitors should spend their time doing janitorial work, not trying to look good). --Interiot 03:46, 15 August 2006 (UTC)

• Cyde, does it occur to you that the "community's common sense" is what gave you the sysop rights? How can you trust them on the one hand, but not on the other? - brenneman ^{L} 05:38, 15 August 2006 (UTC)

Statute of Limitation, my revision

Posting here as I asked Cyde on his talk page for input but no answer as of yet...

"This idea has me intrigued. Do you think RECALL would get needed support to become policy if there was such a thing that said "basis of dispute/primary evidence" must be from AFTER the time the policy is implemented? I.e., say you (hypothetically) did something vile and probably worthy of Recall x days, weeks, months, or years ago. What if the policy specifically said under Limitations "Diffs as evidence for basis of dispute must be dated after 10/01/06, the date this policy became effective. Older supplemental material/background material may be older.""

Thoughts? rootology (T) 22:04, 14 August 2006 (UTC)

I don't think Recall will work because it will be repealed at the Foundation level. As soon as this goes through you just know someone is going to file it on Tony Sidaway or Kelly Martin or one of the other "cabalists". They may have a warped perception in the community, but there's so much work they do behind the scenes to help the Foundation it's ridiculous. I think, from that point of view, the recall thing would be a failure. There's too much going on that people don't know about and they wouldn't be able to accurately judge whether someone is actually a good admin or not. It'd just turn into a "I don't like you" or a "You're too incivil" kind of thing. --Cyde Weys 22:15, 14 August 2006 (UTC)

...and you've just revealed a major component of the problem. That "some" admins are "better" than others, at least in the perceptions of the Cabal. I was practically waiting for someone to say such a thing, and your response of this nature to my suggestion of basically giving every admin a free pass vs. the idea of Recall against any past activity is telling--I raised the ante in a fashion of which the policy will basically only apply to anything you do against policy going forward, and your response in turn is that "Oh, Jimbo will crush this, my friends are too special." Thanks for not helping too terribly much, and cementing many people's perceptions that certain admins would never do anything to put their lifetime appointments in the hands of "lessers". rootology (T) 22:20, 14 August 2006 (UTC)

Cyde, look at how many people vote in RfAs. Several have never interacted much with the candidate, but support anyway. If a ridiculous Recall is filed, will it not be voted down? Getting through 5 admins and 25 other certifiers and then getting a supermajority is not going to be an easy process. People will leap to the defense of TOny, Kelly or many other respected admins. This is not a "only trolls can vote" thing - everyone will have their input. As far as your general attitude, I think rootology is right, you seem to be opposed not to this process, but to greater accountability for administrators. Karwynn (talk) 22:27, 14 August 2006 (UTC)

Side comment: Haveing the policy only apply to acts after implementation might not be an bad idea if this was a code of conduct, since this is merely a process for cutting back on abuse, I don't think that owuld be necessary. Karwynn (talk) 22:29, 14 August 2006 (UTC)

Let me get this straight: over 50% of the "voters" would need to vote "I do not support this user as an admin". So it wouldn't the 75% thing we have with RFA currently, correct? --Cyde Weys 22:36, 14 August 2006 (UTC)

No, it's spefically an 80%+ concensus. I.e., out of 100 "votes", on the Recall itself, a desysop will not happen unless its 80-20-x or higher. No ifs, ands, or buts. It's not meant to be easy to unseat an admin by any means under this policy, and the certification is totally seperate from the RfA by design, to insulate admins from bogus filings... which I keep going over again and again. So, basically, if all 31 certifiers vote "Recall", all that has to happen to squash the whole thing is 20+ people to vote "keep", over the course of TWENTY ONE days. rootology (T) 23:05, 14 August 2006 (UTC)

It doesn't look like the numbers have been worked out yet. (I personally would go for 2/3, or 70% to round it off.) Karwynn (talk) 22:41, 14 August 2006 (UTC)

Just curious, would you agree that people are more likely to vote "Opose RfA" than "Support recall", making it a little tougher to get sysop revoked once given? Karwynn (talk) 22:42, 14 August 2006 (UTC)

More

There's a delay of 6 months for filing these and 9 months in between. I think a year statute is fair, although still short I think. Anomo 22:58, 14 August 2006 (UTC)

I don't want to call it this, because that's the height of Bad Faith, but my suggestion of effectively giving what some might call "Amnesty" for previous actions as the initial basis of dispute/Recall, plus the time limitations, plus either 5 or more admins signing on to certify the Recall, is basically the height of fair play--to be honest, based on ALL of these conditions, an admin would have a better chance with a Recall RfA than they would dealing with ArbCom if they actually did something wrong. Literally, the whole point is to just give users/editors/admins the ability to undo with cause an RfA they feel should not be appropriate anymore. rootology (T) 23:31, 14 August 2006 (UTC)

I just thought up the ultimate in fairness. There's a sort of requirement like 2000 edits and 10 months for people to file this and these numbers will probably change. Maybe the statute of limitations could match the edit requirement. 10 months and 2000 edits would mean the previous 10 months or 2000 edits (whichever is greater) are however far the statute of limitations extends. Anomo 07:10, 15 August 2006 (UTC)

A limitation on the frequency of filing these and a different limitation on how recent the misbehaviour can be don't go well together. Otherwise surviving a recall becomes a free pass for a period of time. If both limits are to be present, they should be the same - so only for behavior since the last recall discussion would be a reasonable limit. Frankly though, such a statute of limitation is unneeded rules expansion. The community will laugh off old items unless they also represent an ongoign trend in a administrator's behavior. GRBerry 14:30, 15 August 2006 (UTC)

Proposed change based on Cyde's idea

• 1 filer, 30 certifiers, 10 of which are admins (10 month/2000 edit requirement remains)
• Basis of dispute/reason for recall is based on actions in past 60 or 90 days (based on diffs/time stamps).
• Evidence section may include previous diffs/history to establish patterns, behavior, etc.
• Keep the nine month ime limitation to eliminate harassment.
• 24 hour notice remains, but if unfiled is just so much hot air, and the named admin can just delete it of course. —The preceding unsigned comment was added by rootology (talk • contribs) .

---

Well, first let's get it accepted, then worry about the details. But (aside from 30 certifiers being maybe a bit high), the ten admins requirement is a red flag. Naturally admins watch each other's backs, that's only human nature. Plus obviously it's not in the interest of any admin to certify. Have you never heard of Omerta or the Blue Wall. They exist for a reason: mutual self-protection. Besides that, admins are a social group, always chattering away on their private IRC channel and what have you. You certify, you're going to be running into the guy at the next admin cookout or whatever. No, if you must keep raising the bar, how about X number of certifiers must have some given higher edit count/longer service than a regular certifier. Not that I'm suggesting that, but if you must. Herostratus 20:41, 14 August 2006 (UTC)

It'll never happen. An admin would have to go completely bonkers for this to be fulfilled, bad enough that an ArbCom case would actually be more likely to close first. That many admins would never certify. Karwynn (talk) 20:42, 14 August 2006 (UTC)

Hero, what do you think would be a good and fair threshold in your opinion? Keeping in mind that this policy is intended for admin use as well as regular editors? rootology (T) 20:48, 14 August 2006 (UTC)

Herostratus, the amount of bad faith above is kind of staggering. You don't think that 1% of admins would be willing to certify on a bona fide recall petition? Cyde above just intimated that 0.3% of admins are whackjobs, which is a different sort of problem, but the point remains...you can't get something approved until at least the basic details are worked out. How many certifiers of each category will be needed is certainly a basic detail, in my world... -- nae'blis 20:51, 14 August 2006 (UTC)

I'm in the dark, where's the bad faith? Karwynn (talk) 20:56, 14 August 2006 (UTC)

There should be no requirement for any admins. The community appoints. If they wish, let the community take away. Experienced users of good standing is all that is required. Tyrenius 21:08, 14 August 2006 (UTC)

The comments about the Blue Wall/Omerta (codes of silence), assumptions about self-serving admin motivations for behavior, and "Naturally admins watch each other's backs, that's only human nature" are what I was referring to, specifically. Anybody who's watched WP:AN/I knows there's vigorous disagreement amongst admins. -- nae'blis 22:00, 14 August 2006 (UTC)

If you can't find 10 admins out of about 400 active admins to endorse an admin recall then maybe the recall isn't supported by the evidence? Any recall of even a semi-active admin is going to easily be able to find thirty users ... if nothing else, thirty users who were previously blocked by that admin or whatever. The reason for admin endorsements is to make sure that it is merited on a level above merely all of the users with grievances coming out to play. --Cyde Weys 21:14, 14 August 2006 (UTC)

So why ten? Why not three? How are 10 admins any more or less likely to be "whackos"? ANd if the recall has little credible evidence, wouldn't it be voted down anyway? Karwynn (talk) 21:18, 14 August 2006 (UTC)

I do like the idea of admin endorsers--it keeps the idiot ones out of the way. But I think ten is too high, and three too low. Five is better. Because, odds are, if five admins decide on a course of action, the other admins will not counter that in good faith. I.e., if 5 decide/agree on a resolution to somehting on AN or ANI, it will likely stand forever. 5? Agreed? rootology (T) 21:21, 14 August 2006 (UTC)

This starts to differentiate admins as a special class, instead of trusting and empowering the community as a whole. I very much doubt that any recall would find 30 users of good standing (a key qualification in this proposal) in 7 days so easily were it not merited. Just because someone has been blocked does not mean they would wish an admin de-sysopped, if they felt general respect for that admin. Besides which, the final result is dependent on the whole community anyway. I think there is a fear of the unknown here, rather than a real need for fear. Tyrenius 21:24, 14 August 2006 (UTC)

The certification would be listed on both RECALL and RFA for the full week of it's shelf life--you don't think it would get enough attention/traffic? I also specifically put nothing in to say "You can't tell people about this..." the same as there is no such limit for RfC, RfA, RfAr, etc., as that would be biased. rootology (T) 21:28, 14 August 2006 (UTC)

Five is a better number. I'd prefer 3, and 25 certfiers too, and a decrease to required edits (why not 1000?) but oh well. Karwynn (talk) 21:34, 14 August 2006 (UTC)

2000 edits is a good value and I think 30 as well. Removing an admin shouldn't be as easy as becoming one, and 2k will also weed out a lot of the stupid stuff. rootology (T) 21:35, 14 August 2006 (UTC)

It's unfair to make such a short statute of limitations of a month or so if there is going to be a 9 month delay to file. Anomo 23:02, 14 August 2006 (UTC)

Time to name names

If the supporters of this proposal are truly serious about making this into policy, there needs to be a discussion about which admins it would have affected. It's great to talk about this proposal in theory and using generalities, but I think it's time for specifics. This proposal is designed to address admins that the ArbCom won't deal with. Which admins (excluding ones already desysoped by the ArbCom) would have been desysoped by this policy? A previous discussion posed this question, but the answer was essentially sidestepped. Is it possible to name 5 admins who would been desysoped by this policy? Aren't I Obscure? 20:45, 14 August 2006 (UTC)

This is an excellent question. I would love to hear some honest answers. --Cyde Weys 20:46, 14 August 2006 (UTC)

Most of my support stems from the fact that I've seen a general lack of accountability for administrators, both in practice in attitude. There have been specific cases that have been particularly frustrating for me and, I believe, disruptive to Wikipedia. Of course I'm not going to name names, there's no reason for me to insult administrators like that. Karwynn (talk) 20:54, 14 August 2006 (UTC)

But I'm looking for someone to name names! There's enough generalities on this page to choke a horse, but virtually no specifics about who it would actually affect. I want someone to show the case of a specific admin who would have been desysoped by this policy, but the ArbCom never acted. I'm not looking for an attack on this admin, just a rundown of why he/she should be desysoped and how this policy would suceed where ArbCom failed. Can we start with just one example? Aren't I Obscure? 21:03, 14 August 2006 (UTC)

Look at what I wrote below. As for the RfA side of things that's an unanswerable question--might as well ask what would happen if I ran for RfA in a year. rootology (T) 21:10, 14 August 2006 (UTC)

Thing is, we can't say they would have been because we don't know what the votes would've been after 3 weeks. Additionally, you won't be likely to find anyone willing to name names, being as ther is often a fear of massive retaliation from said admin(s), as is my belief for my case(s). Besides, it's the general lack of accountability that drives me personally, not so much individual admins. Karwynn (talk) 21:07, 14 August 2006 (UTC)

From just very recently, the situation that was just in WP:AN (you know who) could have factored into this. Looking at some cases, this could have been community resolved faster; this one took three months; this one took three months. There was also the Pedophilia Userbox mess. The following links to previous issues all likely would have passed at least the certification phase for a community review were they processed via Recall: [4], [5], [6], at least one, [7], [8].

EDIT: Recall with a 5 or 10 admin threshold had the policy existed when these events occurred. As to whether the community would have then desysopped, who can say. But it would have certainly reached RfA in some of these. rootology (T) 21:01, 14 August 2006 (UTC)

Those are all bad examples because they refer to cases that ArbCom did deal satisfactorily with. We're asking you to name people who you think this proposed policy would've taken care of who haven't yet been taken care of by ArbCom. Also, I dispute that this policy would've taken care of Guanaco ... he was very popular at the time, because he was one of the pro-userboxers. I'd much rather have the ArbCom dealing with these; they look at actual abuse, rather than merely turning it into a popularity contest. --Cyde Weys 22:12, 14 August 2006 (UTC)

Perhaps, in some cases, the community could have dealt with the situation faster than arbcom. Consider HOTR, for example. I honestly don't follow the Israel stuff so I don't know all the ins and outs of the personalities there ... but I've got to think that the community would have long ago supported a de-sysoping there whereas, if I am reading this right, ArbCom has decided to do nothing. BigDT 22:37, 14 August 2006 (UTC)

For names, go to a forum like wikipediareview or some message board about wikipedia--even usenet and see who they complain about. It will be them. On an unrelated note, I am wondering if administrators could be able to vote anonymously so as to avoid retaliation. If I was an admin and I see an admin who say indef blocks anyone who forgets to sign their name and locks their talk page immediately, I'd want them gone, but I wouldn't want retaliation for it from their friends. Anomo 23:11, 14 August 2006 (UTC)

Wikipedia Review complains about anyone who hasn't sieg heiled at least twice before they pour their Cheerios in the morning. Anyway, I thought this whole thing was about accountability, but you want the people who sign these petitions to be unaccountable? In my view that isn't consistent. --Sam Blanning^(talk) 23:22, 14 August 2006 (UTC)

The whole point of this--I don't know if some admins think I'm going for this--is NOT to crucify people who may or may not be "annoying" to their peers. It's simply to provide admins and users together the ability to undo a previously successful RfA which they feel, with cause, to be no longer appropriate to keep in force. Those Brandt sites are a poor example. rootology (T) 23:29, 14 August 2006 (UTC)

Wikipedia review is just easiest to find. There are many complainer places, but they're scattered. I don't know whose complaining achieves notability. [www.freerepublic.com/focus/f-chat/1549132/posts Here's another] Anomo 23:35, 14 August 2006 (UTC)

Offtopic, why do people call rogue admins now rouge (means red) instead of rogue? I understand there was a joke once, but beyond that maybe it's a learned spelling mistake. Anomo 23:35, 14 August 2006 (UTC)

Anomo, who has been indefinitely blocking people for not signing their name? Have you heard of the Arbitration Committee? I think they'd take a dim view of such silliness. --Tony Sidaway 23:36, 14 August 2006 (UTC)

I expect that someone frequently protesting the fact that their MPOV was being suppressed decried all and sundry as 'rouge admins', it stuck in JzG's mind when he wrote WP:ROUGE, and since then it's been the only acceptable spelling. From playing roguelikes a lot I can tell you that it's one of the most frequent misspellings in English. Who and where made it here, I don't know (if WP:ROUGE gave an etymology it would spoil the joke). --Sam Blanning^(talk) 23:40, 14 August 2006 (UTC)

I'm happy to state whom I believe would be the first on the chopping block if some form of community recall were made policy: Tony, Cyde, Kelly. All three would fail re-confirmation as well. MONGO and Zoe might be under some threat, but would probably squeak through. Looking over Wikipedia:List of administrators I don't see anyone else that would need to worry. - brenneman ^{L} 23:47, 14 August 2006 (UTC)

Aaron is correct, of course. And the only reason they would be on the chopping block is for annoying certain groups of people, which I don't consider a powerful argument for desysopping. All admin decisions are reversible and it takes a hell of a lot of effort to actually be a bad admin. {{subst:/And}} --Sam Blanning^(talk) 23:51, 14 August 2006 (UTC)

P.S. An outside view on Kelly's current RfC calling for her desysopping has a grand total of zero endorsements, indicating that people aren't particularly willing to actually put their money where their mouth is. --Sam Blanning^(talk) 23:53, 14 August 2006 (UTC)

For what it's worth, I would not certify KM based on what spawned her RfC2. rootology (T) 23:54, 14 August 2006 (UTC)

I think that the lack of certification has more to do with the reality of the current dead-minning process than on community consensus. If we think if adminship in terms of giving rather than taking away do we have any real expectation that any of these individuals would pass a request for adminship that happened now? - Aaron Brenneman 23:57, 14 August 2006 (UTC)

Aaron, please present any case for desysopping any of the three you have named to the arbitration committee. Your tireless and unceasing low-level personal attacks on administrators, which span almost the entire tenure of your time on Wikipedia, have not gone unnoticed. --Tony Sidaway 00:06, 15 August 2006 (UTC)

You say low-level personal attacks, I say healthy scepticism. Let's call the whole thing off until one of you wants to actually take it to a forum where people care. This page is about some paddling skeeball process or something. --Sam Blanning^(talk) 00:16, 15 August 2006 (UTC)

Err, did I actually do anything wrong there? - Aaron Brenneman 00:23, 15 August 2006 (UTC)

Nothing at all, but you haven't completed the job. File your arbitration case. --Tony Sidaway 11:56, 15 August 2006 (UTC)

Note that Aaron didn't even say he would certify them, just that he thought those three would be first on the block. I'd have to say he's probably right, which is not to say that it would be a correct decision. -- nae'blis 02:29, 15 August 2006 (UTC)

Kelly Martin has some RFC and complaints and I've read through them and can't figure it out. It's more complicated than that whitewater thing. I just can't figure out what Kelly did wrong. Well, basically I hear one name, starting with an S, mentioned here on every single forum with people saying she's the head of the Cabal, worst admin, blah blah. I've not actually seen S do anything myself, but I see complaints like mad. The complaints I see are less than 20 admins actually. Anomo 00:04, 15 August 2006 (UTC)

What about the admins that are inactive or barely active and like this for a year? I generally think these admins should be desysoped. I read once of someone wanting to de-power the inactive bureaucrats. Anomo 00:11, 15 August 2006 (UTC)

I'm pretty sure they did that on wikibooks a while back, or maybe they didn't (debates tend to take m-u-c-h longer over there). --SB_Johnny | ^talk 00:14, 15 August 2006 (UTC)

For names, go to a forum like wikipediareview or some message board about wikipedia - if you really believe in wikipediareview, which is harrassing people in real life because of what happens in Wikipedia as reliable sources, then that's all that needs to be said about this whole proposal. User:Zoe|(talk) 02:50, 15 August 2006 (UTC)

Ick. Just had a look at that forum... they sure are angry! I wonder if Bobby Boulders monitors that. --SB_Johnny | ^talk 11:22, 15 August 2006 (UTC)

Is this trip really necessary, rootology?

This all seems a bit over-the-top... may I say "WP:CREEPY", if for no other reason than I find it an amusing thing to say?

If the goal is to have admins held accountable to community support, why not just throw out the lifetime appointment thing, and just repeat the RfAs once a year, or maybe on a reduced schedule over time (after 1 year, then 2 years later, then 4, etc.). At least this would be a more "neutral" approach, since good admins would be able to use the opportunity to recieve much-deserved wikilove without asking for it (and hence looking desparate), and bad admins could recieve a well-deserved slapping around without some poor schmo having to open an enormously complicated recall request.

What bothers me about this is that it seems to assume bad faith on ArbCom's part, and no matter how many safegaurds you might add by manipulating numbers and voting rules, these recalls are always going to be proposed in the heat of anger... and sometimes an admin is going to make 30 (or 50, or 1,000) wikipedians angry by doing what they're supposed to do as an admin. A user who's having troubles with an admin need only follow the steps in WP:DR. If virtual handshakes can't be virtually exchanged by the time it gets to ArbCom, at least the user will have had time to cool his or her head, get the arguments together, and present the case without using too many exclamation points (implied or otherwise). --SB_Johnny | ^talk 23:43, 14 August 2006 (UTC)

It's got nothing to do with lack of faith in ArbCom, and I went out of the way to NOT make it sound like that--it's literally to give people the ability to fairly undo a previous RfA if the community feels such an action is warranted. As much as your "rolling and extending" votes sound wonderful, based on responses here (and from previously reviewing the annual call of a different method for the community to "undo" adminship if warranted) I highly doubt the admin body politic will ever come together to agree on anything but ArbCom actually "taking" someone's tools away. I'm not saying that's a bad thing; my goal here was to find a way for the community and admins to police themselves effectively and fairly. rootology (T) 23:57, 14 August 2006 (UTC)

Well, the angry-editor-in-a-snit problem is a pretty serious one, I think... I really think the time and effort involved in following WP:DR provides some time for cooling off, getting opinions, etc. Maybe have this be the step after arbcom, if arbcom either refuses to hear the case, or if their decision isn't acceptable to a large number of users and admins. --SB_Johnny | ^talk 00:05, 15 August 2006 (UTC)

The problem I'm beginning to suspect is that the majority of users for whatever reason are unwilling to put a reverse RfA in the hands of the community--which is odd, as they have no problem with RfA itself as it is, and has been for years. rootology (T) 00:10, 15 August 2006 (UTC)

Not odd at all. When you block people for violations they get upset, even when they clearly violated policies. Enforce policy enough and you will have your own hate fan club. Recall smothers the ability of admins to do the work they need to do, even though the people who brak policy don't like it. I'm sorry, but the reverse RFA is already in the hands of the representatives of the community, arbcom. There is no need to let the angry mob of people you blocked decide your fate. If there was so much admin abuse going on, arbcom would have removed a lot more people by now. But they haven't. pschemp | talk 00:25, 15 August 2006 (UTC)

(Striking part of my previous comment) I think the RfA is in the hands of the community because that's just the easiest way to do it. If the Foundation people had to go looking at every user's contributions to see who would be a good candidate for unpaid work, they'd have no time for anything else. Wikimedia has never been a democracy, and probably won't be in the forseeable future. It uses democratic devices, and has an egalitarian mission, but it's not really a democracy. Even ArbCom is trumpable by Jimbo, and I'm pretty sure this policy could never have teeth unless Jimbo were to declare: "Let it be toothy!" --SB_Johnny | ^talk 00:39, 15 August 2006 (UTC)

As an aside, I find the idea of administrative terms lengths a lot less problematic than introducing a recall system simply because it does seem rather more fair, less bureaucratic, and because it simultaneously reduces the aura of elitism that is the biggest perceptual problem with the current admin system. I think the problem is the difference between explicit, quantifiable violation of specific policies, and a more general lack of respect and regard for non-administrators that skirts the boundaries of policy. Upon appointment, it's expected that administrators will reflect the most trusted members of the Wikipedia community, but there is absolutely nothing in place to ensure that they continue to do so. ArbCom works perfectly well in dealing with concrete violations of established policy, but to suggest that all problems concerning administrators can be boiled down to such relatively clear-cut situations smacks of the sort of bureaucracy that we frown upon. It's not about democracy, or egalitarism, or any other kind of ideology you'd care to name: it's about ensuring that the collaboration that makes Wikipedia work isn't buried under the weight of mutually exclusive cliques and suspicion between self-imposed castes of users. Not that I have an easy solution to the problem, mind you, but I've seen enough that I cannot doubt that the problem exists. – Sean Daugherty (talk) 04:52, 15 August 2006 (UTC)

I agree that something needs to be done; it's ironic to me that I never found myself questioning the judgement of the community or any administrators until I became an admin myself. Personally, I think that this proposal would be hard to be gamed as others fear (I have just read this whole talk page) and I don't really think a lynch mob or the had-a-bad-day-editor is a serious threat -- if you are an experienced contributor, as this calls for, it would be nigh impossible to get 25-30 such editors to agree to "indict" an admin unless there is something seriously wrong. Thus, I cautiously support this proposal. I agree with all of what SD says above. Grandmasterka 08:07, 15 August 2006 (UTC)

(unstriking my comment above, because I'm sometimes fickle that way) I still don't think this will fly, but it's worth considering some more. The thing I'm unsure of is where it would fit into the DR scheme in general. Should it happen before mediation? Before an RfC? Before ArbCom? After ArbCom?

If the latter, then what we're really talking about here is a method for petitioning Jimbo or another foundation member, right? If that's the case, then why not scrap all the "CREEPy" rules, and just set up a template and a vote counting technique (the number of votes and/or the number of days that the Admin Recall Petition (ARP) stays open before it's considered submissible should do the trick). Make it clear and brief and have voters disclose their edit counts (with a link to interiot's tool or something similar). Easy, huh? --SB_Johnny | ^talk 11:18, 15 August 2006 (UTC)

A modest proposal

What the heck. OK, admins can't be recalled, fine. Why not take the obvious next step. Herostratus 07:07, 15 August 2006 (UTC)

Based on the responses here and the utter disdain many have for any extra accountability placed on admins, I expect this will likely never come to pass unless the Foundation level decrees it, at which point no one will have a leg to complain on. For what its worth, if I were an admin I would 100% support the idea of community recall. rootology (T) 07:09, 15 August 2006 (UTC)

Don't be such nancies! Grow some stones, keep working on the proposal, and chin up. Part of the problem is that this was (no offence) terribly written and very little {tl|sofixit}} appears to have taken place. I'm going to do some brutal re-writing. - brenneman ^{L} 07:48, 15 August 2006 (UTC)

Thanks... I'll take another look at it in the morning. rootology (T) 07:49, 15 August 2006 (UTC

Um, hi, excuse me, this section is taken. Get yer own section :/ You're burying my proposed new policy Herostratus

Previously, on Wikipedia talk:Admin recall

This is an attempt at summarising the discussion at Archive 2. Please correct any errors I may have made. - brenneman ^{L} 12:33, 15 August 2006 (UTC)

Well, since this is all now in both places, it's odd but please do still add to the summary to help avoid repeating arguments ad naseum. - brenneman ^{L} 14:08, 15 August 2006 (UTC)

Straw poll

• Some editors said "this is good."
• Several editors like the concept, but opposed the execution.
• Some editors oposed the concept.
• Several editors expressed that it was far too early for a straw poll.

Overall there was nothing like consensus that it should go forward in the form it had at that time.

Objections

• Several editors felt it was open to abuse.
  • Admins make enemies by doing their jobs.
  • Philosophical factionalisation would come into play.
• Some ediotrs felt it was not required.
  • Arbcom can already dead-minn.
  • There isn't a problem with "bad" admins.
• It was pointed out that the initial draft had some basic mistakes regarding existing policy.
  • What b'cats do vs what stewards do.
  • What the Board's function is.

Mechanics

• Lots of discussions about what is the appropiate threshold for "good standing."
  • How many edits: 500, 2000, what about 500 mainspace, etc.
  • Should there be "jury of peers" in: An admin quorum, board, etc.
• Strong feeling that the process was too complex.
• Concerns about the "statute of limitations" being open to abuse by "protective filing."

Re-write

I've hacked and slashed my way though the existing proposal as much as possible without actauly changing anything. Following this and a rest and some food, I'm going to start to summarise the objections/comments from above. I'll also probably merge Dmcdevit/Proposal for desysopping into this, but that's going to be a serious change of intent ans well as language. Lucky this is a wiki, eh? - brenneman ^{L} 08:20, 15 August 2006 (UTC)

My personal view is that the 80 percent threshold is very high, and necessarily high, to prevent abusive attempts at removing admins on popularity grounds - and I honestly don't think that, in practice, it would ever be reachable except in cases of egregious admin abuse. But in such cases, I think the ArbCom would act far faster to stop those abuses. Thus, this proposal is, IMO, unnecessary. However, I think something like this might improve the perception of adminstrators and remove an excuse to hate the evil rouge admin cabal. Of course, the problem is that given the high threshold, it would be just as easy to say "this is just a whitewash, it's still impossible to remove the cabal." Damned if you do, damned if you don't? I also believe any adminship review process should be required to cite specific alleged policy violations - not just "I think this admin is doing a poor job," but "This admin uses his blocking tool abusively here [1] and here [2]." FCYTravis 09:11, 15 August 2006 (UTC)

I actually proposed to even before this to be able to make a list of such abuses and not have the page deleted or the people who ad to it banned. Anomo 09:06, 15 August 2006 (UTC)

There's a line between alleging abuses in a professional and dispassionate manner and screaming "THIS ADMIN SUX LOSER ROUGE SEE HERE [1] HE BANNINATED ME THE BASTARD [2]." One is fair comment, the other a lame personal attack. FCYTravis 09:12, 15 August 2006 (UTC)

I meant the professional and dispassionate manner. Anomo 09:13, 15 August 2006 (UTC)

Right - and such discussions are already extant in RFCs, thus I don't see any particular danger of having said pages deleted or contributors banned, as long as the discussions are kept civil. FCYTravis 09:19, 15 August 2006 (UTC)

Actually that's all covered and I cleared it up by adding these couple lines to Filing section: "Basis of Recall will cite the reasons why the admin is no longer/should no longer be fit to continue on in their duties. Provide a summary. Evidence will be a detailing of links, and diffs detailing the basis of your claim." If this comes to pass we can always just have "closed" certifications and RfAs protected to keep people from messing with them afterwards. rootology (T) 14:58, 15 August 2006 (UTC)

That fails to state that there must be a policy violation - "reasons" aren't enough. Otherwise we're going to have people run through this gauntlet for the henious crime of validly deleting someone's vanicruftspamvertisement page, even if said deletion was entirely within process and based on WP:CSD. This must be about admins who violate policy, not admins who have pissed someone off. FCYTravis 17:03, 15 August 2006 (UTC)

THose recalls will fail, so no harm done. THe admin in question will also be protected for 9 months. Detrimental to the filer's objectives. Karwynn (talk) 17:07, 15 August 2006 (UTC)

White-washing

Okay, now I'm really concerned. I and a few other people took the time to bring up some valid concerns and suggestions, and now everything is "archived"? C'mon. This doesn't feel right. I don't think the way this works is that you're allowed to delete all of the comments until you achieve the outcome you desired. That's shady. --Cyde Weys 13:11, 15 August 2006 (UTC)

Hrmmm ... inbetween the time that I hit the "+" button and when I hit the "Save page" button it looks like the majority of this page was restored. Thanks. --Cyde Weys 13:16, 15 August 2006 (UTC)

Yeah, but the histories are still destroyed. NOt only is it all "archived", it's actually moved. YOu have Aaron Brenneman to thank for that, who felt it acceptable to come in to a few-days-old discussion, MOVE rather than archive, and announce that the discussion was too much, taking the added step of summarizing everyone's thoughts into the fragmented sentences that obviously represent everything our simple little minds came up with. THen when I tried to undo it by nominating this page for speedy deletion so as to move the page back, he expressed his disagreement on the page deleted the tag without notice. This is not cool. Karwynn (talk) 13:22, 15 August 2006 (UTC)

Aww, you have got to be kidding me. The page was 178K of repative discussion, and there was a big bloody note saying "please do correct any errors in summarising". Can you really honestly say that the current page lends itself more to cogent discussion than a summarised version? - brenneman ^{L} 13:21, 15 August 2006 (UTC)

THat is not your place. If you're not interested in discussion, don't disrupt the discussion for your own stylistic preferences. THe people actively discussing can decide when to move archive. YOu are being disruptive - no doubt unintentionally, but stop anyway, please. Karwynn (talk) 13:24, 15 August 2006 (UTC)

You're welcome to summarize if you wish, but please don't interrupt a lot of ongoing discussions by doing so. --Cyde Weys 13:27, 15 August 2006 (UTC)

• "Destroying" history by a page move? What? The "ongoing discussions" was mostly the same five things said over and over again. The straw poll was clearly an error, and a large chunk of the discussion was about a version of the proposal that has been changed a great deal. Philosphically ownership of pages extends to ideas too. Taking a long rambling and diffuse argument, interjected with barbs, and distilling it down so that reasonable dialog can continue is a good thing. Or do we love out clever turns of phrase sooooo much that we cannot bear to part with them? - brenneman ^{L} 13:40, 15 August 2006 (UTC)

• I may not agree with what you say but I will defend to the death your right to not have it archived. --Cyde Weys 13:41, 15 August 2006 (UTC)

Look, Aaron, here's what's going to happen: The "basics" are now laid out. NOw the details will all come p, the individualized problems, the rationalization will all come right back out. ANd the discussion wasn't just a bunch of statements - it was interaction to modify the proposal. THat is now gone. YOu haven't cut down on anything. YOu've just set the discussion process back a couple of days. All without a word of discussion to the people participating. Karwynn (talk) 13:44, 15 August 2006 (UTC)

I'd prefer not to draw this needless min-drama out any longer, but a few things need saying:

1. Page moves are how you do an archive.
2. History is not destroyed by a page move.
3. Refactoring is a vital part of keeping a hot discussion going.

I'd also like to point out that the most recent addition to this page repeats again one of the points in the summary, in a new thread at the bottom of the page. Is it really that impossible to see that this impedes discussion, that the signal to noise ratio goes up every time this happens? I will refrain from commenting in this thread futher, except to ask how many K does this page need to be before it gets archived?

brenneman ^{L} 14:05, 15 August 2006 (UTC)

THe histories to this page are now gone. Any links to them are now broken. DOn't put this "mini-drama" baloney on me. YOu're the one who ignored all the active participants and revamped everything to your liking. YOu archive by cut and paste, NOT BY MOVING. That's like, archive 101. YOu're an admin, so you're obviously not clueless. WHy are you disrupting this process? There was active discussion based on solving disagreements. You disrupted that with this unilateral, thoughtless page move. This is not helping this proposal or its formation. This is counterproductive.

Reversion

Commenting on this thread again, I know. The clean-up I made was reverted. complete with spelling mistakes. Karwynn, you appear a little bit excited and more than a little bit agressive in the editing. I'm a one-revert-per-day boy, and I've now done one on the talk and one on the article in like fifteen minutes... I think that's a first for me. You've expressed several things that betray that you're not terribly familiar with actual practices and the impacts of actions (e.g. your comments regarding archiving) so I'd like to ask, as nicely as possible, that you slow down just a little bit? - brenneman ^{L} 14:27, 15 August 2006 (UTC)

Afraid not. You're not going to nuke this proposal's discussion down and expect me to just sit back and take it. You have disrupted this discussion and all you can do is ask me to slow down and just accept it. THat's not something I'm willing to do. Karwynn (talk) 14:42, 15 August 2006 (UTC)

You're both SO getting recalled for violations of WP:CIVIL and WP:AGF. Fagstein 20:14, 15 August 2006 (UTC)

It's kinda weird, really. I'm sure Aaron is acting in good faith and wants only to give this proposal a chance of making it work (and he has done some excellent work on that). Please, Karwynn, don't accuse Aaron of wanting to nuke the idea. That really isn't what he's trying to do. --Tony Sidaway 22:24, 16 August 2006 (UTC)

Tony is right on this--Aaron is trying to help get a workable policy implemented, Karwynn. It's OK--I'm not glued to my original, it was that--the original. The goal is a working, fair policy, with a fair level of control from the community to implement/call for the process (and have a level of oversight). rootology (T) 22:29, 16 August 2006 (UTC)

Problems with this.

While I like the idea of having a way to remove problem administrators, keeping adminship shouldn't be a popularity contest. As an alternative, I think we should base the removal of sysop rights on clear evidence that sysop powers were severely and or repeatedly abused - in other words, without someone breaching our trust, we have no basis to rescend it.

This shouldn't be a form of community punishment for admins that are disliked, so disputes that aren't accompanied by real abuse would need to continue to go through dispute resolution. I don't support any process that leaves desysopping to the opinion of lynchmobs. - Stephanie Daugherty (Triona) - Talk - Comment - 13:54, 15 August 2006 (UTC)

THe process takes a month, lynchmobbing would die down. BUt if an admin doesn't have the trust of the community, why should they be a sysop? Karwynn (talk) 14:00, 15 August 2006 (UTC)

The thing is, we do have a process for removal of sysop rights based on clear evidence that sysop powers were severely and/or repeatedly abused. It's called ArbCom. If it's not working, than it makes far more sense to fix it than to circumvent it by adding another layer of bureaucracy. On the other hand, ArbCom can, and should, deal in absolutes. Obvious, quanitifiable abuses, like wheel warring, for example. But that's really not the only form of administrative abuse. Continuous incivility towards fellow editors, abusiveness, and controversial actions that may not violate the precise letter of Wikipedia policy and practice can be equally damaging to the collaborative spirit of the project. Administrators are supposed to be "known and trusted members of the community." There should be a mechanism for dealing with administrators who have repeatedly violated that trust. ArbCom cannot make that kind of judgment: only the community itself can. Ideally, this process should supplement ArbCom, not replace it: the difference between a recall election and impeachment, if you must draw the inevitable political comparision.

Concerns over "lynchmobs" are a red herring. As Karwynn says, the duration of the process would smooth over any hasty, kneejerk reactions to beneficial but controversial actions, the kind which Mr. Weys (and others) have raised legitimate concerns over. Any complaints which persist, pretty much by their nature, are not the complaints of a lynchmob, but those of a segment of the community with a serious issue over the way a particular administrator has wielded his or her sysop powers. The outlined process is no more succeptable to a lynchmob problem than any other form of consensus gathering on Wikipedia: it could happen, but it's not particularly likely. This isn't my favored solution, mind you: I'd prefer a term limit system that applied to all administrators equally. But the system proposed here is better than what we're currently stuck with, which fuels accusations of elitism and cabalism against all administrators and damages the working environment of the project. – Sean Daugherty (talk) 15:02, 15 August 2006 (UTC)

This isn't a "let's have an alternative to ArbCom" necessarily. This is, in essence, a new concept: if administrators do not have the confidence of the community, they should no longer b administrators. Karwynn (talk) 15:12, 15 August 2006 (UTC)

But why? Why are we turning administrators into politicians? Fagstein 20:18, 15 August 2006 (UTC)

A previous administrator, karmafist, is a politician. At least I read he runs for office. Anomo 20:25, 15 August 2006 (UTC)

It's not really an issue of "turning administrators into politicians." It's a question of ensuring that administrators really represent the most trusted editors of Wikipedia, which is one of the requirements of the position. – Sean Daugherty (talk) 19:17, 16 August 2006 (UTC)

An admin is just an editor with some more buttons. They can delete and undelete pages, protect and unprotect them, and block and unblock users. They can also see deleted items. They tend to be more prominent participants in the community, but editors such Kim Bruning, Gmaxwell and others prove that a valuable and influential editor need not be an admin, and admins may sometimes give up their sysop bits without losing any status.

We've had some pretty atrocious administrators but they don't tend to last long even when we make a mistake by granting them sysop rights twice. User:guanaco was desysopped by arbcom and took over a year to get back his admin bit, but his conduct after he got it back led to his being the only admin promoted twice by the community but desysopped both times by arbitration committee.

On the idea that administrators should retain the support of the community, I think this is evidently the case of all admins in all but the most pathological instances. However the RFA process isn't a good way to measure it because such processes (much like the deletion and deletion review forums) gather groupies who aren't really representative of the community.

I'd like to suggest a lightweight process for desysopping. An editor presents an application to the arbitration committee for an arbitrator to undergo a deselection process. If four arbitrators support the application this is taken to an RFA-like process that runs for a period to be determined, with a quota to be determined. A bureaucrat exercising his usual discretion then decides whether the candidate is to be desysopped, and a steward performs the desysopping if necessary.

At the arbitration committee's discretion, the application for desysopping may be taken on as an arbitration case instead of being sent to the community. This would help to distinguish conduct issues (which really belong to arbcom) from community support (which should be handled, of course, by the community).

I think this could be a much more streamlined and flexible system than those suggested so far. Typically the arbitration committee will either pass or reject an application inside seven days (often much less). Requiring arbitrator input at an early stage is a good quality control system. The community would still have input, and the bureaucrats would still exercise such discretion as they have. This would address concerns in many old hands, such as myself, that this process would seriously degrade the scope of action of administrators, while ensuring that the community has the option to reappraise adminships. --Tony Sidaway 22:16, 16 August 2006 (UTC)

Utterly pointless

This is simply a troll incitement. There will be lots of attempts to recall unpopular admins - and good ones will have to defend themselves against trolls. But, no-one will be desysopped by the process. So the net result is lots of noise and disruption, but no admins removed.

Why do I say that? Well, I've no doubt that it will be possible to get 30 names to trigger the process in a few cases. But it will be impossible to get the required 80% consensus to desysop thereafter, unless the desysopping is so obviously merited - that arbcom would have done it without a though anyway. I suppose some people are supporting this because they think they can use it to bring folk like Cyde or Kelly Martin to book. Well, news for you, it won't work. Admins who are acting in good faith, even if in unpopular ways, will always manage to get at least 20% of the community to support them. As I say, lots of disruption, noise and discussion will result from this..... but nothing at all will be gained. --Doc 13:57, 15 August 2006 (UTC)

It's not a troll incitement. Admins in good standing and with the support of their administrative peers will have no concerns at all. rootology (T) 14:51, 15 August 2006 (UTC)

Thanks for the assumption of good faith, Doc. It so happens I'm not planning on using this on anyone, and I'm a big supporter. Admins acting in good faith have nothing to worry about. Karwynn (talk) 13:59, 15 August 2006 (UTC)

I'm not assuming anything. I said 'some people'. My point is that any admin may be harrassed by this, but that no admin (good bad or indifferent) will ever fear desysopping through it (other than miscreants who arbcom would hang anyway). So what is the point? What is the actual gain, for all the grief this would cause? Anyway, it is obvious this will never garner consensus - so it is dead in the water already. --Doc 14:19, 15 August 2006 (UTC)

There are currently three proposals for community recall going around, it's an issue that comes up all the time, and there is a lot of active discussion going on. I removed the rejected tag, I think it's too ealry to decide that. - brenneman ^{L} 14:30, 15 August 2006 (UTC)

Fine, but it is never going to fly. --Doc 14:31, 15 August 2006 (UTC)

Unfortunately, I think Doc's right. "To what point and purpose, missy?"

1. If it's supposed to be easier/more reliable in borderline cases than going through ArbCom, then it probably fails that purpose currently. 30 users of significant edits must ratify it, and then it can be defeated by 20% in a wider re-RFA? Take a look at how divided opinion is on the Administrator's Noticeboard, or on any given talk page. This will end up being tyranny of the minority and not what you intend at all, I fear, and will just give those admins a pass for the next 9 months.
2. If it's supposed to be harder than going through ArbCom, then why not file an ArbCom case in the first place and be done with it? -- nae'blis 14:32, 15 August 2006 (UTC)

The goal, then, should be to make it less complicated than ArbCom. Karwynn (talk) 14:34, 15 August 2006 (UTC)

It isn't about simplicity, it is about effectiveness. What is this for? If it is about desysopping really obvious bad admins (which is all the 80$ rule will ever desysop), then it is unneccessary - since anyone who is obviously that problematic will be shot by arbcom long before your process --Doc 14:40, 15 August 2006 (UTC)

Then obviously it's for ones that aren't so obvous. Here's the base premise: Admins who have lost the faith or trust of the community should not be admins anymore. FOr you to call this a troll's incitement is a pretty strong claim. Perhaps you're idea of trolling is too broad. Karwynn (talk) 14:44, 15 August 2006 (UTC)

You misunderstand. I mean, this process will often be triggered by trolls - and I know that's not the intent - but it is what will happen. Of course, the trolling applications for recall will not succeed - but they will still occur. Against that negative, I fail to see any possitive. You say, it is for less obviosly bad admins - but you'll never get 80% to desysopp where it isn't cut and dry, and where it is, arbcom will do it faster than this process. --Doc 14:51, 15 August 2006 (UTC)

The support of your peers in all things is paramount. We have community bans for when a group of admins agree that an editor no longer has anything to contribute to Wikipedia, which is basically set in stone. This is a similar to that in a regard and a natural extension of RfA. The community and it's peers giveth a few extra buttons, and the community haveth the right to review (which is all the process really enables) whether you should continue to have the problem. Part of the issue also is that adminship is "no big deal", but it is perceived to be this Holy Grail of status. It's not and shouldn't be perceived as such. It's simply the community thinking that person x is worthy for a time to be able to keep tabs on other users--if a supermajority of users decides that is no longer the case, why should it continue to be? rootology (T) 14:49, 15 August 2006 (UTC)

If a supermajority think that an admin should be desysopped, then they are probably right. No argument. But you'll never get a supermajority except where it is so obvious that arbcom will beat you to it. So there is not net gain. And what there will be are loads of failed attempts to get that supermajority, some of which will be pain trolling - that will be very disruptive and to no clear gain. --Doc 14:54, 15 August 2006 (UTC)

Doc's correct that the current way this is being worked on is doomed. Why are we futzing around with the most microcopic levels of detail before there is even general agreement on the principle? Stop playing with the numbers, please. Almost no one is going to say "well, I object if good standing is 2000 edits, but if it's 1785 then I'm in!" - brenneman ^{L} 15:01, 15 August 2006 (UTC)

Doc, I think I see what you're saying now. We've discussed this before; it's not in the page history because it got moved out of nowhere. If trolls file, they'll fail. No big deal. THe sad thing is, there is almost no way to prevent trolling, vandalism etc. All systems are vulnerable to trolling. YOu just have to set measures to counter the trolling. Karwynn (talk) 15:07, 15 August 2006 (UTC)

You also need to show that the benefit of the system outweighs the inevitable negative of trolling. I can't see any way that it does. You've got a prnciple of community accoutablity - but no pragmatic outworking. To be honest, your problem is that the 80% rule defeats the purpose. To work this would need 50% - and that will not fly. The only way this will get through is to make it so weak that it won't be worth it. --Doc 15:15, 15 August 2006 (UTC)

Believe me, I know. THe proposal as is is worthless because it will never be successful. But the concept is still there. Trolls should not scare us away from new processes. Karwynn (talk) 15:18, 15 August 2006 (UTC)

Wikipedia is pragmatic. we invent processes to improve things - not because they are noble concepts which won't work.--Doc 15:23, 15 August 2006 (UTC)

All systems are vulnerable to trolling. YOu just have to set measures to counter the trolling. Not necessarily. But any system which is based on numbers will be more gameable than one that is focused on discussion, rational debate, and thoughtful outcome. Giving specific numbers just makes it not scale with growth, and gives the trolls something to aim at. Ever wonder why RFA gets relatively little trolling? It's NOT a Vote, for one thing, so just dragging in raw numbers doesn't help. -- nae'blis 15:19, 15 August 2006 (UTC)

P.S. I am in constant edit conflict here. Now do you understand why Aaron wanted to refactor the page? -- nae'blis 15:19, 15 August 2006 (UTC)

No, with discussion going on with regard to administrator action, there will be all sorts of branches of trolling. Prima makes a comment. Secunda deems Prima to be a troll, and thoughtlessly removes their comment. Prima puts it back, saying not to remove comments. Secunda removes it again, turns whatever heading they're in into a discussion about Prima being a troll. Whatever "side" is in agreement with Secunda begins cheerleading the removal and jeering at Prima. Surprise! You've just created a new problem. WIth a number system, this will still happen, but Prima's votes will still count. See, that's the thing: trolling doesn't always come from the least respected editors. Karwynn (talk) 15:24, 15 August 2006 (UTC)

Top down

Can we agree to stop adding incredible levels of detial to this proposal for right now? There is no agreement on them, several people have objected because of them, and debate over what colour hat the exchecker must wear while squeezing the testes of the whip are totally obscuring any debate regarding the underlying concept. - brenneman ^{L} 15:06, 15 August 2006 (UTC)

Detail prevents wikilawyering. We've discussed this before. I'm fine with resuming discussion, but it probably would've been easier to stick to the active headings. Karwynn (talk) 15:08, 15 August 2006 (UTC)

Details are for wikilawyers. You've no agreement in principle. You've no clear aim. You've not defined a problem. You've failed to demonstrate how this would practically improve wikipedia. Untill you address those issues, this whole thing is doomed. Details be damned.--Doc 15:12, 15 August 2006 (UTC)

Read a lttle more, you'll see the aim of this process is crystal clear. Details on what limitations there are will prevent people from squabbling about deleting recall requests, protecting them, blocking protected users, etc. THe details provide less wiggle room, therefore less wikilawyering. If you're here to just shoot down the proposal, you've made your view clear and we don't need to hear it over and over again. Either discuss to improve or state your opinion and be done. Karwynn (talk) 15:15, 15 August 2006 (UTC)

Oh FFS. If you have any hope of changing policy, as opposed to having a nice accademic discussion about the finer points of a good non-starter, you need to discuss it with critics like me. You need to convince me that there is a problem, that you have a workable solution, and that the net effect will advantage wikipedia. Yes, you can tell me to beat it - and perhaps I will. But if you do, then you really are wasting your time. Now, if you want to be serios, tell me when you think this process would actually work and deal with a problem admin that arbcom can't. Be specific. What is the actual problem? --Doc 15:21, 15 August 2006 (UTC)

I have missed you so. - brenneman ^{L} 15:27, 15 August 2006 (UTC)

The bottom line is that it hasn't raised the kind of supporting comments it would need at this early stage to have a hope of progressing to acceptance as policy. It could be that this is simply a good idea whose time hasn't yet come, then again it may simply be fundamentally unacceptable to Wikipedians. Only time will tell. Feel free to continue discussing it--I wouldn't want to stop that--but perhaps it might also be productive to look at the many opposing comments and decide whether there is some basic problem that, if removed, would turn that opposition into support. --Tony Sidaway 15:24, 15 August 2006 (UTC)

Some changes have been made by discussion, but several objections boil down to this:

• Administrators might lose sysop priveleges unfairly
• Wikipedia is helpless against trolls
• Someone might file a recall against an admin I like
• ArbCom is satisfactory, another process to desysop is not in order

These concerns are inherent to the policy (or any policy with a system of removing de-sysop priveleges) and cannot be fixed. :THe only thing to do is discuss the principle of the objections, and that's getting us nowhere. Karwynn (talk) 15:29, 15 August 2006 (UTC)

Actually, that's not my point. I've no objection to bad admins being desysopped. Of course, the proecess will allow trolling, any process does. But, we usually put with trolling, because the underlying process is useful - not just conceptually virtuous - but actually useful. This process isn't useful, because it will not desysopp anyone who would not be desysopped by arbcom anyway. No net gain. --Doc 15:35, 15 August 2006 (UTC)

Ok, attempting to revive a pseudo-quick poll here: I know that currently Karwynn wants the detials and I think having them is counter-productive. Can we hear from anyone else on the merits of reducing this to bare bones and then building it back up? - Aaron Brenneman 15:40, 15 August 2006 (UTC)

It seems to me that only a complete reappraisal of this proposal could save it. The principal objections seem to be that it would be operationally redundant with arbitration, it would take longer than arbitration typically takes to desysop, and it would be used by trolls to stir up bad feeling about administrators. I want to see bad admins desysopped perhaps more than many others in this discussion, but since I've seen this happen at arbitration I need to know what's broken about the current system that needs to be fixed. --Tony Sidaway 15:50, 15 August 2006 (UTC)

Hi Tony. The ONE thing that a community process would seem to have an advantage over ArbCom is that (if I understand correctly) ArbCom requires all previous steps to be exhausted (RFC, DR, etc) before they will hear a case, and that all steps must apply to the same behavior/instance; thus they rarely if ever deal with chronic cases, rather than acute ones - is that a fair summary from your POV as a clerk? If so, I could see some advantage to a system that somehow managed to thread the needle between being trollbait, and ArbComm (yes, I just made that up). -- nae'blis 16:13, 15 August 2006 (UTC)

If someone's conduct as an administrator is so appalling that he simply must be desysopped in short order, it's quite common for a steward to perform the desysopping pre-emptively--this takes a few minutes or a few hours at most. If the case is misuse of sysop privileges or other cases of unfitness to use the sysop bit, historically the arbitration committee has not demanded prior dispute resolution--at least, not on the same issue. See for instance Cantus vs. Guanaco, Stevertigo, Pedophilia userbox wheel war, and Guanaco, MarkSweep, et al.

On the facts, I don't see any reason to doubt the Arbitration Committee's competence and willingness to act in the case of unfitness to sysop. You don't see any chronic cases of bad sysopping going to arbitration precisely because the arbitration committee acts long before such cases have had a chance to develop. Abuse of the bit sticks out like a sore thumb and is such a blemish on Wikipedia that the current Arbitration Committee has typically dealt with such cases in a more timely fashion than other issues. The Pedophilia userbox case, for instance, took considerably less than a week from start to finish, and resulted in two desysoppings and sanctions against several other administrators. --Tony Sidaway 17:08, 15 August 2006 (UTC)

Aha, there was some discussion up-page that intimated otherwise, so I'm glad I asked. Thanks. -- nae'blis 17:26, 15 August 2006 (UTC)

I say go for it, Aaron and everyone

As initial author of this I support Aaron's idea to strip it down and rebuild. The ultimate goal is simply to empower the community to undo a previously granted RfA, for the express purpose of removing an admin who 1) has lost the support/trust of his peers; 2) is not trusted by his peers; 3) peers are both regular editors and admins. My number-based criteria was simply an attempt to Troll Proof it from the get-go. rootology (T) 20:41, 15 August 2006 (UTC)

"Good standing"

I really don't think this term ought to be used in the policy. It's used on so many policies that a) it's meaning might be ambiguous and b) this policy might help set a precedent for what "good standing" is, something many particularly venomous editors seem to be pushng for (none in this discussion though). THoughts? Karwynn (talk) 15:33, 15 August 2006 (UTC)

The lead links to the bit in the text that explains exactly what "good standing" means so there is no ambiguity. Change the naming convention if you'd like. But saying things like "particularly venomous editors" doesn't help keep things calm, ok? - brenneman ^{L} 15:38, 15 August 2006 (UTC)

I'm quite calm. There are venomous editors who throw around the idea that if you have such-and-such edits, you're alright, and if you have less, you're a nobody and likely a troll. I also specifically stated that I'm not referring to anyone in this discussion, so if you think I'm referring to you, I'm not. I see no reason to get upset about this. As far as my concern goes, I'm aware that you linked to it. it's the naming I have a problem for the reasons above, and thoughts on those reasons?

Good standing is an established term, though; it's not as though someone just pulled it out of a hat. -- nae'blis 16:19, 15 August 2006 (UTC)

Yeah, I'm aware of that. In fact, it's part of my concern. This should be familiar -

It's used on so many policies that a) it's meaning might be ambiguous and b) this policy might help set a precedent for what "good standing" is

If no one's going to comment on this, I assume I won't hear any objections to changing it? Karwynn (talk) 16:51, 15 August 2006 (UTC)

This is absolutely not the place to define good standing, so yes I object. However I would support if you want to either make a Glossary edit, or start Wikipedia:Good standing/Wikipedia:Editors in good standing... -- nae'blis 17:22, 15 August 2006 (UTC)

Ok, enough confusion. We're in agreement (I oppose the setting of precedent for good standing, hence my concern), so I'm changing it back. Karwynn (talk) 17:26, 15 August 2006 (UTC)

Oops, forgot to kill the section. Good catch. Karwynn (talk) 17:29, 15 August 2006 (UTC)

Workable policy

OK, so far I've been negative. Here is what I think could work - and might just get support.--Doc 16:07, 15 August 2006 (UTC)

It is very simple:

New policy (or perhaps a restatement):

• Any administrator who is deemed to be damaging the project, or who has overwhelmingly lost the confidence of the community, may be desysopped.

Process:

1. The initial step is a request for comment - to allow an attempt at solutions by community discussion.
2. If this is obviously failing to address the concerns, after 3 days, any administrator may initiate a 'recall petition' on the RfC.
3. Providing that this motion is supported by at least six other administrators, it may be refered to Arbcom.
4. Arbcom shall then consider whether an administrator: 1) still enjoys the confidence of the community, and 2) is still of net benefit to the project.
5. If Arbcom deems the administrator to fail one of these conditions, it may order the desysopping of the administrator, or impose other restrictions or remedies as it sees fit.

That could work quite nicely, but I'm wondering how the ArbCom would determine if the admin in question "still enjoys the confidence of the community"; the obvious methods—straw polls or re-RFAs—would bring us back to the original point, so presumably some other method would be adopted? Kirill Lokshin 16:21, 15 August 2006 (UTC)

A quick look at the prior RfC, the talk page of the admin, and the numbers endorsing the recall petition should suffice. Polls are evil and often not reflective. Arbcom members are part of the community, so I guess they'll have a pretty good idea.--Doc 16:24, 15 August 2006 (UTC)

It's got some merit, although I'd like to hear from at least one ArbCom member about their viewpoint on workload vs. payoff for such a community-started system. I'm also leery of an "admin-only" step, but it's probably too tough to define a single member in good standing without making it trollbait. -- nae'blis 16:27, 15 August 2006 (UTC)

This is a very good idea--it's basically an Escalation Function for converting an RfC into a direct ArbCom matter. rootology (T) 20:44, 15 August 2006 (UTC)

And as RfC is basically a toothless wonder, this would grant additional legitimacy (quite a great deal) to that process as well. rootology (T) 20:45, 15 August 2006 (UTC)

I like the "lightly" approach of Doc's too, but I'm going to split this into a "working copy" box, with my proposed changes in bold. Anyone who wants to can make an edit in this box keeping the first as our current point of reference. Sound ok? - brenneman ^{L} 16:29, 15 August 2006 (UTC)

Much better, IMO, but I don't think it will address rootology's initial concern, because it still stops the buck at arbcom, which he (justifiably or no) has doubts about. --SB_Johnny | ^talk 16:35, 15 August 2006 (UTC)

I'd like to hope that they would certify all but the most outrageous cases. Someone has failed to get the consensus and a b'cat has stamped it, but they'll still say "no way?" I have more faith in them than that. - brenneman ^{L} 16:38, 15 August 2006 (UTC)

What's a b'cat? --SB_Johnny | ^talk 22:03, 15 August 2006 (UTC)

Nothing to do with ArbCom, again--it's to provide a means for the community and the admin's peers to reevaluate directly whether he should continue on as admin based on his actions. rootology (T) 22:08, 15 August 2006 (UTC)

Alternative

Make it even simpler.

1. When a groups of editors in good standing feel an admin should be desysoped, they write a petition on the merits and sign it.
2. The admin in question responds.
3. After that, a week talk page discussion to get the feel of the community.
4. Vote by the ArbCom on the merits.

End of story. (Arbcom has the right to kill frivolous petitions in any stage). This might result in two or three desysoppong events as soon as it becomes implemeted. It is clear, it can nt be highjacked by a group of trolls or so, and it leaves the power with the ArbCom (where I think it should remain).-- Kim van der Linde ^{at venus} 16:33, 15 August 2006 (UTC)

Please edit me

• Any administrator who is deemed to be damaging the project, or who has overwhelmingly lost the confidence of the community, may be required to stand for reconfirmation.

Process:

1. The initial step is a request for comment - to allow an attempt at solutions by community discussion.
2. If this is obviously failing to address the concerns, After at least number Z days of discusion, any administrator may initiate a 'recall petition' on the RfC.
3. Providing that this motion is supported by at least arbitrary number X other administrators and arbitrary number Y, a request for reconfirmation will commence.
4. This reconfirmation will have the structure of an existing request for promotion, and consensus will be interpreted by a bureaucrat.
5. Arbcom shall then consider whether an administrator: 1) still enjoys the confidence of the community, and 2) is still of net benefit to the project.
6. If Arbcom deems the administrator to fail one of these conditions, it may certify the outcome and order the desysopping of the administrator, or impose other restrictions or remedies as it sees fit.

No good. ADmins can delete RfCs about other admins without consequence. Karwynn (talk) 16:37, 15 August 2006 (UTC)

Admins can delete any page. There's logs, though. I don't understand this objection, unless you just think all 1000 admins are corrupt. -- nae'blis 16:40, 15 August 2006 (UTC)

Ok, I created the category of admins open to recall, are you seriously suggesting that I'd just ignore it if someone deleted a request for comment? I think that I can count on my hand the number of certified request for comment pages that were deleted. - brenneman ^{L} 16:44, 15 August 2006 (UTC)

This is my experience:

Prima makes well-put together RfC about an admin complete with diff links.

Secunda brands everyone supporting it a troll and deletes it after a couple of days

Triada brings it up on the DRV. All admins who disagreed with it come to endorse the deletion, based on the fact that they don't agree with the files (I'm not making this up, they say that)

Triada is labelled a harasser, and is told he will be blocked if he ever brings it up again.

Not all admins are corrupt, but none care enough to keep the others in check. RfCs should never be speedy deleted, but they sometimes are. NOne of this was hidden, or conspired. It's just the way things happened. ANd nobody gave a damn. Karwynn (talk) 16:48, 15 August 2006 (UTC)

(The light finally dawns on me). So this is one of these 'admins can't be trusted' rants? If that's the case - goodbye. --Doc 16:52, 15 August 2006 (UTC)

NO. DId you even read a word of what I said? Karwynn (talk) 16:55, 15 August 2006 (UTC)

"ADmins can delete RfCs about other admins without consequence". That's a very telling comment about the motivations behind this procedure. If the underlying assumption is that all admins are corrupt and untrustworthy, and massive sets of rules are needed to keep them in check, then there's no point in keeping an eye on this: it's based on a false assumption, and consequently will never amount to anything worthwhile. --Carnildo 19:07, 15 August 2006 (UTC)

Well, good thing that's not what I think then, as I've said already. And there weren't any consequences. Look what I said: Not all admins are corrupt - NOne of this was hidden, or conspired - no conspiracy theory, no fight the man mentality here. Karwynn (talk) 19:19, 15 August 2006 (UTC)

[9] I'd ask anyone else commenting here not to mischaracterize my comments, thanks. Karwynn (talk) 19:22, 15 August 2006 (UTC)

OK you've pointed out one case where an RfC was deleted. However, in this case the RfC apparently (according to the deletion reason) consisted of nothing but flames. Do you have any actual examples of non-frivolous, certified RfCs being deleted by an admin without consequence? If an RfC was part of this desysop process and a rogue admin deleted it, do you really think 1000 other admins would refuse to undelete the page? If an admin can supposedly delete an RfC without consequence, why couldn't they also delete a recall page? Aren't I Obscure? 19:47, 15 August 2006 (UTC)

The purpose of this policy idea is NOT any silly assumption that all admins are corrupt monsters--anything but, I think at least 99.9% of you wield a fair and generous mop. Like I've said throughout the idea is simply to empower the community directly in some fashion to very fairly undo what it no longer holds to be a meritous RfA. rootology (T) 20:48, 15 August 2006 (UTC)

There are slightly less than 1000 admins. If you believe that "at least 99.9%" of admins are fair and generous, that leaves 1 (one) admin that might deserve to be desysoped. Do you think that an entirely new process is required to deal with a single subpar sysop? Aren't I Obscure? 21:37, 15 August 2006 (UTC)

No, what I mean is that I'm not trying to craft a policy to "get" admins. The point of this is to give the community the right to decide for themselves directly who whould be an admin if they feel collectively someone is no longer fit for the job. rootology (T) 21:48, 15 August 2006 (UTC)

The reason I took what you said literally was that it's important to perform a cost-benefit analysis when undertaking such a proposal. In this case, the cost includes the misuse of the process by trolls, ill will caused by recalls and all the bureaucracy involved. The benefit is having an additional mechanism (besides ArbCom) to remove an admin. If there were dozens of rogue admins that weren't being dealt with by the ArbCom, the benefit would be fairly substantial. However, if only .1% of admins deserve to be recalled, the benefit is essentially non-existent. Processes on Wikipedia are designed to address a problem and thus far it hasn't been shown that there's a problem to address. No problem to solve means no benefit. As long as the cost outweighs the benefit, this proposal isn't going to fly. Aren't I Obscure? 22:02, 15 August 2006 (UTC)

COuld you please explain a hypothetical situation where a troll could successfully abuse the system? Karwynn (talk) 22:05, 15 August 2006 (UTC)

I only said they'd misuse the process. I highly doubt they'd suceed, but that's not their goal. This is from our internet troll article: "The ultimate goal of most trolls is to cause a reaction and/or receive attention, whether positive or negative." What better way for a troll to waste everyone's time than to start up a recall. Aren't I Obscure? 22:11, 15 August 2006 (UTC)

So do what's best for trolls: ignore them. A filing needs 30 certifiers, 3 admins. That's 30 trolls, 3 of whom would have to be admins. Ignore trolling recalls completely and they won't get certified, and away they'll go. THere are all sorts of systems for trolls to use - RfA, AfD, RfC, etc. - but s the answer to close them all down and live in constant fear of trolling? Just ignore them and they'll go away. That makes more sense to me than not implementing any more new processes for fear of trolls.Karwynn (talk) 22:22, 15 August 2006 (UTC)

I'm an extreme case, but my activities in enforcing the basics of the image use policy (all images must indicate their source, and must have a license tag) has resulted in the deletion of images uploaded by around 40,000 users, including several hundred established users. Based on what I've seen on my talk page, people would be bringing up recall petitions once or twice a month, and if someone were willing to take the time, they could easily come up with thirty people to certify, and a thousand to vote against me. Do we really want a system that punishes people for enforcing policy? --Carnildo 23:54, 15 August 2006 (UTC)

Which is one reason this proposal is fatally flawed - it has no requirement that the recall proposal even allege that the admin's actions somehow violated a policy. So yes, theoretically a bunch of people could put you through the whole recall mess because they don't like the fact that you are enforcing policy. FCYTravis 02:00, 16 August 2006 (UTC)

Where/how did you come to this conclusion? The filer has to include the basis of the dispute plus evidence. rootology (T) 02:03, 16 August 2006 (UTC)

What part of "dispute" requires that a policy have been violated? There are lots of people who would "dispute" legitimate and necessary admin actions that in no way come anywhere close to violating any policy. "I dispute the fact that this admin deleted all the copyrighted, non-fair-use images I uploaded." No admin should have to go through such an ordeal unless there is actual suggestion that he or she has undertaken actions contrary to established policies. FCYTravis 02:11, 16 August 2006 (UTC)

Any active admin has violated a strict reading of policy a few times. Usually it's speedy deletion of material that doesn't meet a specific criteria but nonetheless doesn't have any place on an encyclopedia. The actual reason people are voting to recall doesn't need to bear any resemblance to the evidence given in the petition. --Carnildo 02:39, 16 August 2006 (UTC)

I see it in the same way as sysop rights. A burocrat can add them, only a steward (a level higher) can remove them. Giving the rights is determined by the community, removing them by the ArbCom (a level higher). This is a simple principle of safety. The way forward is to find a earier way to deal with abusive admins, without a ful blown ArbCom case. -- Kim van der Linde ^{at venus} 03:58, 16 August 2006 (UTC)

I looked at Internet troll you quoted and the entire introduction was changed by someone without even a talk page (whose sole edit was that article) and an anon IP the same day it was quoted to replace its previous first paragraph that has been there for at going back a year. Anomo 04:26, 16 August 2006 (UTC)

From this comment and your edit summary "rv new user without even a talk page and anon IP who changed the article purely so someone could quote it" [10], it appears you're accusing me of changing the article so I could quote it? You can't be serious. Are you really suggesting that I altered an article because I might quote it 20 hours later? That's one way to throw assuming good faith right out the window. To be perfectly clear, I have no idea who edited that article, nor any idea that the line I quoted had only been there for a short time. However, I still find the quote to be a rather accurate description of a troll's goal. Aren't I Obscure? 15:21, 16 August 2006 (UTC)

I'd almost say you quoted a vandalized version by mistake, but it wasn't that bad of an edit, just it changed what other people agreed on and that the changed version was copying redundant things found later in the article. I reverted the article back. I was tempted to try to mix them, but how the first paragraph had been was the result of lots of people trying to fix an NPOV version of it over time and I did not want to mess it up. Anomo 17:23, 16 August 2006 (UTC)

Cool off a little obscure, I highly doubt he was accusing you. Jumpng to that conclusion just makes it look fishier anyway ;-) Karwynn (talk) 17:40, 16 August 2006 (UTC)

It's difficult to not view it as an accusation when Anomo explicity stated in his edit summary that the article was changed "...purely so someone could quote it." If the purpose of the revert was just to return to a consensus version, why phrase the edit summary like that? Aren't I Obscure? 17:48, 16 August 2006 (UTC)

Oh yeah, I guess you're right.  :-( Still, could I politely ask you guys to take this elsewhere? Karwynn (talk) 17:53, 16 August 2006 (UTC)

Some principles

These are the principles as I understand them. Keep in mind this is not stone-hard fact here. It's the principles this policy is bult on. This is really rootology's policy, so he's more correct about this than I am, but here goes:
Sections below moved out.

The problems

1. SOME administrators act abusively (per Wikipedia conduct policies), with no accountability (a central part to Wikipedia), and/or incompetently. The number/percentage is unspecified. THese behaviors are sometimes the pattern and not isolated cases.
   1. Abusive conduct is detrimental to Wikipedia, per reasons outlined in conduct policy.
   2. Lack of accountability is detrimental to Wikipedia - consensus and discussion facilitate rational thought and better decisions
   3. Administrative incompetence - lack of knowledge of policy, or foolhardy actions using tools based on a lack of context about a situation or a lack of knowledge of all the circumstances, including rushed decisions - adversely affects blocks, deletions, and protections, inhibitting the ability of Wikipedians to build an encyclopedia.
2. Bad administrative actions, as defined above, causes the community to lose trust/faith/confidence in said administrator.
3. There is currently no system for the community to come together to desysop someone they have no confidence in. - Current dispute resolution steps are in the control of admins for blocks on participants, deletion, and protecton.
   1. There is no current involuntary recall process
   2. Filing an RfC will get it flooded by fellow admins, who in my experience generally proceed to throw trolling accusations around rather than listen to the discussion.
   3. RfCs about admins can be speedy deleted on sight, even if they've run for days.[11] There need not be discussion about this.
      1. Asking for the deletion to be reviewed gets you branded a harasser.
   4. People who complain about admins may be subjected to accusations of harassment or trolling, and are ignored they ask how they are harassing/trolling.
      1. As a result, civility, no-personal-attacks and respect are thrown out the window. Blocks may follow.
   5. There is a "community ban" for removing the status of "editor" (i. e., blocking) in the blocking policy, but no process for community removal of admin status
   6. Arbitration fails to fully desysop poor admins (meaning it gets some, but not all)
      1. ArbCom cases filed by less experienced users are less likely to be accepted
      2. Previous discussion on desysopping is considered incivil and is subject to penalty in Arbitration.
      3. "Failed" complainers are branded as trolls by administrators, decreasing the filing of potential cases.
      4. There are only 6 arbiters - don't have the time
      5. Arbitrators only desysop in cases of severe, severe abuse.
      6. Arbitrators are not intended to gauge the community's view of the admin

Result: If an admin is not trusted by the general community and only by his fellow admins (and sometimes not even then), he remains an admin, despite the fact that an RfA on his/her part would not succeed.

The solution (in theory)

1. Provide a method of allowing a consensus decision to desysop based on lack of faith/trust/confidence in the competence/conduct. Assume good faith in predicting the motives of non-troll users.
2. Provide a system to weed out trolling or frivolous filing
   1. A certain time period and number of edits are required to certify
      1. "Survival" factor - many trolls permabanned before achieving required time/edits
      2. Assumed greater knowledge of Wikipedia policies/guidelines
      3. Assumned committed interest to improving Wikipedia
   2. Failed Recalls "protect" admin for 9 months
      1. Makes filing in bad faith or for dumb reasons counter-productive for the filer
3. Provide a system to weed out emotionally-charged filing
   1. Needs 30 certifiers
   2. Needs consensus
      1. Unlikely for undeserved recalls, notices displayed at high-profile pages
   3. Takes a month to complete the process
4. Provide greater accountability for administrators - based on the community's trust (this is both a solution and an advantage)
   1. Assume that the community at large has decent judgement

Result: Good-faith consensus-backed desysopping of administrator deemed to be in poor standing

The advantages

1. Removal of poor sysops - less use of tools outside the scope of policy
   1. Removal of abusive administrators
      1. Less abusive blocking means more productive editors
      2. Less abusive deletion means more legit content for Wikipedia and greater capacity for discussion in the Wikipedia: namespace
      3. Less abusive protection means more productive editing/discussion
      4. In general, this will mean less conflict on some scale
   2. Removal of incompetent administrators
      1. Less poorly judged blocking, even when done in good faith, means more productive editors.
      2. Less poorly judged deletion means more content for us
      3. Less poorly judged protection (not a problem so much, but anyway) means more editing/discussion
2. More administrator accountability
   1. More accountability does not mean stricter polcy/guidelines on use of tools
   2. There is nothing wrong with an admin having second thought based on "Is this in line with Wikipedia polcy?"

Result: More expansion/improvement for Wikipedia

What this policy is NOT

1. An excuse to desysop
   1. Duh, voting procedure prevents undeserved desysopping
2. A reason to desysop based on an isolated action/ set of actions
   1. If the admin doesn't show a pattern of the bad acts, it will get voted down
   2. This is especially important in cases of perceived incompetence. One or two bad decisions is not enough - it needs to be a pattern to show a lack of improvement to the project
3. A discussion of WHY an admin should be desysopped
   1. It's a discussion of WHETHER to desysop someone - meaning, there are no specific criteria AND there is ample chance to oppose the desysopping
      1. In other words, it's not like initiating the process means a desysopping, it means it's going to be discussed.
4. Any disciplinary action beyond desysopping
   1. In other words, even if desysopped for abuse, admins won't get blocked or whatever (leave that to ArbCom)
   2. Again, this is especially the case for perceived incompetence - the admin may be acting in bad faith, even if his/her judgement is less than helpful
5. ANYTHING other than a process to improve Wikipedia and the means by which it is written - the recall process is not to be used if you just don't like someone.

Discussion

So... thoughts? Karwynn (talk) 16:32, 15 August 2006 (UTC)

I moved your signature up here so to keep it together if we start having threaded discussions on the below. I hope you don't mind. - brenneman ^{L} 16:41, 15 August 2006 (UTC)

That's fine. What are your thoughts? Karwynn (talk) 16:53, 15 August 2006 (UTC)

So we're back to "ArbCom doesn't work so let's have non-admins decide". Fagstein 20:41, 15 August 2006 (UTC)

The policy as drafted by me (2nd or 3rd draft in like 5 hours) required admin endorsement of the certification of the recall--the original was 25 "users" with 3 months/300 edits, then it was 30 with 5 months/1000 edits, then I settled on 30 with 10 months, 2000 edits, and 3+ admin certifiers of the 30 to eliminate stupid trolling. The nine month exemption was in it from draft 1. rootology (T) 20:50, 15 August 2006 (UTC)

I think this is going to face an uphill battle, mostly because of the first section: Problems. I think you'll find many editors/admins who will disagree with it, and I'll post my own opinions here to give you an idea of why.

1.1. Abusive conduct I do not agree abusive conduct is, absolutely, harmful to the project. I believe upholding some of the core principles sometimes requires what is, essentially, abusive coduct towards elements which would challenge those principles.

1.2 Lack of accountability Between arbcom, Jimbo and the Foundation/Stewards, I do not agree there is a lack of accountability.

1.3 Administrative incompetance I agree this is a problem. As I feel accountability is adequate, I think the solution to this issue is to strengthen the RfA process.

2. Bad administrative actions This may be true, but I don't feel it's helpful. Unpopular decisions can well have the same effect. And with the number of active editors, there's usually a good number of people who find any action unpopular.

3.1. No involuntary recall process Again, I feel arbcom fills this role. 3.2. RFCs while part of this dispute resolution process, should not be considered part of the admin recall process. 3.2 - 3.6 show a lack of faith in the current system. I doubt anyone could say anything to restore someone's faith, and for this I'm sorry. Even if you manage a consensus on these points, though, I don't think the solution is another layer of policy & process. If the current system is as flawed as you say, adding to it will only increase the flaws as the new policies and processes are "exploited."

Instead, as I said, I'd hope to see existing process refined and fixed, starting with RfA. If arbcom is too small to handle its workload, maybe it can be expanded or divided into an ArbCom for admin disputes, an Arbcom for policy disputes, an Arbcom for content disputes, etc., etc. This proposal, I fear, has little change of life because I feel there will be very many editors and admins who will concur on some if not all of my points in some way or another. --InkSplotch 21:14, 15 August 2006 (UTC)

Notice: WE'RE TALKING ABOUT THE COMMUNITY, NOT INDIVIDUALS. That means that saying "well, any good admin action is going to be unpopular among some people" is a waste of time because all that matters is the popularity to the COMMUNITY. Additionally, this is not about popularity, but about confidence in the admin's enhancement of Wikipedia's quality.

I agree about RFA, I think people ought to tighten up their RfA standards. And that doens't mean bump up the required edit count, it means saying "Is this guy really exceptional or just another experienced editor?" Additionally, I find it disturbing that you feel that breaking policy is not disruptive/detrimental to WIkipedia (since the principles I put up define abuse as deliberately breaking policy). Three questions: do you mean ANY editor breaking policy is not detrimental, or just admins are not detrimental when they break policy? Second question: Would you support, then, removal of the parts of the policies and guidelines that state that adherence to that particular is beneficial to WIkipedia? And third: Are you saying you don't think that abusive or incompetent decisions made by admins don't cause the community to lose faith here?

Also, I want to clarify: THis is not saying ArbCom don't have good judgement, only that they don't always get to everything because not everything is filed, for the reasons stated in the principles. It's not so much a lack of faith in the committee, more in the system. Karwynn (talk) 21:45, 15 August 2006 (UTC)

I appreciate you're focused more on the community than individuals, but then, how do you tell? We have an extremely large community, but not that many vocal members the Wikipedia space. So I see people try to draw a line, 10, 20, 30, people to certify...maybe admins too...I just don't believe there's a realistic number of editors you can find to make this process reliable enough. I think RfA is flawed enough, and I'm much more concerned about a process to remove admins making mistakes than the existing process to promote admins making mistakes.

To address your questions, and I apologize if the answers aren't as direct as you expect. (1) I don't believe all policies are created equal, or that all are meant to be enforced equally. So if some are broken in the enforcement of others, I'm OK with it. I look at the end results. (2) Not necessarily, although I do feel certain policies might benefit from updates or clarifications. Alas, I'm not able to provide you a list :) (3) No, I'm saying I think proper and competent decisions can cause the community to lose faith simply because they may or may not be understood properly. I don't think admins need the faith of the community to do their jobs. Be elected, yes, but once there they're called on to do some very unpopular, often contentious jobs. I have faith in our current resolution process (Foundation, Stewards and ArbCom) to watch out for the abusive ones. I think the incompetent ones are harder to spot, just because incompetent doesn't always lead to abusive actions, which is why I feel efforts are better spent on RfA.

I think the faith of the community is better placed in themselves to make a better encyclopedia. Focus on that, and while conflicts will arise, conflicts will be quickly resolved and the project will get better faster. --InkSplotch 22:05, 15 August 2006 (UTC)

Respectfully, I don't agree. I've gotten crap from admins when I'm (in intention, at least) NPOV-izing articles and upholding other policies like verifiability, original research etc, or by supporting the keeping of encyclopedic articles that are not to an admin's liking. If I disagree, I'm threatened with blocks and such. It's dangerous to dismiss any inter-editor conflicts as unproductive. But then, how is it decided whose edits stay and whose go? Essentially, by saying I'm to avoid all conflict with administrators, you're saying (or implying) that the non-admin side of a conflict (me) is the wrong side. And once we all accept that, the encyclopedia will benefit. So why not just cut the uselessness then and eliminate all non-admin users? I know that's not what you're suggesting, but that's where (I think) that line of logic takes someone. Point is, it's not always productive to ignore conflict - conflict is inherent to cooperation on some level, and to let a bad edit or something bigger, like the blocking of several productive editors, slide is not productive. Karwynn (talk) 22:12, 15 August 2006 (UTC)

Actually, I don't mean to suggest the editor is automatically wrong versus an admin. Just ask yourself, are your actions more or less disruptive than the harm to core principles like verifiability, neutrality, etc. you're trying to prevent? Is it an immediate need? Take it to AN/I, and realize you might ultimatly be wrong. And I mean wrong, which could result in blocks, bans, permanent community bans, or other restrictions. You might learn from this and go on, or the encyclopedia could lose an editor. On the other hand, you could be right, other admins might support you, and you gain respect as someone beneficial to the project. Any damage to the project gets fixed quickly, thanks to the exposure. It may sound callous, and I apologize if it does, but ultimatly Wikipedia gains from it all. How much you gain, is in how you conduct yourself. --InkSplotch 22:40, 15 August 2006 (UTC)

Archival

You really shouldn't be archiving discussions which were active just last night. User:Zoe|(talk) 18:42, 15 August 2006 (UTC)