Jump to content

User:Linuxbeak/RFA Reform

From Wikipedia, the free encyclopedia

The goal: By the end of this week, I would like to see a rough draft of a modified RFA system. If the community agrees, we will take the new RFA out for a test drive, and then let the community decide on whether or not to keep it.

Note: Make sure you sign your comments. Also, if a comment already exists, don't just go "me too" and sign it; make sure you write out the entire thing. No half-donkey comments :-) Linuxbeak (drop me a line) 17:54, 17 January 2006 (UTC) See also Wikipedia:Admin accountability poll.

Analysis

[edit]

See User:Linuxbeak/RFA Reform/Analysis.

What is good about the existing RFA?

[edit]

Simplicity

[edit]
  • Simple process. Users just provide a reason (in theory) and sign their comment. Carbonite | Talk 17:54, 17 January 2006 (UTC)
  • Even though RfA is not intended to be a vote, people are familiar with the concept of voting to support or oppose a candidate. Because of this, users would likely be more comfortable with the current process than one which was more focused on discussion and consensus. As the history of RfA has shown, a vote can often be a decent approximation of consensus (whether it exists or not). Carbonite | Talk 18:27, 17 January 2006 (UTC)

Effectiveness

[edit]
  • Very low rate of false negatives (promotion of unsuitable admins). Only a few admins have ever been de-admined. Carbonite | Talk 17:54, 17 January 2006 (UTC)
    • That's an indictment on dispute resolution and arbcom processes, not praise for RfA. - brenneman(t)(c) 01:51, 18 January 2006 (UTC)
  • Relatively easy for bureaucrats to decide whether or not to promote. — Ilyanep (Talk) 17:56, 17 January 2006 (UTC)

Organization

[edit]
  • I like the sections (oppose, support, neutral, comments); it helps conformity, and it keeps opinions and views in an organized manner. Also allowes the closing beuracrat easier decision making regarding the nomination. -ZeroTalk 17:54, 17 January 2006 (UTC)

Administrative

[edit]
  • Lacks instruction creep.  Grue  18:04, 17 January 2006 (UTC)
    • I think I know where you're going with this, but could you clarify it a little more? Linuxbeak (drop me a line) 18:14, 17 January 2006 (UTC)
      • I like that simple usage of an over-used expression. To me (and not necessarily to Grue), it means this: That, by being very simple, very open and very transparent up until a bureaucrat closes, it is a very free process. A very free process is unlikely to make colossal errors very often since it is almost never hamstrung by its own rules. We don't accidentally exclude a good editor from making an important comment because they have 999 edits rather than 1000, for example. It also exposes those who are being trolls — and it doesn't matter if your RfA gets trolled since good users and 'crats should see through it (and if you can't swallow a troll, then what are you doing R'ing A?). Of course, the more sophisticaed trolls will slip by such sanity checks, but then they would do in any system. In its simplistic brutality, RfA can be pretty sure it is an effective process for weeding out mose of those not up to the task. As long as the 'crats get things right, too, of course. -Splashtalk 18:28, 17 January 2006 (UTC)
        • There's a problem with this. If the process is made subjective, then only those who are fit in the eyes of the community will be "promoted" (it's nothing special). This creates exactly what is going on right now. There appear to be at least a couple informal "alliances" of people who are entrenched in their own culture, and wish to perpetuate it. The very nature of the current process ensures that nobody will "rock the boat" and that change is less and less likely to happen with each new administrator created. Having a hard, defined threshold for when consideration is possible means that people can be judged on the basis of their contributions in addition to the opinions of their peers. Avriette 18:42, 17 January 2006 (UTC)

It was demonstrated that it works

[edit]
  • All other possible ways are not. If it is not broken ain't fix it abakharev 04:10, 19 January 2006 (UTC)
    • I'm afraid that's conjecture, not fact. Please provide solid facts. Linuxbeak (drop me a line) 12:52, 20 January 2006 (UTC)

What is not-so-good about the existing RFA?

[edit]

Low standards

[edit]
  • In my opinion, standards are lower than they should be. I expect admins to be at least somewhat familiar with process, because they will likely end up working with it. This familiarity is lacking in several of the recent candidates, that are probably best described as friendly and enthousiastic newbies. Radiant_>|< 17:52, 17 January 2006 (UTC)
Good point, but there is no good solution. Any hard requirements as the minimum number of support votes, or minimum requirements for voters, or other recently circulated ideas are not going to solve it. What would help things I believe is less tolerance for abuse of admin powers. Oleg Alexandrov (talk) 03:32, 19 January 2006 (UTC)
What about guidelines rather than standards? I like the idea of jury-style voting; for fairness and uniformity, and to support the voting committee, standards or guidelines will be needed. The problem with standards is that they tend to be (too) rigid. Lambiam 09:02, 23 January 2006 (UTC)
I think some of the standards are too low, and some of them are too high. Standards based on editcountitis are inherently flawed IMHO. That doesn't mean editcounts are useless as a qualifier, but only so much as they make the bar high enough that its hard for sockpuppets of malicious users to be accepted as admins. None of the admin actions are irreversable, and truely rogue admins have been historically rare, even when standards were much lower. Why WP:AGF doesn't seem to apply to the RFA process I don't know, but it does worry me. That said, we need to be looking more carefully at the potential for abuse, potential for error, and potential for benefit from each adminship candidate. Opinions and personal grudges need to be removed from said standards, and replaced by objective examination of a candidate's edit history (history, not count!), their dealings with others, and their participation in at least one process which would be aided by admin tools. - Stephanie Daugherty (Triona) - Talk - Comment - 10:09, 10 August 2006 (UTC)

Unfair to veteran [non-admin] users

[edit]
  • It has been said that long-time users that aren't admins yet, have accumulated too many enemies to ever have a chance at adminship, despite the fact that their behavior is not better or worse than that of existing admins. Radiant_>|< 17:52, 17 January 2006 (UTC)
    • While I agree with this statement in principle, I think that it misses a finer point -- that despite everyone saying adminship should not be special, or viewed as a promotion, it actually is, and is presently being used to drive struggles between "warring" cliques. I'm not sure what valid reason people would want to have adminship. Almost all the responsibilities they have can be accomplished through non-admin interfaces. Adminship seems to foster the need to crusade through the wikipedia looking for things to "fix". Ah, the saying I'm looking for is, "When you have a hammer, everything looks like a nail." Perhaps to correct this it would be good to either have term limits (so that a correction can be made if somebody loses sight of the "lack of specialness" that adminship has), or better screening during the RFA process to find out what a user is really seeking it for. "I want to help stomp out vandals and RC patrol" is not a very good reason. Avriette 18:48, 17 January 2006 (UTC)
  • The current procedure is tuned so as to make it easier for users with 2,000 edits to get promoted than users with, say, 20,000 edits. That's why unexperienced and clueless admins seem to predominate in the community.--Ghirla | talk 19:45, 17 January 2006 (UTC)
That's a silly comment. Yeah, you do have 20575 edits, but edits are not everything when deciding who should be an admin. Oleg Alexandrov (talk) 03:30, 19 January 2006 (UTC)
I was not talking about myself. Since I first came to Wikipedia ten months ago, I always repeated that I do not seek adminship and declined nominations. Neither do I state that the number of edits is important for a potential admin. I just say that it's easier for less experienced users to get promoted, and that's a fact. If you want a bet, I may start another account and get that account promoted within three to four months. --Ghirla | talk 15:55, 20 January 2006 (UTC)
it may be silly, but it's been empirically verified at least once: [1] -- nae'blis (talk) 17:04, 19 January 2006 (UTC)
Durin's data didn't verify nothing at all. It is all the same whether you have 3,000 and 5,000 edits IMHO. In my experience, those wikipedians who actually write a lot make much better admins than those who do not and therefore seek to interfere with those who do (e.g., Oleg Alexandrov). --Ghirla | talk 15:55, 20 January 2006 (UTC)
  • I updated the data to reflect nominations from June of last year through current. You can see it here . Also note that the success rate of admin nominees with more than 10,000 edits is 78%, which is inline with the success rate of 2-3k, 3-4k, and 4-5k edits success rate. This data disproves Ghirla's assertion that it's easier for someone with 2001 edits to get promoted than users with >10,000 edits. --Durin 17:39, 26 January 2006 (UTC)

Use of RFA as a soapbox

[edit]
  • People vote not on the candidates but to prove a WP:POINT. — Ilyanep (Talk) 17:56, 17 January 2006 (UTC)
    So true. Redux 02:07, 19 January 2006 (UTC)
I don't see a problem here. Oleg Alexandrov (talk) 04:14, 19 January 2006 (UTC)
??? Elaborate please? isn't the RfA on the person, not the process/policy/POV being soapboxed? -- nae'blis (talk) 17:08, 19 January 2006 (UTC)
  • There is a number of bad-faith editors who destroy great nominations by voting to prove a WP:POINT. The finest example I'm aware of is User:Oleg Alexandrov whose opposing votes are habitually acompanied by silly comments that the candidate doesn't use edit summaries frequently enough for his liking. IMHO such phony votes have nothing to do with the candidate's potential admin performance and should be discarded by Bcrats in the future. --Ghirla | talk 16:13, 20 January 2006 (UTC)
  • You might think someone who votes opposed based on a lack of edit summaries is voting in bad faith. You're certainly welcome to your opinion, but a number of people feel this is a valid criteria. It's fine if it's not a criteria for you; you're welcome to your own standards. I wish you would allow others the same leeway. You might want to have a look at User:Durin/Admin_criteria_comments#Edit_summaries. --Durin 17:55, 26 January 2006 (UTC)

Lack of efficiency

[edit]
  • While the process itself is simple, it takes bureaucrats 5 edits (I think) all on different pages just to close an RfA — Ilyanep (Talk) 17:56, 17 January 2006 (UTC)
    • Yeah, but once you get used to it, it becomes like riding a bike: you don't really need to think about it too much, just watch out for lampposts and slick spots. -- Cecropia 23:06, 17 January 2006 (UTC)
Do five edits, then regret a lifetime. :)
We need more bureaucrats then, to have individual people work less. :) Oleg Alexandrov (talk) 03:37, 19 January 2006 (UTC)
It's not that much work in the long run (such as to merit more bureaucrats at the moment) but it's just annoying (especially since I've just performed my first batch of renames since last August. — Ilyanep (Talk) 03:41, 19 January 2006 (UTC)
      • What Cecropia said. Especially using tabs, I can promote in about 5-7 minutes, and close without promoting in less. This is hardly the biggest concern about the system (at 5 minutes average, 10 RfAs a week is under an hour - compare that to the time used in, say, closing VfDs and other maintance tasks. -- Pakaran 04:04, 19 January 2006 (UTC)
        • Once again, it's not time that concerns me (5 minutes! Why that much?) Just having to have like 6 tabs open just to promote is akward at times. — Ilyanep (Talk) 04:15, 19 January 2006 (UTC)
        • Meh...I wrote a script to automate this.Voice-of-All 19:37, 9 September 2006 (UTC)

Impersonal

[edit]
  • I don't know many of the users recently popping up and it takes a long time to investiage. — Ilyanep (Talk) 17:56, 17 January 2006 (UTC)
  • Although Bureaucrats may have the authority to make a judgement call on any RfA, they almost always go by raw votes when the support percentage is lower than 70%75% or higher than 80%. For example, an RfA with 90 support votes lacking reasons and 10 well-reasoned oppose votes would almost certainly pass under the current system. This is an extreme example, but the main issue is that, in practice, bureaucrats really only use their discretion in the 75-80% support range. Outside of that, it's basically a simple tally of the votes. Carbonite | Talk 18:16, 17 January 2006 (UTC)
    • Good point. Linuxbeak (drop me a line) 18:17, 17 January 2006 (UTC)
    • Please, Carbonite, discretion is generally 75% to 80%, not 70%. One new Bcrat promoted someone at 72% after a different new Bcrat edited the Bcrat page to say 70%-80%. If we want Bcrats to use broader discretion, we can straw poll it to accumulate community opinion. -- Cecropia 23:09, 17 January 2006 (UTC)
  • Voters are not entitled to quote diffs supporting their opinion and vote, i.e., what they found so helpful or repulsive among the candidate's contributions. The current procedure leaves a door wide open for the trolls who start generalizing that the candidate is a "likely POV-pusher" or they don't trust him because he supported someone who later opposed them, thus misleading stray users who would vote "oppose" because the candidate is "too controversial if he generated so much opposition" and repeat the fables that we have "dozens other worthy candidates", while actually in some areas (e.g., Eastern Europe) dedicated contributors capable of a large-scale patroling work may be counted with fingers of my right hand, as they say in Russia.--Ghirla | talk 19:45, 17 January 2006 (UTC)
    • Can you clarify "not entitled to quote diffs"; do you mean not encouraged? -- nae'blis (talk) 21:10, 17 January 2006 (UTC)
      No, I mean, it should be required that everyone opposing a candidate should provide diffs to the nominee's controversial edits, not just generalizing that "the candidate is too controversial" or "is likely to push his POV", as we've seen in some recent votes. We need stubborn facts rather than one's opinions in order to judge the candidate's previous behaviour. --Ghirla | talk 07:52, 18 January 2006 (UTC)
      Ahhh, but I thought this section was talking about current practice problems, not solutions (as of yet). That's why I was confused; thanks for clarifying. -- nae'blis (talk) 17:07, 19 January 2006 (UTC)
Of course one better provide diffs for serious accusations. But the fact of the matter is that some people are indeed too POV pushing to make good admins. Oleg Alexandrov (talk) 03:39, 19 January 2006 (UTC)

Maintainance

[edit]
  • People are constantly try to fix it, wasting valuable editors' time and effort. (Not an attack on this page, just sayin' is all.) --LV (Dark Mark) 18:17, 17 January 2006 (UTC)
  • Indeed. RfA is almost always the subject of reform discussions. The current system has gone through the crucible of time and countless reform suggestions. It's done a pretty good job of evolving over time. I spend way too much time trying to get people to root their reform ideas in factual evidence rather than suppositions. See my comments above regarding the success rate of nominations for example. People can postulate what's wrong all they want. Refuting comments by basing such refutations in fact is time consuming. --Durin 18:00, 26 January 2006 (UTC)

Lack of standards for voters

[edit]
  • Absence of suffrage results in proliferation of voting spam on user pages both here and in the alternative wiki projects. --Ghirla | talk 19:45, 17 January 2006 (UTC)
Any standards for voters will be very hard to implement in a consistent manner. Besides, I don't think it would be fair. New users, unless they are sockpuppets, fully deserver to participiate in voting for admins, just like any other people. Oleg Alexandrov (talk) 03:42, 19 January 2006 (UTC)
As I said many times before new editors are too unexperienced to vote judiciously. I suppose that new editors come here to edit and to write articles. If they come to Wikipedia and start with voting, it makes a bad impression on me. Some sort of RfA suffrage is indispensable. --Ghirla | talk 16:19, 20 January 2006 (UTC)
  • Candidates who campaign usually have a negative impact on their RfAs, not a positive one. RfA regulars look very poorly on campaigning both on Wikipedia and #wikipedia. I find the voting spam comment to lack merit in a role of affecting the outcome of RfAs. --Durin 18:01, 26 January 2006 (UTC)

Herd mindset

[edit]
  • Many votes originate from the existing vote--that is, people vote because of who has already voted. Demi T/C 21:01, 17 January 2006 (UTC)
    • "Everybody likes a winner." -- nae'blis (talk) 21:10, 17 January 2006 (UTC)
    • I vote to support what Demi said. --Rudolf Nixon 23:08, 17 January 2006 (UTC)
Demi's right! From here on it is downhill, expect a bunch more comments agreeing with Demi here! Oleg Alexandrov (talk) 03:43, 19 January 2006 (UTC)
    • (after Edit conflict, darn it Oleg!) Support what Rudolf and Demi said — Ilyanep (Talk) 03:46, 19 January 2006 (UTC)
  • Agreed. But, there's no way to fix it which is community friendly. If you had a panel of people charged with reviewing candidates for RfA, cabalism claims would be made left and right. You can't force people to review candidates; they either do or they don't. I do, which is why I don't vote often; time consuming. When I make a nomination myself, I spend considerable time reviewing candidates. Few RfA nominators go through the same trouble. --Durin 18:04, 26 January 2006 (UTC)

Focuses on wrong qualities

[edit]
  • The current process favors editors who stay out of edit/revert wars or any contentious issue. Then they are expected to uphold bans, decide on vandal bans, 3rr bans, use discretion after arbcom decisions etc. RfA candidates shouldn't be expected to get involved in wiki fights, but they should show some familiarity and interest in dispute resolution when that is their most discretionary decision involving real people. SchmuckyTheCat 21:43, 17 January 2006 (UTC)
People involved in dispute resolution are very valuable as future admins. People who are involved in disputes, and behave badly, are not wanted as admins. :) So, being involved in wiki-fights is good on a future admin's resume, as long as that person behaved properly. :) Oleg Alexandrov (talk) 03:45, 19 January 2006 (UTC)
I agree, but people who were involved in wiki-fight have some history to be judged on. People who were involved only in noncontroversial topics (e.g. tropical fish) and avoided any confrontation maybe a perfect editors but their behavior as admins is absolutely unpredictable as admins deal mostly with controversial or at least confrontational. May be they would be perfect admins, maybe they will through a tantrum maybe they will continue to ignore anything remotely controversial (not the best scenario). abakharev 04:26, 19 January 2006 (UTC)
I'm with Alex Bakharev here. Being currently on RfA nomination myself, and having followed that of Ramallite (and a few others), I really get the impression that anyone who has a long history editing "controversial" topics, no matter how impeccable their behavior, has received a "kiss of death" for promotion. Well, that's an exaggeration: Ramallite squeaked by, and I might too. But generally, you can't try to promote NPOV on politicized topics without gaining "enemies" among those who want articles to push a particular and strident political POV. Lulu of the Lotus-Eaters 21:20, 19 January 2006 (UTC)
Absolutely agree with Lulu. --Ghirla | talk 16:29, 20 January 2006 (UTC)
  • I recall an admin who wrote excellent articles but eschewed editing controversial topics. Several weeks after promotion, he was dragged into a dispute, started frantic revert warring, got blocked and then was astonished to learn that Wikipedia had some sort of dispute resolution procedures at all. In other words, we don't have a method to decide whether a candidate has admin skills and really knows how Wikipedia works. Three questions practiced today are not exams if you know what I mean. --Ghirla | talk 16:29, 20 January 2006 (UTC)

Waves of oppose votes based on one edit

[edit]

In my time watching RfA candidacies here, I've seen several candidates whose RfA's were going well for the first several days but failed after one oppose voter brought up one diff. Often, the diff is of an old edit (>6 months ago), and sometimes even of vandalism they did while still a new editor. Most contributors are not perfect, and if one voter who researches candidates more than the other voters do uncovers a bad edit, a pretty good candidate can have their RfA fail rather quickly. I suspect this might be because some voters, while well-intentioned, still vote somewhat blindly and don't research what else the candidates have done. (I'm not calling out individual users on this, since I read votes rather than the user names they're attached to, but I have noticed this pattern on several failed RfA's during the last several months.) --Idont Havaname (Talk) 23:26, 17 January 2006 (UTC)

If that is a recent bad edit, one may question future admin's ability to judge things well. You don't want somebody as an admin who would abuse his tools even 10% of the time. That is, it makes a lot of sense to vote oppose for such a person. Of course, this is a rule of thumb, but I would be against overlooking major errors of judgment of a future admin, even if that is a single instance. Oleg Alexandrov (talk) 03:24, 19 January 2006 (UTC)
I tend to agree with Oleg here. --Ghirla | talk 16:33, 20 January 2006 (UTC)
  • A single diff causing the wreckage of an RfA certainly happens. But, there's good reason for it. There's some essay somewhere that notes that RfA isn't really an affirmation of positive editing; it's the opposite. "We haven't found anything bad, so we'll promote". Sometimes, single edits can be a peak behind the curtain, so to speak. It can give us more insight into the true attitude/behavior of the nominee than 1,000 edits by the editor. Case in point; I voted on NickBush24's RfA. Prior to my voting, it was 36-0 and well on the way to a very successful RfA. Then, I voted [2] oppose because of a diff [3] that I found that he'd made about a week before his RfA began. But, even if the diff had been from months before, I would have found it objectionable to treat people like that, most especially newcomers. I didn't have a problem with my vote and still don't. I was, however, shocked that his RfA collapsed after that point and finished up at 43-22. I think what that served to highlight more than anything was that quite a number of people who were voting were not checking his editing background, even through a couple of weeks into the past. --Durin 18:16, 26 January 2006 (UTC)

The voters

[edit]
  • Voters who like seeing their names in polls
  • Voters who have no concept of whether the person they are voting for has a clue what to do
  • Voters who want to see if they can push the poll numbers to a new high
  • Voters who say "nice guy"
  • Voters who say "won't do any harm"
  • Voters who oppose from spite
  • Voters who say "if User:Toadbrain says he's good, me too!
  • Your ad here -- Cecropia 23:15, 17 January 2006 (UTC)
    All those are definately problems (I'm tempted to edit 'your ad here' to add something humerous) — Ilyanep (Talk) 00:46, 18 January 2006 (UTC)
    • I'll take you up on that :) Radiant_>|< 01:10, 18 January 2006 (UTC)
  • Voters who don't say anything but simply oppose everyone
  • Voters who oppose candidates because they don't like the way the wiki presently works (regardless of whether the candidate had anything to do with that)
  • Voters who support everybody because their nomination is also up
  • Voters who oppose because the candidate once supported some user who they don't like
  • Voters who say "the candidate is too controversial" or "we have plenty less controversial candidates"
    • My turn.
  • Voters who oppose based on gender, race, religion (or lack thereof)
    Lack of gender? — Ilyanep (Talk) 02:58, 19 January 2006 (UTC)
    Probably meaning those who do not wish to reveal their genders.—Ëzhiki (ërinacëus amurënsis) 02:04, 20 January 2006 (UTC)
  • Voters who support in response to an oppose vote
  • Voters who blanket vote (support or oppose) without giving reasons
  • Voters who oppose for reason of "We don't need any more admins" (I've seen this, amazingly)

A small group can manipulate the vote

[edit]

A relatively small group of 15 or even 10 well-organized users can have a huge influence on the results of voting. Voting in the consolidated manner as oppose they can stop almost any RfA, voting support, they can significantly increase the probability of success. Maybe I am paranoid, but I am surprised how much interest some disruptive users (with very short history of content creation) have in the RfA process. The talk page of a user whose blocklist is almost of the same length as the list of contributions is filled with wishing success and congratulations for some admin-candidates and mustering oppose votes for the other. I would speculate that people who are often blocked and/or often fill frivolous complaints have vested interest in administrators. abakharev 04:45, 19 January 2006 (UTC)

Since I succeeded in derailing Halibutt's nomination after dozens people voted in support, I feel that Alex is right. Another example. Once I voted "neutral" and quoted a diff, which led four or five other people to vote either "neutral" or "oppose", citing that diff as their reason. Then I received apologies from the candidate and voted in support, but the votes of others, based on mine, remained valid. Moreover, other editors later opposed because the candidate appeared too controversial now. --Ghirla | talk 16:41, 20 January 2006 (UTC)

Encourages vote counting

[edit]

While vote counts can be important, they are too often treated as the most important metric of the current nomination. The problem is that all votes are generally treated as equal, when they definitely should not be. While obvious sockpuppets are excluded, other factors should be taken into consideration as well. See the "problems with voters" section for a more complete list, but "me too" votes, "why not" votes, no-reason-given votes, "I support|oppose everyone" votes, votes from people themselves unfamiliar with Wikipedia policy, etc. are too often given the same weight as more substantial votes from known users who give valid reasons for their vote. Turnstep 14:27, 22 January 2006 (UTC)

Different ways of doing RFA?

[edit]

RfC style changes

[edit]
  • As per WP:FAC, create a section for "arguments in favor" and another for "arguments against". No voting. Radiant_>|< 17:52, 17 January 2006 (UTC)
    • I agree, either this or a customized WP:RFC page for RFA's. No voting. Rx StrangeLove 17:59, 17 January 2006 (UTC)
    • The cons to such a method might be that the bureaucrats have a greater say in that case, and CABAL conspiracy theories start popping up. — Ilyanep (Talk) 18:01, 17 January 2006 (UTC)
      • I'm inclined to agree. In theory, this proposal is all well and good, but obvious problems begin to emerge if you stop to think about borderline cases. Even well-meaning bcrats are likely to be put in a position of earning public disfavor and sparking counterproductive controversy. Plus, I'm concerned that the increased complexity here seems to be a guaranteed recipe for creating an unworkable logjam of RfAs. – Seancdaug 21:10, 17 January 2006 (UTC)
  • A spinoff of the FAC method; people provide agruments for and against for a certain amount of time, then a period of voting. — Ilyanep (Talk) 17:55, 17 January 2006 (UTC)
    This is the model I've supported for some time. Compare the second Kelly RfC to the first. If we had say five days of "evidentiary" editing, followed by two days of "polling" I'm sure we'd get more nuanced outcomes. If we had a template to fill in, "evidence of understanding policy", "evidence of dispute resolution", "evidence of incivility", that would be good as well.
    I know two days of polling isn't long, but a if the facts are laid out people could make a more informed recmomendation in a shorter time, and b anyone interested would probably have been watching the page anyway. I'd bet that the S/O/N ratios don't change much past the first two days except for those unfortunate cases where a single diff brings a nom crashing down.
    This would also help alleviate those. Often a bad diff shows up, theres a pulse of opposes, but then some sensible discussion nullifies the diff but not the oppose votes. They don't come back to look at the chat, and the nom fails. This would allow any poison to be purged before voting started.
    brenneman(t)(c) 02:26, 18 January 2006 (UTC)
    This actually seems very reasonable. On reflection, I can't figure out how to 'tally' RfA discussions in a purely FAC-like system. When you're editing an article for Featured Status, you can go and make changes/improvements so that the article now meets criteria; it's a forward-looking process. By contrast, most RfA behavior is backward-looking, at bad/controversial edits that may have been made in the past and cannot be changed. "Absence of evidence is not evidence of absence", and all of that. I'd change it to 4 days comment and 4 days voting or something similar, but I like it. It makes for a pretty good incremental improvement over the current system, without throwing out the old. It also may temper some of the snowball effect both on the oppose and support sides, and gives newer voters something to base their opinion on. -- nae'blis (talk) 22:18, 19 January 2006 (UTC)
    I've place a proto-template at User:Linuxbeak/RFA Reform/Adminship nominee discussion. It might be required to have it be two transcluded sub-pages, one protected at a time. Edit away everyone! - brenneman(t)(c) 06:35, 21 January 2006 (UTC)
  • How about a purely evidentiary method, with the ability to override certian findings through consensus? - Stephanie Daugherty (Triona) - Talk - Comment - 10:19, 10 August 2006 (UTC)

Tuning the voting

[edit]
  • Providing clear reasons and concensus for one's standpoint is a must. Any votes without any elaboration are discounted. And "half-ass" votes such the classics, "I never met him", "I don't like him", "I know him from another website" are to be discounted as well. Diffs should also be strongly implemented to back up one's POV. -ZeroTalk 17:54, 17 January 2006 (UTC)
    • Yeah, I would agree, though while I would encourage diffs, I'm not sure I'd require them. – Seancdaug 21:10, 17 January 2006 (UTC)
  • Suffrage. 1000 500 edits or more. No ifs, ands, or buts, no conditions. Keep out the trolls and clueless newbies. ([4]) -ZeroTalk 17:54, 17 January 2006 (UTC)
Note: Until this weekend, I had less than 1000 edits. I'm atypical (a relatively casual wikipedian who frequents project space), but I had a grasp of what was going on around here by 700 edits or so. I'd think 500 would be sufficient to tell someone is not a newbie, and really, if we require reasons, wouldn't cluelesness correct itself? -- nae'blis (talk) 18:09, 17 January 2006 (UTC)
Note: I respectively change my opinion on suffurage then. -ZeroTalk 18:24, 17 January 2006 (UTC)
Thank you. It just felt funny that ArbCom election requires only 150 edits, and you were talking about one or two thousand. -- nae'blis (talk) 18:56, 17 January 2006 (UTC)
Yes, wouldn't this make it seem like Rfa is more important than the ArbCom election? — Ilyanep (Talk) 19:13, 17 January 2006 (UTC)
  • No rfar voting. Strong POV and "revenge" basis. Best to stay in the white. -ZeroTalk 17:54, 17 January 2006 (UTC)
  • No "spite votes" ([5]) -ZeroTalk 17:54, 17 January 2006 (UTC)
  • When an editor is (politely) inquired to explain his/her standpoint, viewpoint, an elaboration must be necceasary. However, hounding is unacceptable. -ZeroTalk 17:54, 17 January 2006 (UTC)
    • Please be careful over use of the word consensus. How does one provide a 'consensus' for one's standpoint? The process itself generates a consensus, and a consensus cannot, by definition, be rested on a single editor's opinion.
    • There is not always a need for diffs. It is, e.g. adequate to oppose someone as being too new (imo) and that has no diff possible.
    • Editcountitis can be fatal. It's no good as a raw, unadulterated measure in a forum like RfA since it takes no account of the quality of those edits.
      • I agree. Having a minimum amount of edits for sufferage is risky at best. If there *was* to be some type of sufferage, make it something like 50 edits or at least a week of experience. Linuxbeak (drop me a line) 18:09, 17 January 2006 (UTC)
    • What is RfAr voting? All but about 12 of us cannot vote in RsfAr.
      • I'm confused on this one as well. Linuxbeak (drop me a line) 18:09, 17 January 2006 (UTC)
      • I'm pretty sure, from the discussion over at RfA talk, that he means users under the subject of an RfAr cannot vote. -- nae'blis (talk) 18:12, 17 January 2006 (UTC)
    • Elaboration cannot be necessary since every edit to Wikipedia is voluntary. If someone has already provided the 'elaboration' of your first bullet point, why should someone be able to invalidate their 'vote' by asking for further explication? -Splashtalk 18:06, 17 January 2006 (UTC)
  • Informal suffrage. Put in the rules that only votes of people with demonstrated history of contribution to Wikipedia who were not exceedingly distributive counts. Then require voters to introduce themselves and specify their main contributions to Wikipedia. Also the opposite side of the debate could provide examples showing the destructive behavior of voters. The closing bureaucrat can use the info to dismiss some votes, also the voters may use the info to decide if they want to believe the judgement of preceding voters. PRO: User with 10 edits can create a long useful article, maybe priceless contribution a voter with 3000 edits can bring little effort, nothing useful and a lot of distractions (e.g. by participating in revert wars), this format will make the distinction, a simple edit count will not. Contra: To much stress on bureaucrat - it is difficult to tell users that their contribution does not worth the problems you created, so your opinion does not count. Also the RfA could degrade into a flamefeast. abakharev 12:33, 21 January 2006 (UTC)

Nomination changes

[edit]
  • Many societies use a nomination committee charged with finding good candidates. If the rules are democratic, voters are not restricted to the committee's choices, but they may find them helpful. Perhaps--and I'm brainstorming here--we could create some voluntary committee or project devoted to finding and nominating good candidates for admin? Jonathunder 18:22, 17 January 2006 (UTC)
Absolutely not. Too much risk of POV pushing and more of that "Cabal" nonsense. -ZeroTalk 18:25, 17 January 2006 (UTC)
Well, it's a brainstorm: feel free to improve the idea. But you did catch that I said voters should be free to reject the nomination, right? Jonathunder 18:34, 17 January 2006 (UTC)
The problem is not just rejecting approved candidates, but having the option to approve rejected (by the NomCom) candidates. An open process means no one can be excluded unfairly from trying. -- nae'blis (talk) 18:56, 17 January 2006 (UTC)
Just as under Robert's Rules, nominations in addition to those of the nominating committee are always in order. Jonathunder 19:02, 17 January 2006 (UTC)
Ahhh, but you never specified Robert's Rules (in fairness, I know you weren't trying to present a full proposal in one paragraph). This makes more sense now, but I'm still not sure it's needed. -- nae'blis (talk) 20:00, 17 January 2006 (UTC)
  • some ideas from Lar... not all necessarily compatible with each other, not all necessarily practical here (although I've seen some used elsewhere) but intended as thought starters. Lar: t/c 18:41, 17 January 2006 (UTC)
    • ---> I was hoping some other people might comment too, these are meant as brainstorm starters. Should I move them to the individual sections where they apply, move them to the talk page, or not worry about it? Lar: t/c 16:27, 18 January 2006 (UTC)
    • The vote part of Suffrage should not be just a count test, it should be a demonstration of a good distribution. 1000 user page edits and nothing else shouldn't count for much.
      • I agree, but what kind of distribution did you have in mind? 100 non-minor article edits (just as a random number)? One way or the other, we need to come up with a broad guideline, if not hard-and-fast rules, to determine what qualifies as "good distribution." – Seancdaug 21:10, 17 January 2006 (UTC)
    • Competency test. Some sort of process/procedure exam to be given to ensure basic competency in areas (decided by consensus) are important for all admins to know. A "passing grade" would be required, or alternatively, the score would be provided as an evaluation criterion.
      • I can't imagine how this would work. You could probably improve the boilerplate questions that all admin candidates currently have to respond to, and doing so might provide a better picture of "basic competency," but a formal (or even informal) "test" just doesn't seem workable to me. – Seancdaug 21:10, 17 January 2006 (UTC)
    • Take an oath to uphold the Wiki. The last step after everything else is complete, is an affirmation of intent to abide by the 5 pillars, put on a page somewhere (with the corrolary that people could call admins on it)
      • I'm not sure this will actually mean much, in the larger picture, but I certainly don't think it would hurt, and is probably good principle. – Seancdaug 21:10, 17 January 2006 (UTC)
    • Recertify every so often (time driven at a min, or if the above call on not upholding happens). With 800 admins we don't have a shortage, do we? The recert would be a different process, involving providing several representative actions for evaluation. (the admin to supply some, and the community to supply some as well). Would not necessarily be a pleasant process!
      • I think this would probably create too much overhead. I would support a "vote" (if you'll pardon the term) of confidence in situations where the admin has been accused of misusing his or her position, but "term limits" seem to be a little too much. – Seancdaug 21:10, 17 January 2006 (UTC)
    • Come up with a different name for the process of providing input so people stop calling it "voting". :-)
      • I think we need to give up on this. Wikipedia may not be a democracy, but every time someone tries to propose a way to "de-democratize" our processes they seem inherently flawed, either because the proposals are just too complicated to be workable, or because they're far too open to (real or perceived) abuse. That said, I think "polling" (as in straw poll) is probably a better term, anyway. – Seancdaug 21:10, 17 January 2006 (UTC)
    • Scrap community consensus entirely (OK this one is radical). Formalise a cabal and let them decide.
      • Would kill the project outright. The idea of a "cabal" is not a particularly popular one, and it's not in the interests of the project to create a caste system. Doing so would only alienate newcomers and increase tensions between editors. In all honesty, this invites problems far more dangerous than any perceived problems with the current RfA system. – Seancdaug 21:10, 17 January 2006 (UTC)
    • (This one is REALLY radical!!!) Scrap consensus for this! Make it out and out voting, but votes are weighted by how much money you contributed to the foundation in the last year, or by money and time, or by how many articles you had make it to featured status, or any of several others (these all are getting at the idea of a meritocracy... the more useful contributors (by whatever metric, there are lots of metrics possible) get more say. after all corporation stockholder votes are weighted this way)
      • I don't think we can get away with scrapping consensus (nor do I think doing so would be a good idea), and, besides, I'd be worried that basing the process on financial contribution is a potential privacy concern. And judging based on number of articles you've contributed to that make it to featured status seems to do little more than transfer the problems of the RfA process to FAC process. – Seancdaug 21:10, 17 January 2006 (UTC)

Jury-style voting

[edit]

It looked stupid, but the more I think about it the more it make sense. In the courts of law the verdict is not done by voting of friends of the plaintiff and the defendant. Instead we randomly choose people to hear the proceedings and decide who is right. Maybe we should do the same with the RfA? Lets randomly select 10..20 people among wikipedians in good standing (say more than 500 edits in the article space since the latest block or ban). Lets all the interested parties to present their evidence and discuss the admin-candidate. Then lets the jury decide. The bureaucrat may act as the judge stopping unrelated discussions and explaining the jury their duty.

Alternatively, we can introduce some suffrage, removing disruptive people, sockpupets and people voting due to loyalty to some non-Wikipedia organizations (been it a wikiboard, language wiki or an outside internet forum). The number of edits is a bad indicator of the productiveness and the number of blocks is a bad indicator of disruptiveness, but we do not have something better yet. Thus, lets use something like the average number of edits in the article space between blocks (that is divided by the number of blocks plus one) should be more than 500..1000, or even better would be to use the number of bytes inserted into the article space (counting revert as one byte). This way, even if a some sort of cabal would be formed it would not be at least the cabal of disruptive non-productive users (aka trolls and vandals). abakharev 05:27, 19 January 2006 (UTC)

I like this idea. Perhaps create a 'voter pool' that anyone over 500 edits and a certain amount of time can put their name in. After for-and-against arguments are presented, a BCat picks 10 or so random voters to give their say. I'm sure a query could be written to provide a BCat with an entirely random list of 10 users in a 'voter pool' category. --^demon 18:37, 19 January 2006 (UTC)
I support the jury-style voting, provided that other users are allowed to present evidence in the RfAr style. It is better than a mob curcus scene which RfA reminds me these days.--Ghirla | talk 16:46, 20 January 2006 (UTC)
Good idea. Anyone has arguments against? Let's hear them. Lambiam 08:52, 23 January 2006 (UTC)
Well, I find it intriguing, but then we are not allowing everyone to express their opinion. Say, for example, I had a huge problem with you, Lambiam. So you decide to try for Adminship and I'm not on the jury. Do I get a say? How do I know my opinion gets heard? --LV (Dark Mark) 14:58, 23 January 2006 (UTC)
Perhaps to reduce the overall load of picking new jurors every day...a set of jurors is picked for a week, and they vote on all RfAs that happen to fall within that week. Then they go back into the jury pool. --^demon 15:24, 23 January 2006 (UTC)
Hmmm... how would the voting go within the jury itself? Would you need to get a simple majority of the jury, or a larger percentage of approval? Could we force jurors to recuse themselves where they may be involved in a conflict with the user? Would we have reserve jurors to fill in their spots? Would the nominee have peremptory challenges to jurors as in a voir dire hearing? --LV (Dark Mark) 15:46, 23 January 2006 (UTC)
I say that if anyone comes up with a conflict of interest (with cited evidence), then a juror should be dismissed. At which point, I supposed we'd just pull the next name from the pile. --^demon 15:50, 23 January 2006 (UTC)
Let's start with a simple majority, but clear guidelines on what qualities an editor needs to have to be given admin responsibilities. If, in spite of the guidelines, there are persistent issues of administrator problems due to too many jurors taking the guidelines too lightly, we can consider tightening the requirement to two thirds. Lambiam 20:17, 23 January 2006 (UTC)
In response to LV (Dark Mark): Before the jury deliberates, there is an opportunity for "all the interested parties to present their evidence and discuss the admin-candidate", including You know who. Lambiam 20:07, 23 January 2006 (UTC)
Got that right! ;-) Well, I just want to make sure the jury won't ignore good reasoned editors. My guess is that most will be judging in good-faith, but you can never be sure. --LV (Dark Mark) 20:15, 23 January 2006 (UTC)

Merging all votes

[edit]

Instead of having separate "support", "oppose", and "nuetral" areas, have a single discussion thread, similar to the way that AfDs, for example, are done. This makes it harder for people to do a quick vote count and pile on their own vote, and it encourages people to actually read through everyone's reasons. It would also promote discussion more: not only would everything be sequentially arranged, but there would be more pressure to write a "reasoned vote" than in the current system in which all you have to do is add an entry to the end of "your camps" vote block. Turnstep 14:54, 22 January 2006 (UTC)

Consideration in proposing changes

[edit]
  • Don't make the cure worse than the disease. -- Cecropia 16:54, 18 January 2006 (UTC)
    Let's delete RfA ;) — Ilyanep (Talk) 03:00, 19 January 2006 (UTC)
I don't think that's funny. That if you remember the deletion of Wikipedia:Articles for deletion. Oleg Alexandrov (talk) 03:25, 19 January 2006 (UTC)
I think it's funny specifically because of that. Yes, the whole thing was inapropreate, against policy, caused bad database problems, etc. But it was a slightly amusing ordeal (especially when watched from the sidelines). — Ilyanep (Talk) 03:45, 19 January 2006 (UTC)

Existing RFA process is not the problem

[edit]

I said it in several other places, so sorry if this is the second or third or fourth time you see it. But electing admins who are not so experienced is not the biggest problem. People learn on the job. The big problem is that there is too much tolerance for admins abusing their tools. And they are not always newbies; rather they may be people who have been here for a while and may think that what they do is right.

So let adminship be "no big deal", and let people be deadminned more easily if they abuse their tools. Oleg Alexandrov (talk) 03:50, 19 January 2006 (UTC)

I didn't take it easily when one new-promoted admin "learned on the job" by blocking myself when I fended off Bonaparte's socks. Another admin (whose nomination you supported) issued his first block to Mikkalai, despite his previous 50,000 edits. This "learning on the job" will be long remembered, as Mikkalai does not edit Wikipedia since then. Plenty of other examples of "learning on the job" come to mind. But you cannot escape the fact that the majority is not satisfied with current RfA and deadminship procedures. See the ongoing poll and the history of RfA talk page. --Ghirla | talk 16:57, 20 January 2006 (UTC)
I'd like to second this. I've been blocked once. It was by an admin promoted after 3 months, on 'no big deal' grounds. He then wheel-warred on the block. It really pissed me off, to an extent that has since amazed me. This admin was later hauled before arbcom (by another admin) for running off User:Name withheld through bogus blocks and page protects, and being positively obnoxious to other admins who tried to straighten him out (case was settled in IRC negotiations). This fellow learning on the job ran off one truly excellent, vastly experienced, wikipedian, and nearly ran me off too. Learning on the job is fundamentally unacceptable. It is a big deal when you have a bad admin, no matter what Jimbo says — I suspect he's never been the victim of an admin 'learning on the job.' Derex 23:09, 25 January 2006 (UTC)
Since there isn't any way for a user to use admin functions without being an admin, there's no way in which they can learn everything they need to understand without doing it on the job. RfA isn't failing on this point. What's failing is the lack of training for new admins, and the lack of any...ANY...feedback loop on admin behavior. --Durin 17:07, 26 January 2006 (UTC)
Looks like we have two problems, then: the voters themselves and the candidates. We have unqualified voters and unqualified candidates. Good show, Durin, I think you just identified what we were looking for. Linuxbeak (drop me a line) 17:10, 26 January 2006 (UTC)
Durin is insightful as always, Ghirla's example of the perils of "learning on the job" above notwithstanding. There's a strong reaction on the WP:AAP that if adminship is no big deal, de-adminship should likewise be no big deal. There's also the problem of "adminship" encompassing all of the roles of vandal fighter, copyvio fighter, welcomer, defender of the wiki, IP blocker, and deletion reviewer. That's a lot of hats to assume "untested". -- nae'blis (talk) 17:21, 26 January 2006 (UTC)
I think the only really dangerous power is that of blocking registered and established editors. Any other abuse can be rapidly overturned by another admin. It may seem a frivolous complaint, because a block can also be overturned (albeit with much more effort, since you're blocked). However, we _do_ lose good editors over it. And we irritate many more immensely. To me, the solution is to separate out the janitor functions from the police functions. Janitors can learn on the job; police can't. I'd vote a lot more RFA's through if police functions were only granted after say 3 months of adminship in good standing, as determined by bureaucrats. (Speedies are another dangerous power, but that should be solved through a 'tag and bag' approach, thus requiring agreement of two.) Derex 17:52, 26 January 2006 (UTC)
  • Frankly, the only way to decrease the number of administrators who are abusing their tools is to make it easier to get rid of such administrators. One of the ways to do it is to have some re-voting after a fixed term (a year, or two years). It should not be 80% consensus voting (God forbid, no good admins will get it), but 50% or 67% maybe expected (maybe only admins should vote in the confirmation voting, as many of the details of the admins work are not visible to non-admins). abakharev 05:04, 19 January 2006 (UTC)
    • It seems to me that admins would be more inclined to participate in some sort of 'vote of confidence' for existing admins, since they will likely have become more familiar with that person compared to their time as an ordinary user (I hope that made sense). That being said, limiting it to admin-only suffrage, and fixed time reconfirmations, are both bad ideas IMO. A (valid) complaint triggering a Vote of confidence might not be bad, though (see Wikipedia:Admin_accountability_poll#Existing_administrators). -- nae'blis (talk) 17:15, 19 January 2006 (UTC)
"Votes of confidence" would be an unnecessary bureaucratic complication. The problem is not affirming existing good admins (which are the greatest majority), but to make it easier to get rid of existing admins who abuse their tools. But I don't know how to make that work. Oleg Alexandrov (talk) 17:38, 19 January 2006 (UTC)
As a brainstroming idea: Abusing of admin tools is likely to be redone by another admin and quite often is undoing of another admin actions. Lets define admin actions as controversial if it is undoing of actions of another admin or if it was undone. Admins who made more than a fixed number of controversial edits in a fixed time frame (say 10 controversial actions in 30 days) should survive a vote of confidence. If they survive (that is receive a community support for their actions) the count of controversial actions should be reset. Obviously, if a sysadmin acts against the consensus of the other admins, he would be a frequent subject of the confidence voting. The additional benefit would be some curbing of the Wheel Warring between admins. abakharev 00:42, 20 January 2006 (UTC)
As a brainstorming idea this has a lot of legs! Lar: t/c 03:18, 21 January 2006 (UTC)
Oooooh. This is an interesting idea to play around with a little. How would we run the scheme? Would we allow e.g. one 'free' reversal each time before we add points to their license? The particularly interesting thing about it is that the (fairly small) set of admins would be key to this ever actually going through to a 'recall', and it is thus (we hope) harder to troll around with. -Splashtalk 03:32, 21 January 2006 (UTC)
  • Possible drawback - admins becoming gun-shy about reversing bad actions by other admins. I'm just playing the devil's advocate, since this is the best and freshest idea to come down the pike in a dog's age. [Kudos]] to Alex! - brenneman(t)(c) 06:40, 21 January 2006 (UTC)
I feel the devil would be in details. The idea might be useful if the numbers to trigger and the procedure of the Confidence Voting are right and it may be a disaster if the details are wrong. Since I do not have the first-hand experience as sysadmin, I can only guess the right numbers, we could select the right values based on the past experience, so choosing the numbers that would generate some Confidence Votes but not an unmanageable amount of them. The Confidence Voting rules itself should be tuned so to be not a big deal if the admin is obviously right, but to be a big deal if the admin is wrong. There should not be 80% supermajority required, probably just a 50% majority. The gun shy sysadmin maybe a problem, another possible problem maybe a War Reverting. Then a sysadmin A reverts sysadmin B not for good of the Wikiproject, but to trigger Confidence Vote of the sysadmin B, that he for some reason dislikes (OTOH, A would go this way in the direction of his own Confidence Vote, so there should be reasons). abakharev 12:06, 21 January 2006 (UTC)