User:Friday/bcrat

some problems
We promote bad admins sometimes. A certain amount of this is unavoidable, but I don't understand how we got to a situation of "When in doubt, promote" (citation needed) when this action is not reversible. Do we not agree that bad admins are a huge time sink?

Some people seem to want crats to judge only what the community thought. Some people want crats to judge the suitability of candidates. These different notions can probably lead to all manner of disagreement.

Would people feel better if the bcrats made decisions as a group, rather than one individual sticking their apparently single neck out? It looks like this is already being done sometimes, in questionable cases. This is almost certainly a good thing.

Certainly there is no shortage of criticism and panic when crats act boldly. One probably legitimate point: if crats are planning to promote outside a normal range, they should so say well in advance. Otherwise, many people who disapprove will refrain from opposing, in order to not pile on. Once done, this skewing of results is unknowable and unfixable.

Particularly bizarre was the disagreement with some of the opposes on Riana's RFB. Some people wanted to reject any that mentioned her nomination of a particularly unsuitable candidate as being irrelevant and not about Riana. On the contrary- someone demonstrating very questionable judgement on who should be an admin is very relevant. This is the area that crats operate in. How could it not be relevant? This must be due to people making a huge separation in their minds between the crat's opinion and them "reading consensus". I don't believe such divisions are realistically possible. I want crats to have good judgement and not make decisions that are contrary to it. Community consensus is mainly a rationalization to please the masses.

I think crats should promote only when there are no substantial, valid objections. In AFD, we expect closures to be policy-based, and to overrule majority vote as needed. We can't easily do this at RFA because there are no policies about who should be a sysop. There are some best practices, but the community already commonly ignores then.

Examples
Purpose of examples: try to demonstrate that when candidates have been promoted despite objections for good cause, this has usually been a mistake. (More importantly, it's usually been later recognized as a mistake.) Objective: get the RFA voters to open their eyes and start actually evaluating candidates. Alternately, get people to recognize that it'd be useful to have a way to remove the bit without going to arbcom.

Examples of controversial crat decisions:

Requests for adminship/Carnildo 3

Requests for adminship/^demon 3 (see also Wikipedia Signpost/2008-02-25/Controversial RFA)

Requests for adminship/Elonka 3

Requests for adminship/Gwen Gale 2

Requests for adminship/Philosopher

Requests for adminship/Ryulong 3 - was eventually desysopped for doing the sorts of things the opposers predicted

Requests_for_adminship/Majorly - weird. he was apparently promoted after this with no additional RFA? how did that happen?

Requests for adminship/Ecoleetage 3 - clear and ample evidence of unsuitability of the candidate. will likely be closed by the numbers, as a success. Some unknown quantity of chat room buddies number among the supporters. what a catastrophe. retirement stalking accusation At least people see now that we dodged a bullet there. [last useful version] talk page discussion highlighting the incompetence of the rfa process

Requests for adminship/Foxy Loxy 3 - ongoing maturity concerns, even among supporters. yet, a big crowd of "great candidate, hooray!" voters turned up.

Other weird stuff:

Requests_for_adminship/Archtransit - not "controversial"- nobody saw this coming.

this travesty is good evidence that yes, things are screwed up with RFA. How can we discourage fan-club-ism and encourage actual evaluation of candidates?

Requests_for_adminship/IMatthew_2 - highly questionable rfa. some of the supporters, and even the nominator, had "concerns" yet they put their vote in the wrong column.

What happens when people ask crats to do something about a problem admin? We don't have much data. I have two ill-fated attempts at User_talk:Kingturtle/Archive6 and User:Raul654/archive14. All we can do is keep trying, backed up with better discussion and a good demonstration of a community lack of confidence. Altho, it appears that no argument can stand up to the objection of "But, we've never done that before."

Requests_for_adminship/Majorly_2 another case of misplaced opposes. Some of the people voting support come right out and admit he's a drama queen, and they support anyway? I strongly disagree.

Requests for adminship/Pastor Theo significant opposition for good cause, yet by simple vote counting it's successful

Requests_for_bureaucratship/MBisanz - successful, really? wow. only a plain-old vote count could come up with that result.

Requests_for_adminship/Steve_Crossin - a case of "la la la I can't hear you" among the supporters. bizarre.

Requests_for_adminship/Kww_3 should have succeeded on the merits of the arguments, but some of the crats were afraid to depart from simple vote-counting.

Repeat RFAs
If at first you don't succeed... Examples of repeat RFAs with questionable results. Does simply repeating the process lead to different results, through voter fatigue or something else?

Requests_for_adminship/Juliancolton_3

Requests for adminship/ST47 5 - maybe not controversial, but a good example of how all you have to do is keep running, and people will give up on opposing.

Requests_for_adminship/FlyingToaster_2 - no evidence of anything except trying again and making chat room friends. and now we have this, which, along with a bunch of drama-mongering from the usual suspects, may contain legitimate criticisms. it also points out the bad side effects of the whole misguided "you should write x articles to be an admin!" crowd. Combine this with young eager-to-please editors, and all you get is incompetence in writing. take people who probably know they're not really writers, and peer pressure them into doing it anyway? not good.

Requests for adminship/TenPoundHammer 6 - unlikely to be closed controversially, but a good demonstration of some problems. We've got solid evidence of the unsuitability of the candidate, yet the "he's my buddy, I support" votes keep rolling in. People who vote this way should rethink their participattion in RFA. This also shows that a huge bloc of voters who think "he tries hard" is a good reason to support. Evidence of the myspaceification of wikipedia?

Requests_for_adminship/Aitias_3 - all you need to do is keep trying?

Requests for adminship/The Transhumanist 6 bizarre

Requests for adminship/Timmeh 3 - yet another case of weirdness. even some of the supporters have concerns about maturity and judgement, yet they support. and, a bunch of people are saying really unreasonable things like "Oh no, you only waited 3 months! That's not good enough.  Wait another 3 months, and this by itself will make you a good candidate." More evidence that we need to restrict who can participate at RFA, because so many voters are so completely irrational.

principal of least surprise
Why don't crats have the technical ability to undo a promotion? I always assumed this was a historical technical accident, which later turned into a social expectation of behavior. Some people think it was a deliberate design choice, perhaps. Is there any evidence to support either conclusion?

crat overhaul
The simplest thing that could possibly work: Crats hand out admin tools liberally to those who seek them. There should be some opportunity for community objection so that editors who've already displayed problematic behavior don't get them, but nothing resembling RFA. Perhaps some objective standards in terms of (sigh) tenure or (double sigh) edit count. Or maybe they need a nomination from a trusted editor who's willing to vouch for them.

Then, the crats should remove them either 1) upon seeing a prima facia case for misuse of the tools or maybe 2) upon community showing a lack of confidence in the admin in question. Perhaps #1 is already covered by emergency desysop procedures and the crats should focus on 2. I don't think 1 is adequately covered- this happens only in extreme cases. Negligence or incompetence should get the bit pulled too.

This is too big a change
Too radical; can't be done in one step. Second best(?) option: Get crats to take responsibility for their promotions. If a promotee goes off the tracks, the crat tries to fix it. But how? Worst case: The crats asks arbcom to ask a steward to desysop, by saying "I promoted this guy but have since then changed my mind, given new evidence. I don't have the technical ability to do it myself.  Wanna help me out?" Would arbcom go for it? (Would a steward go for it directly? Maybe, but it's be best to OK this with established authorities.)

Could an individual crat do it differently?
This is questionable. It's going to be seen as a problem if different crats have vastly different standards. How a given RFA is closed is going to depend on who grabs it first. This is bad, but then again it's good if the other crats are OK with it. Could allow a more gradual change in process- there could be an unspoken agreement between crats. What if the other crats don't like it? We don't really know what happens much if crats disagree on crat matters- it's pretty rare (or else it's not very public).

Possible platform
One way to approach crat overhaul is for someone new to run for crat, along with a platform of "here's how I plan to do things, if given the ability to do it." The downside is this blends things together- people will be voting on the new crat and the new process at the same time. Some people may like one but not the other. It may still be worth trying, though, as a way to cut through the stagnation that otherwise gets in the way. A simple platform for overhaul may be: "I will promote whoever I think is a suitable candidate, without RFA. Then I will take personal responsibility to ensure that the new admin stays on track.  If they show themselves unsuited to the tools, I'll ask a steward to demote them.  The steward should be able to see that there's no consensus for them having the tools in a case like this, so the demotion should be uncontroversial." Could it possibly really be this simple?

What to focus on today
The de-adminship side of things is too difficult to move forward with. Perhaps the best a crat can do is mentioned above- recommend that a promotion be overturned. It's possible this could become unnecessary over time if the simple application of social pressure becomes effective. What a crat can do right now today is try to be conservative in promotions. Working with the other crats is probably necessary here. Not all of them see it as their duty to try to make sure unsuitable candidates are not promoted. Is it worthwhile trying to encourage them to see this as a proper responsibility of a crat? Anything that smells of "protecting the community from itself" will be subject to unpopularity and cries of elitism.

blah blah
(post this somewhere useful when it's more baked) What is the role of bcrats? Well, they're the only ones who make new admins and bcrats. (The other duties are not controversial as far as I know).

Why don't we just have admins close RFAs? Well, admins have software functions we like to protect. Therefore, we protect who becomes one. (Apparently admins wouldn't be trusted to judge consensus? I'm not so sure about that, they do it on Afd, which, granted, is less of a big deal) If there are so few bcrats and they guard important access, I think we should be able to reasonably depend on them being rather conservative with how they wield their abilities.

Also note that other bcrats cannot undo a promotion. They could ask the stewards, but wouldn't the stewards be in an odd position if some bcrats told them to promote and other bcrats told them to undo it? So, while almost always, what's done can be easily undone, perhaps this is less true in this case? Again, this sounds like all the more reason for bcrats to be conservative. (But if the crat who promoted asked the steward to undo the promotion, this should work.)

So, famously in the somewhat recent Carnildo case, the bcrats were way less conservative in the use of their power than they'd previously been. This was a surprise to many.

Should they be conservative or not? Should they rely on feedback from non-bcrats to decide this question, or should it be all amongst themselves. Those who may "vote" on future RFBs want to know, I would guess.

newer older stuff
Crats judge consensus for adminship. Consensus can change. Why have we ignored the obvious consequences of this till now? There's a cultural resistance to the idea of a de-adminship process, but from where I sit, it's clearly seen as necessary also. Why the insistence that only arbcom or Jimbo can do this? Is it because they're the only ones who ever have? Big fucking deal! Every established process had a first time- that's how it became an established process.