Wikipedia talk:Discussions for adminship

All the old pages have been archived here. Please take a look to see the developing ideas, and also what went wrong. I am about to archive this page and we're going to do something!!! — Ilyanep (Talk) 19:39, 15 April 2006 (UTC)

Okay...

Okay we're starting from scratch.

Here's what we need to figure out before we start writing anything:

• We've pretty much settled on the xx day discussion, xx day voting period.
  • How long will each be?
  • Will the discussion be a plain talk-page like discussion or more of a narrative?
• What are we going to put on what pages, so that it's not a stinking dump like WP:RFA is?
• How can we reduce the workload for Bureaucrats?
• Suffrage
• The name -- Discussions for Adminship may have been a bad idea.
• Attempt to fix issues with RfA (some of which are listed below)
• Minimize Instruction creep

(Items in italics added 22:12, 15 April 2006 (UTC))

Anything else? — Ilyanep (Talk) 19:43, 15 April 2006 (UTC)

BTW, In case you don't know what this was founded upon, see the old pages (linked to up at the very very top of this talk page) and archives to see where we started. — Ilyanep (Talk) 19:48, 15 April 2006 (UTC)

Is there a large workload for bureaucrats? People don't seem to think so. Maybe this is a post-Cecropia phenomenon. If there's a large bureaucrat workload and we need more bureaucrats, I've certainly shown that I'm willing to do the job. Andre (talk) 19:49, 15 April 2006 (UTC)

What I'm mainly concerned about is that there's so many instructions; one has to have 5 tabs open to promote a candidate. Although it's not that much, it adds up. — Ilyanep (Talk) 22:08, 15 April 2006 (UTC)

I am personally in favour of preserving the RfA-style system for voting (which is what I think we're doing), but adding a period for discussion beforehand. I think that the discussion should be talk-page-like, but that when discussion concludes, the bureaucrat should summarise the salient points for either side. Whether others should be allowed to edit the summary is a good question, but I'm kind of inclined against it -- instead they should ask the 'crat to modify it. This is instruction creep and more work for 'crats, though, so it's a downside. (It's also something for Durin to thwack us on the head with for unnecessarily complicating the adminship process. :p) We should just stick to the RfA name, since I don't see the need to change the title. Johnleemk | Talk 19:51, 15 April 2006 (UTC)

That was essentially what we were aiming for, but we need to work out the details. — Ilyanep (Talk) 22:08, 15 April 2006 (UTC)

This policy won't succeed unless you find a way to implement it with minimum instruction creep.--Urthogie 19:53, 15 April 2006 (UTC)

That's the thing. We're trying to minimize instruction creep so that it's even less than the current RfA. — Ilyanep (Talk) 22:08, 15 April 2006 (UTC)

Here's what you haven't figured out...

In the above you state "Here's what we need to figure out before we start writing anything:" and in that list is nothing about any effort to determine what is wrong with RfA. My objections to this process therefore remain the same. --Durin 20:37, 15 April 2006 (UTC)

I'll have to agree with Durin. Before trying to "fix" something, please explain what's wrong with it. User:Zoe|^(talk) 22:03, 15 April 2006 (UTC)

A huge amount of instruction creep. m:Polls are evil. People bringing up false accusations, votes piling on, then the accusations being rebutted but nobody ever comes back to change their vote. Ridiculously increasing high standards. Trolls. Rouge admins. We need to find a way to fix all of those, and hopefully with very little instruction creep, and being able to merge it into the process we currently call RfA. — Ilyanep (Talk) 22:09, 15 April 2006 (UTC)

...and if anyone else has any problems we need to look at, please tell us. — Ilyanep (Talk) 22:11, 15 April 2006 (UTC)

Here're some more: Use of RfA as a soapbox (Boothy443, anyone? And others as well), and RfA is starting to be widely seen as merely a popularity contest. — Ilyanep (Talk) 22:14, 15 April 2006 (UTC)

That sounds about right. And suffurage would sound lovely as well. I personally think about 2000 would be great. -Zero^Talk 22:29, 15 April 2006 (UTC)

• What's the goal of RfA? To weed out people who probably would not make good administrators. That's it. That's the only goal. How do we measure how successful it is? Answer that, and then derive the results. Here's one way; how many admins have been forcibly de-adminned since the establishment of WP:RFA in 2003? Very, very few. In the meantime, 750+ admins have been created using the WP:RFA process. If we say 10 admins have been forcibly de-adminned (I think it's less), then that is a WP:RFA mission failure rate of 1.3%. WP:RFA is 98.7% successul. What is it you were saying is wrong with WP:RFA again?

• Trolls: And how has that affected the success of WP:RFA exactly?
• Rouge admins: Very few have been forcibly deadminned. Doesn't seem there are too many rogue admins.
• Polls are evil: RfA has been a poll since June of 2003. Seems to have worked just fine for hundreds of admin nominations.
• Increasing standards: I'm collecting more data on this. However, I think the notion that standards keep going up and up is false. Guess what the average edit count was for the first 20 nominees on RfA? Approximately 4900. An illustrative chart on this is this: Image:AverageEditCountatRfA.png. Over the last eight months, the average edit counts of nominees who have been rejected has only slightly risen.
• Instruction creep: It's pretty simple. The section in the front matter that covers how to vote is rather short. The instructions to nominate are also pretty simple. It remains a fairly simple process.
• False accusations: Do you have evidence that shows a number of RfAs being undermined by false accusations or are these isolated incidents?
• Votes piling on: Uhm, people are supposed to vote, are they not? What would you suggest as a remedy to throttle the number of votes and why would this be a good thing?
• Nobody coming back to change their vote: Again, do you have evidence that shows a number of RfAs being undermined by people not coming back to review additional information and perhaps modify their vote or are these isolated incidents?
• RfA as a soapbox: I think we've done a rather good job of weeding out the people who are using it as a soapbox. Boothy443 as a reason to redo RfA doesn't seem particularly strong.
• RfA as a popularity contest: Widely seen? Interesting. Any data to back that up? If it's a popularity contest, then why is it that standards are, according to you, going UP and not DOWN? Surely it being a popularity contest would have a downward effect on standards, not an upward one.

• I have noted on many occasions people indicating that RfA has this, that, or the other problem. I have been keeping statistics on RfA for quite some time now, and have firmly rebutted in fact what people claim is happening at RfA. To further understand what RfA was and is, I have begun collecting data all the way back from when WP:RFA began in June 2003. I have already found evidence that undermines a long held thought, that WP:RFA used to be self-nomination only. Utterly false. The majority of early noms were self noms, but 20% were not. More data needs to be collected, but I would not be at all surprised to see standards haven't changed all that much since 2003. If standards were skyrocketing as much as people continually claim, we'd be at 10,000 edits and two years as a user now possibly. No, I think the crux of this protest isn't that the standards keep going up (which I doubt) but that the standards are too high for some people to accept. C'est la vie. This is a community based effort, and sometimes our own ideas of how things should be will not mesh readily with how things are at Wikipedia. --Durin 22:43, 15 April 2006 (UTC)

As Wikipedia grows in size, there is less chance that every active Wikipedian will know each other's work. IMO, we need to continue to assume good faith about editors unless there are obvious behavior problems. The process in place now, RFA, follows this model. I don't think any changes are needed. --FloNight ^talk 23:24, 15 April 2006 (UTC)

Durin, the problem I have with your analysis is not that it's wrong, as far as it goes. You've ably demonstrated that the current process works well enough. The problem I have is that your analysis doesn't address whether the process could be better. ... I'm thinking about some rather different (and much harder to quantify) metrics than admins forceably deadminned or edit counts, such as "how efficient is the process" and "how happy are the editors that don't get selected", to pick two. (there are others)

1. I spend a fair bit of time analysing candidates. Under the current system, my analysis isn't leveraging anyone else's much and it isn't leveragable by anyone else. The DFA idea in part is to package up this analysis so that others can benefit. Is it workable? Maybe it isn't. Is it measurable? Maybe it isn't. But I don't think it's something yet quantified.
2. I am worried about the effect on premature or ill advised candidacies on the retention of otherwise good editors. Getting ripped apart (even validly, even civilly) is not pleasant. In a perfect world, with perfect people, no one would leave in a huff, and it's true if we had an infinite supply of editors maybe we wouldn't care about those that did. (harshly: we would be better off with out them) Is this a perfect world? How many good editors do we lose after failed RFAs? How many people do we lose that are turned off by stuff in general? Dunno. Is it measurable? Maybe it isn't. But I don't think it's something yet quantified.

So then... In my view we have a process that's "good enough" now. But "better" is the enemy of "good enough". I think that even if this current process works pretty well, there may nevertheless be better ones out there. Will the DFA proposal get us to them? I'm starting to have my doubts that it will, but it's not because it isn't a good idea to front load analysis and discussion and have a nice package of stuff for "nonvoters" to look at, it's because changing a process that works is hard, the hurdle to do so is high, and unlike a broken process which can only be made better by change, a working process could well be made quite a bit worse by change. ++Lar: t/c 00:27, 16 April 2006 (UTC)

Precisely, Lar. Nobody has said that the RfA process is fubared. What we have said is that the RfA process could be improved. I have brought up nominations before on WT:RFA where people piled on, and when the facts changed, didn't alter their votes. No, this doesn't matter much, but the situation could be improved. (IIRC the last time I brought this up, all Durin said was "I disagree.") The only question for me is whether or not the extra effort and instruction creep involved in a DfA (or new RfA, or whatever we call it) will outweigh the advantages of DfA. I don't think it's that much of a big deal, though.

Here's a new idea: Retain RfA. However, double the current time for an RfA to run to two weeks. The first week is for discussion and for the candidate to accept. Discussion takes place on the talk of the RfA subpage. The main page will be the standard template, except it has the nomination, etc., a prominent link to the talk, and a list of salient points summarising reasons for and against the candidate having the mop and bucket. Anyone can add to the list of points, as long as the point being brought up has appeared on the talk. If this point is disputed, the disputer may either remove it for further discussion, or add a counter-point (as long as it is also raised on the talk). This sounds complicated and instruction creepy, but it's not. To me, it's the natural thing to do. After a week, we move to voting, although discussion may continue. The voting page should retain the prominent link to the talk, and the summary section should maintain that if voters want to be informed and have the full picture, they should read the talk.

One concern would of course be that voters might get fatigued of discussion, etc. But it's worth a shot. I'm going to go ahead and add my proposed format to the page. Johnleemk | Talk 05:56, 16 April 2006 (UTC)

• Johnleemk, I have repeatedly detailed my objections well beyond "I disagree". People insist RfA could be improved. Fine. It could be improved. I've never said it's perfect. I've never said it couldn't be improved. The fact is, it's a very good system. Without a concerted effort to diagnose what is wrong, backed up with verifiable fact, there's a very high likelihood that any new system will make it worse. Above, Ilyanep outlined several things that he feels is wrong with RfA. It's opinion. Opinions are fine. But, I prefer not to base a new RfA on opinion when the existing system is doing a rather incredible job of weeding out those who simply aren't ready to be administrators.
• As to your suggestion, WHICH page are you intending on adding it to? There was an earlier attempt to ramrod DfA down the throat of RfA without any notice or discussion on WT:RFA. I do hope you're not planning a similar move?
• Your idea might have merit, it might not. But (and I'm rewinding the tape recorder...hang on...just a few more seconds...) there's still no consensus on what's wrong with RfA, nor any idea of how this idea addresses any of the supposed shortcomings. I could just as well suggest that we reduce RfA to 2 days worth posting, and have people say they second the nomination (as opposed to voting). Guess what? That's the way it used to be here. I therefore assert that my suggestion has at least as much merit as yours does. Problem is, neither has any connection at all to addressing what the supposed shortcomings of RfA are. Also, I've noted before that having a discussion is at least somewhat likely to lead to edit wars as people attempt to portray their favorites/anti-favorites in a particular light. --Durin 17:28, 16 April 2006 (UTC)

The only time I brought up the RfA could be improved point with you, your response was just that: "I disagree". Perhaps you could link us to the relevant WT:RFA archive? Altering an existing and important process is never a good idea. Where did you think I was going to add it to? (Hint: It starts with a D and ends with a p.) Anyway, if there's no consensus on what could be fixed with RfA, what's wrong with me trying to build it through discussion? Your argument seems rather circular: There is no consensus on what could be improved on at RfA -> Johnleemk is giving his two cents about what could be improved, contributing to a possible consensus -> Johnleemk's action is wrong because there is no consensus. Johnleemk | Talk 17:41, 16 April 2006 (UTC)

• Please review the archives of this talk page for extensive commentary from me. The argument is not circular. Perhaps I am, despite my efforts, not being clear. Several people, yourself included, keep coming up with grand ideas to 'fix' RfA. Yet, nobody has yet come up with any reasoned list of what is wrong with RfA that needs fixing and been able to back it up with verifiable fact. It isn't circular at all. I'm trying, apparently vainly, to convince people that we should go about this in a rational, well thought out way. Instead, there's an apparent over riding interest in proceeding ahead with whatever ideas we can come up with, and to heck with any analysis of what's wrong with RfA and whether or not the bright shiny new ideas will do anything to solve those problems. It's like people saying "Hey let's make this car go faster!" and the response being "Ok, let's put square wheels on!". When I note the failure of that mode of thinking, I'm taken to task for not recognizing that RfA could be improved. Sigh. Me thinks I'm talking with the Golgafrinchan marketing execs ;)

• Bottom line, I'm just once voice here. All of you are bent on replacing RfA. Ok. Go for it. I strongly suggest you get consensus at WT:RFA before attempting it again. I also think any effort along these lines will utterly fail if you refuse to do any analysis of what's wrong with RfA and how the new system will fix what's wrong and not generate a slew of problems itself. But, I've voiced these opinions multiples times before apparently to little effect. So, I'll stop wasting your time on this :) Have a great day! --Durin 20:01, 16 April 2006 (UTC)

suggestion

I know wikipedia is all about consensus, but you'll probably succeed most at creating this proposed process by starting it out concrete, and letting us hammer away at it, and seeing if it can survive consensus. Just my experience, --Urthogie 22:30, 15 April 2006 (UTC)

What is the problem?

What should the focus be here? Is RFA producing acceptable admins in a process that needs improvement? Or is it producing a non-trivial number of admins that should not have been accepted?

I know what I think, but if the admins getting promoted are acceptable is the process so bad that it needs an overhaul? Personally I think we're promoting admins before they are ready. I think it's become a popularity contest. But...I'm curious what other people think about which needs the work, process or product? Rx StrangeLove 06:37, 16 April 2006 (UTC)

Well yes, it's a popularity contest. Also, I think Lar raises a few good points above. — Ilyanep (Talk) 15:55, 16 April 2006 (UTC)

• Ilyanep, you haven't answered my response to this above. If it's a popularity contest, then there should be a downward trend in standards. You assert there's an upward trend. So which is it? Please note that I've yet to ask who DfA (in any form) addresses these issues. DfA won't stop it be in a popularity contest, if that's what you think it is. It also won't stop trolls, rogue admins, and etc. I am at a loss as to understand why there is so much resistance to wanting to clearly identify and verify what problems RfA has as a basis for developing a system to replace it. --Durin 17:32, 16 April 2006 (UTC)

• Why does it necessarily mean that standards have to go down for it to be a popularity contest? In fact, it seems the other way around (and perhaps popularity contest isn't the best term, perhaps a good ol' boys club is more fitting). People are expected to be here for a year, to have more or less equal contributions in all namespaces (some people on 1.4.06 were jokingly voting oppose for 'not enough portal talk namespace edits', and this requirement isn't easily measurable), oh and god help you if you've ever stuck your neck out to do anything around here and ruffled some feathers around here. At least with the discussion period, people can see what people's exact concerns are. Which is another point. I frequently find that I don't know most of the candidates running (I only ever know one or two people running at a time), and the discussion would help me as a voter decide where to go. The current format is not formatted for discussion, it in fact discourages it.
• Either way, rest assured that I won't try to shut down RfA again. Perhaps we can try some time to run RfA and DfA parallel, but only when we're absolutely sure that DfA is refined enough. Yes, RfA is good enough, but so was asking for sysophood over the mailing list. No, it hasn't caused major problems yet, but could we do it better? Yes. And this is also our opportunity to reduce the instruction creep that is prevalent on RfA (CSCWEM's 2nd nomination was a victim of it). — Ilyanep (Talk) 20:38, 16 April 2006 (UTC)
• By the way, Johnleemk made a good starting write-up for the main page. — Ilyanep (Talk) 20:41, 16 April 2006 (UTC)

Why we need discussion seperated from voting

If a nominee has ever been involved in a "controversial issue", then it will sink their RFA, and possibly any to follow. Example: people caught up in the userbox issue. What we need is time for discussion, followed by time for voting, and voting should not include any comments; that's what the "discussion" section would be for. Alphax ^τεχ 04:02, 18 April 2006 (UTC)

If there's no comments on the "voting" or poll section, then it's not leveraging consensus anymore. At which point -granted- it'd still not be *entirely* pointless (I'll concede that much), but it'd definately be a lot less useful. Kim Bruning 08:04, 12 May 2006 (UTC)

• I don't see how discussion would help with an issue like that, exactly. The reason that major involvement in controversial issues is bad for an RfA is because it earns you enemies. Those people are pretty likely to vote against you regardless of any discussion. Christopher Parham (talk) 07:16, 15 May 2006 (UTC)

RfA Floor

Hi All - may I have your views on a possible floor for requests for adminship, in order to prevent the onerous processing of inherently futile RfAs? Virtually every week there are some good faith and poor faith nominations of inexperienced Wikipedians, many a times by naive nominators. For example:

Floor Proposal

1. Users with fewer than 1,500 edits may not be permitted to create or receive a nomination.
2. Users with under two month's presence on Wikipedia may not be permitted to create or receive a nomination.

These qualifications are based on the basic consensus that has emerged amongst Wikipedians on what qualifications and experience an admin should possess. While it is clear that good potential admins are often identified below these limits, what is the harm in waiting for 200-300 more edits, or a few more weeks? Looking forward to your views. Rama's Arrow 07:57, 12 May 2006 (UTC)

An advantage of these qualifications is that

1. It encourages the aspiring, yet young and naive users to study Wikipedia policies and work with the community rather than getting discouraged or getting wrong ideas.
2. It builds more respect for the process - no joke nomins, so people take everything else seriously. Rama's Arrow 08:02, 12 May 2006 (UTC)

Actually, if someone has been around and done 500 to 1000 edits, you generally already have a good idea of what they'll be like.

It's also extremely difficult to thoroughly review over 1500 edits, and people have just about given up on that.

So ironically, objectively, admin standards have been eroding ever since people started edit counting. Kim Bruning 08:19, 12 May 2006 (UTC)

Imposing arbitrary limits on requesting adminship is instruction creep and, to be honest, a bad idea. Amgine, one of the best editors of Wikimedia, couldn't have had an RfA under these limits! --Rory096 04:53, 3 June 2006 (UTC)

Stevage's recent edit

While I appreciate the cruft cleanup, the fact is that anything short of driving home several key points again and again gives DFA's detractors a chance to snigger and say, "See, DfA is wrong!" Two of the most common objections WRT DfA are "RfA isn't broken" and "You're using theoretical possibilities/one-in-a-million cases; show us some examples". I think we at least ought to provide examples, because they will come up at a future point in time; they already have many times before. Johnleemk | Talk 12:26, 12 May 2006 (UTC)