User talk:Dominic/Proposal for desysopping

This is just an idea that's been forming in my mind for a while now, so I decided to put it to words. I'm not planning on making any formal proposal yet myself, unless anyone else wants to use this idea. In fact, I'm not quite sure whether there's any use in having any more process to desysop problem admins than we have already, which is not terribly unsatisfactory in my mind. However, if we were to have a process, I tried to find one that addresses both the fears of losing good admins to "lynch mobs" and of having only a "cabal" making the decision. As long as there's a conservative body to act as a check against unwarranted desysopping, the community should have its say. In any case, feedback appreciated, and hopefully this will at least stimulate some interesting discussion. Dmcdevit·t 20:41, 24 May 2006 (UTC)[reply]

A comment on reaffirmation standards

I hope you don't mind my adding my thoughts on this aspect here. If you do, I will move my comments to my desysop idea page.

Most consensus based processes on Wikipedia have an underlying assumption of the preservation of the status quo unless there is consensus to overturn it (ie. a consensus must be for the deletion of an article, a lack of consensus results in keeping it). Since the status quo in the case of desysoppings is that the person is already an admin, it would seem reasonable to require a consensus for the removal of the status to take place, rather than requiring a consensus to maintain the status quo. One of the failures of past deadminship procedures has been applying a standard to the select few that are forced to go through the procedure (an RfA 80% to pass standard), that many admins could not meet if they were forced to go through it. It puts too much of the onus of desysopping people on those who decide whether or not someone goes through the procedure, because it is often a foregone conclusion that most who go through it will fail.
Secondly, if admin actions are involved, it seems reasonable that there be a consensus or at least a majority of those who have experience in using admin actions (namely admins) agree that those powers have been misused. If one's fellow admins believe you are misusing your powers, that says a lot more to me than dozens of users, many quite new, many having a strong belief in "all use of admin powers is too much", saying that the admin is misusing their powers.
Giving someone the presumption of good faith, I would like to see someone desysopped only when at least 50% of the community, along with at least 50% of the admin community who participate, agree that the person should lose their adminship rights. This measure of community support will allow us to protect those who are targeted unfairly (without requiring unreasonable 80% supermajorities), and yet allow the community to easily remove those who have legitimately lost the community's trust. NoSeptember ^talk 21:11, 24 May 2006 (UTC)[reply]

No problem at all with you giving your thoughts here; I put this up hoping it could cause a fruitful discussion. I'll respond to each point in turn.

It's an interesting question what the consensus should be for: deadminship for reaffirmaion. Let me make plain, though: this proposal is heavy on process, and more than is typical of me. I have no illusions about what "consensus" on RFA means; bureaucrats typically go no furher than calculating percentages. 80% is "consensus" only because that's what the community has defined as an acceptable number for bureaucrats to promote under. In that sense, the role of the bureaucrat is to feel out where the community stands on the issue of consensus, and act accordingly. Saying "a consensus for reaffirmation" may very well mean, depending on the community's feeling on this, ony 60% or so. Three main reasons why consensus for reaffirmation still seems like a good idea for me: 1) I'm not so sure if it's status quo or the conservative option that we should be favoring (keeping an article in a disputed deletion is more conservative, not promoting a disputed RFA is the conservative option, and removing a disputed adminship that can't gain consensus is the conservative option), when we're talking about admins, do you really want to keep the one that 40% or more would like demoted? 2) ArbCom: in the case that too-high community standards get the better of a useful administrator, the check of Arbitration Committee consent is designed to be there to prevent just that, and only that, occurence. 3) Step 1: discussion/evidence and certification thereof comes before the RFA for that reason; there has already been an attempt to resolve the situation, there is already demonstrable community concern over the admin's judgment, and there is already compiled evidence for the very first people to the RFA to evaluate. Once that stage is reached and that community concern already exists, it seems it can hardly be unwarranted to at least see whther the community/arbcom still has faith in their position.
I agree. A consensus of administrators is far more likly to be sensible. There are no trolls or newbies, and hopefully none that are ill-informed or of poor judgment. But this is true of any decision. For the same reasons, a consensus of administrators is more likely to be sensible of RFA, DRV, etc. In fact, a consensus of a subset of even more sensible than average administrators would be even better, and a consensus of Dmcdevit even best (of course). But I'm not convinced that this means the community, even incorporating the newbies, would really go wrong, any more than they do on RFA. At least, I'm not sure that arguments for suffrage are any morre compelling here than they are elsewhere: do the drawbacks of having newbies, trolls, and people who are wrong commenting outweigh the problems, even if it is just one of appearance, of restricting the commenting?
I would consider someone with only 55% of the community behind them to have lost the communty's trust. Maybe a near majority supports them, but that doesn't demonstrate trust from the community (which would imply agrement on the matter, not a near split) so much as uncertainty. Considering what adminship is and what it means to the community, I do still have a hard time accepting that a "no consensus" decision should lead to an admin retaining their tools. There is no longer consensus because the community's trust has changed to uncertainty. Dmcdevit·t 07:41, 26 May 2006 (UTC)[reply]

Let me respond to a few things you wrote for now:
"when we're talking about admins, do you really want to keep the one that 40% or more would like demoted?"

This is the wrong question. We routinely keep admins with these low levels of support. We do it by not forcing them to go through a reapproval procedure. If we are willing to keep unpopular admins routinely, what has really changed just because someone decides to put an admin through your procedure? We are only going to desysop those admins who have a campaign against them started. It's like the lottery, you can remain an unpopular admin as long as you don't pick up a troll who is dedicated to pushing your buttons and driving you out of adminship.

"ArbCom: in the case that too-high community standards get the better of a useful administrator, the check of Arbitration Committee consent is designed to be there to prevent just that, and only that, occurence."

This is why I have said in the past that ArbCom should just bite the bullet and make desysop decisions themselves, instead of going through the pretense of having a community decision. Admins desysopped because of a controversy just can't succeed in a community decision, and ArbCom members know this. So if the ArbCom thinks the admin shouldn't be desysopped they will find a way to refuse to consent to the process. We should be honest that ArbCom is the group that is really making the key decision here, any community vote followup will be totally predictable (certain defeat). I don't think it is bad for ArbCom to be making the final decision btw. ArbCom is made up of active long term admins, so in a sense can be seen as a substitute for my idea above that a majority of admins must agree to desysopping another admin. There is some community control of ArbCom, since the community can (and has) refused to reelect some of them. Also, ArbCom members will know what the community wants, there are plenty of outside parties who comment on these things, you will know if the community is really against someone remaining an admin.

On the idea of a 50% majority. If ArbCom can make decisions by a simple majority, why can't the community? I see the community affirming the adminship of someone by 55% as putting them on probation, letting them know that if they screw up again, they are likely going to fail the next time. Since we don't automatically reconsider admins here, any higher standard turns the focus on the process itself, instead of on the admin (you go through the process = you lose, you avoid going through the process = you win). People would be much more willing to go through this process and have others go through this process, if the standard is one where many will succeed and some (the true bad apples) will fail. That is so much better than a process where even good admins fail because they can't swing a 4 to 1 support ratio. NoSeptember ^talk 14:24, 26 May 2006 (UTC)[reply]

Unthreading the above

I'm going to make a little section for each of the thoughts we're following before it gets too messy. - brenneman ^{L} 05:37, 27 May 2006 (UTC)[reply]

ArbCom oversight

While I understand the idea behind it, I shudder to think about the consequences if this actually needed to be used. It's implicit, but just to be sure: This would require a simple majority of non-recused active committee members? - brenneman ^{L} 05:44, 27 May 2006 (UTC)[reply]

It is useful. Many people would shudder to think of a community-based process without such a check, conjuring up nightmarish visions of good admins kicked out by swarms of trolls and cliques. Also, the Arbitration Committee is already vested with the power to remove adminship at will, and so this is not a large leap to make them oversight for this process (with conservative use of sucj oversight). As to the logistics, not so important, but a simple majority of non-recused active committee members is the likliest and most sensible way to determine it. The other mechanism we could use, which I think could streamline things, and would go more towards my idea of the unobtrusive oversight, is that no objection means no arbcom inolvement. the process could go along undisturbed, until an arbitrator disagrees with a particular decision and brings it up to the Committee, whereupon a consensus approach is decided. This is largely how checkuser and oversight permissions discussions have happened (eg: Dmcdevit: "I've talked to Taxman, we all know him, and think he should have oversight. He's ready and willing." Kat: "Absolutely." Mark: "I agree." One week later, no objections, he gets it.) How the ArbCom decides to do it is trivial, I think, as we can trust them to do something that works and that represents their tru opinion, and that's all we want. Dmcdevit·t 08:02, 9 August 2006 (UTC)[reply]

Jury selection pool

We don't have that much chatter at RfA now, and the B'cats sort it out ok. I suggest starting without any sufferage and if there's a problem, tighten it up later. I'd be strongly opposed to having any sub-group be chosen. (Barring the G = {Dmcdevit} set mentioned above, of course.) It's a Ceaser's wife issue, and part of what I'd like to see come of something like this is a lessening of the percieved distance between admins and non-admins. I'd hazard a guess that self-interst would mean that all admins kept a very close eye on these, and thus we'd get a higher sysop turnout than for a standard RfA. - brenneman ^{L} 05:55, 27 May 2006 (UTC)[reply]

With NoSeptember's sensible point above noted, I still agree more with your view. The drawbacks of suffrage aren't worth the benefit, at least until we see how it pans out as is. Dmcdevit·t 08:02, 9 August 2006 (UTC)[reply]

Dead-minn cutoff

I'm going to further split this one, as the bar we choose depends upon which way we're cutting it. - brenneman ^{L} 06:00, 27 May 2006 (UTC)[reply]

Personally, I think it should be a community dynamic with bureaucrat discreation to interpret the community's will, just as RFA is. There should be no preset cutoff at all to restrict us. Dmcdevit·t 08:02, 9 August 2006 (UTC)[reply]

Defaults to keep, so to speak

Defaults to delete, as it were

This is the current "regular" RfA method. I'd prefer this, with the same rules for staying an admin as for becoming one. It's more straightforward, but simpler in execution. I also believe that we'll not know how bad a time a "normal" admin would have here until we try one.
brenneman ^{L} 06:00, 27 May 2006 (UTC)[reply]

I would consider those listed here to be examples of how admins will fare, the involuntary ones, since this proposal is designed to handle "habitual misusers". NoSeptember ^talk 06:13, 27 May 2006 (UTC)[reply]

However, it is important to look at that pool of involuntarily desysopped. Most admins don't get desysopped until after they've done some egregiously stupid things. Now, I've been called just today one of the hanging judges on arbcom, but I think that there is a certain amount of laxity on arbcom towards this kind of thing that the community would not have in an RFA. Wheel warring with Jimbo, indefinitely blocking administrators without warning or just cause, unblocking yourself and blocking a pile of admins who blocked you, etc. The fact that this small number hasn't had high success in re-RFA does not mean that th3e community is overbearingly harsh, but that your sample is skewed. If we were to put the many admins who have weathered arbcom after some misjudgment(s) and come out with an admonishment or even a binding remedy, I think many would survive. I also think most average admins would definitely survive, but we would just have to see waht happens. I prefer that an admin needs consensus to remain an admin. No consensus = no admin makes as much sense in initial promotion as in reaffirmation. An administrator needs to have the community's confidence, not lack of definitive distrust. (Most failed RFAs are never so to the level that there is consensus that the person should not be an admin.) Dmcdevit·t 08:02, 9 August 2006 (UTC)[reply]

I cannot however help thinking back to Sean Black part two. I would not have had an inkling that he'd need a "discretionary" b'cat closing. - brenneman ^{L} 11:51, 9 August 2006 (UTC)[reply]

Very true. I guess my reason for such irrational optimism is that I'm imagining bureaucrats feeling out the community idea of what constitutes an acceptable threshold and that threshold being lower than new admins (but certainly not majority or anything so low). 75-80 (though I abhor putting it into numbers at all, even bureaucrats do it, sadly) is where they put it currently, which is absolutely arbitrary, but fits the community's needs and desires. 70 for reaffirmations may very well be the community's own standard, especially considering the lack of controversiality od Sean's promotion, unless I didn't see it. Sounds reasonable to me, at least. Dmcdevit·t 08:12, 11 August 2006 (UTC)[reply]

Poke

As there is some moevement at the station regarding "open to recall" anyone have any thoughts on this? How about a more pubic showing? - brenneman ^{L} 17:52, 7 August 2006 (UTC)[reply]

what kind of comments were you looking for? There are parts of this I like but the pseudo RfA part, maybe not so much, I'd prefer a more consensus driven model without even erstatz voting... ++Lar: t/c 22:13, 7 August 2006 (UTC)[reply]

This seems like a decent idea. Don't let it die off, but leaving it to the hands of a privledged few won't really do anything. Attic Owl 02:51, 9 August 2006 (UTC)[reply]

Poke noted. I guess I'l start by making some responses to the months-old comments above I seem to have forgotten about. But remember that it's a wiki, if you like my idea, or some part of it, take it and run with it. I'm not owning it; I'm just not pushing it (and have reservations along the same lines as Lar). Dmcdevit·t 07:18, 9 August 2006 (UTC)[reply]

perennial

"This is a perennial proposal, and one that has major problems that prevent most editors, admins especially, from getting on board." What is a perennial proposal? The proposal itself, or does the proposed system have some kind of perennial aspect to it?--Paraphelion 07:22, 15 August 2006 (UTC)[reply]

The idea of having a formal process for desysopping administrators other than through arbitration is the perennial proposal I meant. It is seen frequently brought up (usually by relatively new users who don't know it has been suggested before) frequently on the village pump of WT:RFA, but usually dismissed as one of those irresolvably perennial proposals (hence my wariness). This particular proposal has never been stated in this way before, that I know of. Dmcdevit·t 05:39, 16 August 2006 (UTC)[reply]

Perennial proposal. You recognise it yourself, so there you go. Nice try though. :-)

Having a look at it, well, using Requests for comment and Administrators noticeboard as a basis for your own process was not a good start :-/ . Those are processes that are not working as designed, and they probably even need to be removed at some point (I'm not looking forward to the wikidrama).

Hmm, I agree admins could do with more accountability, but maybe we need a different angle to achieve that.

Here's an idea... it's counter-intuitive, but have you considered making *more* admins? If we get our admin percentage back up a bit, admin actions will tend towards being the Right Thing more often, due to the wiki-principle. Right now there's often only one admin looking at any particular situation, and they often have to make the call all on their own. No wonder they sometimes make mistakes!

Kim Bruning 09:24, 14 September 2006 (UTC)[reply]