Wikipedia talk:Requests for adminship/Archive 80

• Maybe polls are evil. I think that misses the point here. There's plenty of discussion above and seeming general agreement that this was a good idea. But, nobody seemed to know or be willing to go ahead with implementation. In other places where I've seen such lack of motion, a poll has a strong tendency to get things in motion. Lastly (and tongue in cheek) if polls are so evil and we could further conclude that what descends from them are evil, then the Wikimedia Foundation board is all evil, as is all of ArbCom :) --Durin 20:47, 27 January 2007 (UTC)

• The idea is good, but I'm seeing some pollish-ness in this. bibliomaniac15 01:21, 31 January 2007 (UTC)

m:Polls are evil, discuss instead

• Discussion: I was tempted to oppose, because the proposal was blatantly anti-consensus. But then I noticed the irony of opposing in a poll. ;-) If you throw out the numeric requirement, what's left is for bureaucrats to always confer. Well DUH! Why aren't they doing that already? It *would* be wise for bureaucrats to always confer with one other bureaucrat before promoting, just like admins often confer with each other too. Besides which, this would stop bureaucrats rushing to promote or deny and actually look at merit. Also, working in pairs is known to reduce error margins. Finally, more bureaucrats would stay involved and up to date on the process, reducing the likelyhood of an "everyone competent is on holiday" event. If some bureaucrats just read this and say "duh, why didn't I think of that" and like, just do it, then we're already done, finito, basta, klaar; problem solved. --Kim Bruning 18:13, 27 January 2007 (UTC)

• But the error margins are already incredibly low, perhaps one or two cases a year that there's disagreement on, and they always come when there's a promotion with a raw approval percentage of below 75%. Requiring two every time is solving a problem that doesn't exist, in other words it's needlessly bureaucratic (and it's impossible to discuss this kind of thing without an irritating number of puns). Trebor 18:28, 27 January 2007 (UTC)

(edit conflict, replying to Kim) I think the point of attaching numerical values to when bureaucratic discussion is required is so 'crats won't have to confer about whether to promote the next Brad or Sharon. Always conferring will lead to a backlog, and therefore more 'crats will be necessary. But until the atmosphere changes to one of trust in bureacrats' decisions (which this proposal might help do), no more bureaucrat candidates will get consensus for promotion. So it's something like a catch 22, no? Picaroon 18:32, 27 January 2007 (UTC)

We've already established that RFA isn't a vote. I don't want to make any more rules that give the erroneous impression that it is. --Tony Sidaway 18:37, 27 January 2007 (UTC)

Firstly, RfA is a vote. You're using "vote" as shorthand for "majority vote" (which RfA certainly isn't), but the word actually has much broader meaning.

Secondly, it's ludicrous to pretend that the numbers play no significant role in the outcomes. There are other factors as well, but they come under great scrutiny primarily in numerically borderline cases. —

David Levy

19:07, 27 January 2007 (UTC)

Hmm, in this case you contradict yourself. Either "the numbers" play no role and this is not a majority vote, or the numbers play a role and this is a majority vote. --Kim Bruning 19:23, 27 January 2007 (UTC)

No. If the numerical tally were the only factor, it would be a majority vote. —

David Levy

21:37, 27 January 2007 (UTC)

Which other factors come into play? And at what times? (if they only show up in <5% of cases, nevermind, it's just a majority vote then) --Kim Bruning 16:04, 28 January 2007 (UTC)

No, that isn't what I wrote. The closing bureaucrat always should consider the other factors (the participants' identities, the validity and comparative weight of the arguments presented, the candidate's response to objections, et cetera), but these go hand-in-hand with the numbers. In cases in which the ratio doesn't strongly point to a particular outcome, it's more difficult for the closer to evaluate these factors. —

David Levy

19:44, 28 January 2007 (UTC)

We want Bureaucrats to think darnit, else we could just replace them all with a bureaucratbot. So either we have them confer, so that they actually can apply their grey matter, or we could (ultimately) just replace them with a bot. Those seem to me to be the two possible outcomes here, roughly speaking.--Kim Bruning 18:54, 27 January 2007 (UTC)

Rubbish. You know darn well that there's no need for bureaucrats to confer with one another in clear-cut cases. Furthermore, there'd be nothing stopping a bureaucrat from opting to engage in such conference when he/she isn't required to. —

David Levy

19:07, 27 January 2007 (UTC)

Define clear-cut. Is that by %, by number of edits, by

WP:100, by previous higher level privileges already held, or by other ? --Kim Bruning

19:56, 27 January 2007 (UTC)

• This strikes me as a bad idea: (1) it invokes specific percentage requirements, which does damage to the ideal that this is a consensus discussion and not a numbers game. (2) It has a distinction between promotion and failure that strikes me as odd -- 70% failures are, in my view, much more of a problem at RfA than 70% promotions. (3) Is in direct and immediate response to a specific incident that made people unhappy; we ought to wait for a month and re-evaluate whether changing our policy in reaction to a single incident is really helpful. Christopher Parham (talk) 19:30, 27 January 2007 (UTC)

1. While the numerical tally is not the sole determining factor, it's ludicrous to pretend that it isn't a determining factor.

2. An unjust RfA failure does not significantly harm the community. It directly affects one user. If he/she allows this to demoralize him/her to the extent that it affects his/her editing (rather than simply continuing along and trying again in a few months), he/she obviously isn't cut out to be an admin.

3. If Ryulong's promotion were the only such incident, we wouldn't be having this discussion. —

David Levy

21:37, 27 January 2007 (UTC)

Of course numerical tallys are a factor in a general sense -- obviously there's a big difference between 100/0 and 50/50. But this proposal is claiming that 76/25 and 74/25 are different enough that they must be treated differently under policy. That strikes me as arbitrary. To your second point, the failure to promote people who would be perfectly good admins certainly hurts the community. Not only does it reduce our overall ability to combat vandalism, to deal with speedy deletions, etc., the fact that people fail RFA for spiteful or stupid reasons is generally detrimental to the well-being of the process here. Moreover, the fact that there are, in my view, many more improper failures than improper promotions in the 70-75% range means that the failures are a much bigger issue. Christopher Parham (talk) 22:02, 27 January 2007 (UTC)

1. Indeed, "75%" is an arbitrary and imperfect cutoff point. So what? It still would cover most controversial cases. (In other borderline cases, the closer could choose to consult another bureaucrat.)

The most important thing to keep in mind is that such consultation is never a bad thing. I ask the opponents of this proposal to explain how requiring additional bureaucratic conference will be harmful.

2. The community stands to sustain far greater harm from the promotion of one unsuitable candidate than it does from the failure to promote ten suitable candidates. As I noted, the latter individuals will continue contributing and try again in the near future (and likely pass with flying colors).

Having said that, I have absolutely no objection to the idea of applying the requirement to all borderline cases. —

David Levy

04:05, 28 January 2007 (UTC)

The problem with the arbitrary application of the rule is that without justification, a certain set of closures are subjected to special scrutiny -- this implies that they are more questionable and problematic. The current proposal indicates that <75 promotions require special scrutiny, but <75 failures do. The message being sent is that promotions in the 70-75% area are suspect. I'd rather send the opposite message -- looking back at the past couple months, every single candidate in the 70-75% area has been more than qualified for adminship. We shouldn't send the message that promoting those qualified people is suspect, and failing to promote them is not. Your last point is not a concern for me: I don't think that people promoted controversially are more likely to be unsuitable than candidates passed uncontroversial -- at least, I see nothing in evidence to suggest that this is the case. Christopher Parham (talk) 08:00, 28 January 2007 (UTC)

Yes, certain RfAs would receive special scrutiny. So what? I strongly disagree with your assertion that this automatically reflects poorly on the candidate. It merely reflects the fact that bureaucrats are human (and therefore more capable of dealing with such a situation when they have some help). If this results in a promotion, it means that the bureaucrats have determined that consensus exists. (In fact, the involvement of a second bureaucrat actually makes the promotion appear less questionable). If the candidate isn't promoted, it means that consensus doesn't exist (and it's a good thing that the original bureaucrat solicited assistance). Either way, an extra pair of bureaucrat eyes is never a bad thing. (And I've already stated that I have no objection to requiring such conference in borderline cases in which the closer intends to deem an RfA unsuccessful.)

I'm a bit taken aback by your belief that "people promoted controversially are [no] more likely to be unsuitable than candidates passed uncontroversial." When a candidate is opposed by numerous users in good standing who believe that he/she is likely to misuse his/her sysop tools, of course there's cause for concern. As a specific example,

David Levy

19:44, 28 January 2007 (UTC)

I've always found the hyperbole of slogans like "polls are evil" somewhat off-putting, though I agree that they should generally be a last resort that's best avoided. The problem is that the above proposal has to be seen in light of the fact that RfAs are treated by many, many editors as if they were polls (and – mea culpa – I think that I've done the same in the past), and it's not clear that bureaucrats always (ever) stand out against that. Looking at current RfAs, for example, one can find many, many bald votes ("support" or "oppose" with no reasons given), often responsible for pushing the tally from well under to well over 70%; when taken in conjunction with people who repeat short, slogan-like reasons given by others even when those reasons have been challenged or even refuted, and it's clear that the idea that an RfA is a discussion rather than a poll is unrealistic. I can't remember seeing a case in which a clear numerical majority wasn't followed by the closing bureaucrat, regardless of the quality or even presence of reasons (though I admit that I could have missed such cases, and I'd be happy to be given enough evidence that I'd change my mind and strike out this comment). --Mel Etitis (Μελ Ετητης) 23:58, 27 January 2007 (UTC)

I largely agree with Kim here. Aside, that is, from the stuff about "blatantly anti-consensus", which doesn't add up in the least: true consensus is a considerably stronger condition than a 'supermajority', not a weaker one, where one person's discretion overrides copious opposition. (Except in cases where an actual consensus is numerically 'masked' by irregular or just plain nonsensical voting, which is certainly a real possibility, or by excessively "individualistic" admin standards.) In cases where there's an absence of consensus in the community, at the very least there should be a consensus among the bureaucrats, and it's a tad hard to determine that unless two or more are gathered together... However, I think that the magic 75% figure is firstly, offering up a rhetorical hostage to fortune to those who'd seize on any such as a reason to oppose; and it isn't even the right magic number, IMO. I'd prefer either a generalised statement that "consensus is not clear", or a figure of 80%. I'd also prefer that this not be cast as a "process that must be followed", as an expression of expectation of the community. A vaguer statement might be less satisfying to some, but it might well fix the perceived problem. If some BCs ignore it, and assert that this wasn't what they were !elected to do, or if they interpret "borderline" or "lack of clear consensus" very differently, then perhaps that's something best dealt with under the heading of "cross that bridge when we come to it"; too little change is easier to fix than too much. Alai 03:01, 28 January 2007 (UTC)

• Oh. Yeah. It's a good idea: when there's doubt, ask if somebody agrees with you. Sounds like common sense, and we really could do without such lengthy debates as this one in the future. I'm sure that the 'crats can see that by using a little diplomacy they can avert a lot of heated tempers. I'd say that if we reword it to sound less legalistic, that'd placate several oppose-voters above. >Radiant< 16:47, 29 January 2007 (UTC)
• Numbers are counterproductive. One significant reason for denying admninship should be enough to kill a nomination (usually it does because people agree). The opposite is also true, if a nomination has a low percentage because people piled on to defeat, and the claims of offensive behavior are bogus, that should not be reason to defeat the nomination no matter the percentage. Whenever we create a numerical standard it encourages less discussion. The same thing is true with all the Xfd's. We should decide all discussions by the force of argument, and not by counting votes. --
  Samuel Wantman
  20:57, 29 January 2007 (UTC)

  I can see the point of saying that when there is significant dispute, a bureaucrat should post a note explaining his or her reasoning and/or should confer with other 'crats. I don't have a strong opinion about whether there should be specific numerical triggers for these steps or what the triggers should be, but this particular proposal doesn't ring any alarm bells - if 25% of the respondents oppose a candidate and the 'crat promotes, it seems reasonable to ask the 'crat to explain the reasoning, just in the interests of transparency. (Similarly, when editors close contentious AFDs, it make sense to explain why they went the way they did.) I think Radiant!'s proposal to reword is a very good idea - if people are uncomfortable with numerical triggers, just say something like "significantly contested RFAs." TheronJ 21:09, 29 January 2007 (UTC)

Comments

• The intent with this proposal is only to address those nominations coming in under 75% where the bureaucrat intends to promote, simply requiring a second bureaucrat to confirm the promotion. This does not affect the 75%-80% discretionary range normally exercised by bureaucrats, nor nominations above 80%. There's no intent here to erode the intent of RfA to develop consensus. That still remains. All this does is require the presence of a second bureaucrat to confirm the consensus evaluation of the first bureaucrat if they intend on promoting when the raw vote percentage is less than 75%. --Durin 16:44, 27 January 2007 (UTC)

  Let me suggest a hypothetical: what if the bureacrat who intends to successfully close a 72% nomination asks two different crats who are currently online to come verify his/her choice, in the hopes that he or she will get a faster response than if he only asked one? One comes along and indicates their support for the proposed promotion. Four minutes later, the other comes along and opposes it. Will they then wait for even more bureacrats? Or will they just say the side with 2 people behind it wins? Picaroon 17:58, 27 January 2007 (UTC)

• Remember, we have a one approval system now. Hopefully moving to two will be enough. It's at least a good start. In reality what would happen in that case is the first two would make every effort to take into account the third's reasoning. - Taxman ^Talk 18:08, 27 January 2007 (UTC)

Heya Durin. Though I often agree with you, in this case, hmm, I don't think that adding more numerical requirements exactly improves the use of consensus. On the other hand, asking bureaucrats to team up a bit... well, that might not be a bad idea. How about we just do that? (see also above) --Kim Bruning 18:13, 27 January 2007 (UTC)

(not meaning to speak for Durin) I think that's just as good depending on how asking them to team up is implemented. The only reason I mentioned percentages at all is to define what borderline meant. But I think borderline is one of those things that you know when you see it, I think that this works just as well without mentioning any numbers at all. Either way, I think this idea has a lot of merit.

RxS

18:57, 27 January 2007 (UTC)

• One problem I see is with introducing the need for a second Bureaucrat in order to close a particular RfA. Two points: 1) this kind of procedure would contradict one of the fundamental aspects of the position, which is a combination of judgement and discretion to analyze independently community consensus and render a final result; and echoing Oleg Alexandrov's post on the "Neutral" section about how this could introcude a "caballistic" aspect that wouldn't benefit the process; 2) Another key aspect of Bureaucrat action, very well put by Cecropia back in the day, was that Bureaucrat decisions provide closure to community discussion, and as such must be done in a timely fashion, since both the community and the candidate deserve a result in a reasonable time, which is why, as Cecropia once said, "Bureaucratship is not a debating club". This can be severely hurt by this requirement: I often close RfAs at times when no other Bureaucrat is online. If we must wait for someone else to go online and double-check before we can close, RfAs can at times be severely delayed without a need for it. The way I see it, it would be more useful to come up with some sort of effective review for cases where a promotion is done with a lower [than 75%?] support ratio. For instance, we could create a subpage of the BN for that, and in it, those cases could be listed whenever they take place. If any given case can obtain from two Bureaucrats other than the closing Bureaucrat an agreement that the promotion should not have happened, then it would be undone (by a Steward) at once, with the result overturned to "failed". I know it's also charged with a degree of cabalism, but at least the RfA would be closed, and we would not be debating the candidate her/himself anymore, but rather the analysis that was made of the consensus reached. It's just an idea at this point. Redux 19:04, 27 January 2007 (UTC)

THe thing, redux, is that the 75% threshold is just a guideline, it's not a magic number such that > 75% cannot be discussed and < 75% must be. We elected bureaucrats because we trusted their judgement. You do know that hard numerical limits is not the wiki way -- Drini 19:13, 27 January 2007 (UTC)

I have no problems waiting a little longer for an RFA to close. I do have issues with bureaucrat decisions being questioned. I don't mind having to wait extra hours if that means the bureaucrat decision will be certain to stand. In practices such as Pair programming, it has been shown that people make less mistakes, and do a better job, when they work in teams of 2. (A concept that was, not entirely coincidentally, developed by the same people who were involved in the early development of the wiki concept.) --Kim Bruning 19:15, 27 January 2007 (UTC)

(after edit conflicts, response to redux):*In response to 1), this isn't trying to change the role of the 'crats significantly, but is an acknowledgement that combined decisions tend to lead to better results. In response to 2), I'd prefer that a day or two be given to come up with a definitive decision than the days and days wasted arguing afterwards; in other words, I think there is a need for it. A variant on your idea was suggested further up the page. Trebor 19:16, 27 January 2007 (UTC)

Actually, the 75%-80% margin was something that Cecropia started doing for borderline cases, and was later embraced by the community — granted, at a time when not everything was controverted. It is done in the interest of Bureaucrat discretion. Creating RfA committees because certain decisions are questioned does not seem like a solution to me. Being questioned sporadically is pretty much in the job description for Bureaucrats. If I fail a RfA with a support margin of 79.5%, there will be people complaining. If apparent consensus is at 81%, but I review the support side and decide that certain participations cannot stand, lowering consensus to under 80% and decide to fail it, there will be a storm just the same. Bureaucrats are questioned, that's a fact of life. Anyone who's in this position knows it; anyone who wants to be in this position had better be well aware of it. The bottom line is that Bureaucrats are chosen on the basis of sound judgement, so that their decisions can be trusted. If there's a case where a Bureaucrat demonstrates very poor judgement, and appears to be knowingly and willingly abusing the tools (meaning, the Bureaucrat's explanations for the course of action taken are nowhere near convincing), this is not solved by creating instruction creep on Bureaucrat work, but rather by holding the Bureaucrat in question accountable. That's what ArbCom is for,

I'm very unhappy with the use of percentages, as opposed to solid reasoning. It seems strange that currently Articles For Deletion (the most criticised process on wikipedia, so much so that criticism even shows up in the press), now actually has a better system in place than requests for adminship. Do you speak for all bureaucrats? --Kim Bruning 20:59, 27 January 2007 (UTC)

So do you think those unhappy with Raul's closure should take it to ArbCom or Jimbo (I can't see how

WP:OFFICE

would be relevant)? That, to me, would appear to be creating a far uglier mess than already exists. I think most people will accept the closure as one of those things (provided Ryulong doesn't do something incredibly controversial), but that doesn't mean they want it to happen again. Bureaucrats are chosen on the basis of sound judgement, not on the basis that they're infallible. I can't see how this is unnecessary bureaucracy: it would affect maybe two cases a year, and a discussion between a couple of 'crats wastes far less time than the free-for-all further up this page.

In response to Kim Bruning, I don't think the parallel with AfD works. For AfD we have policy and guidelines which one can argue from. The closing admin can see good reasoning and bad reasoning in a far more clear-cut way. But there isn't a comparable set of criteria for RfAs. People vote for all sorts of reasons, all of which boil down to them asking "Can I trust this candidate to do a good job?" Trust is an inherently subjective thing; the level at which I might trust someone could be very different to your level. It is thus much harder for the 'crats to determine a "good" or "bad" vote (outside of obvious sockpuppetry or bad faith). There is no consensus as to what "solid reasoning" for RfAs actually is. Trebor 23:41, 27 January 2007 (UTC)

Actually, at times Articles For Deletion (then VFD) was very similar to requests for admiship. Sometimes we had to step in and kill over-zealous majority voting, and we sometimes used requests for adminship as a template. Note that all this "body of policy" didn't come from nowhere, it's basically just written down from old AFD discussions. AFD came first and worked quite well, even without anything in writing at all. Requests for adminship is only slightly different even today, in that every person writes their own criteria, instead of people working on common criteria. That and some silly person instituted percentages (that's a big oops). --Kim Bruning 06:31, 28 January 2007 (UTC)

I wasn't around in those days, so I'll take your word that AfD (VfD) worked well without anything in writing. I still don't think there's a particular agreement on what a good admin should be/have, so I'm not sure there's a way around imposing percentage on some level. Trebor 13:30, 28 January 2007 (UTC)

How about we write such a document for RFA? (Btw, you can check old versions of the AFD documents in their page history, of course) --Kim Bruning 16:08, 28 January 2007 (UTC)

I had a quick look through the oldest VfDs (y'know, for fun and stuff) and was somewhat amused by

this; I'd hope AfDs wouldn't go through like that these days. I think such a document for RfA would be good, if you could get people to agree on what should be included (although like AfD, I suppose you could work off previous discussions). Trebor

17:03, 28 January 2007 (UTC)

Quick question: If this modification is accepted, when a couple of bureaucrats decide that a 55% nomination will be successful, will people just accept it or will they claim for the bureaucrat's heads, a new RFA process, threaten to leave the project and begin saying they don't trust anyone now? -- ReyBrujo 21:01, 27 January 2007 (UTC)

The latter. Whenever bureaucrats make decisions that are outside of the margin Cecropia established, things like #Ryulong's RFA passing at 69.4%!? are most likely going to happen. (For example, I knew what I would find on this talkpage when I caught sight of Raul's edit summary). 61%, 72%, and 74% will usually lead to the same sort of result, just of varying magnitudes. This bureaucrat-discussion proposal might or might not work out, but it probably won't lead to the elimination of this sort of talk page outburst - which is why I'm somewhat wary of it. Picaroon 21:53, 27 January 2007 (UTC)

I can't really imagine a situation in which a candidate with barely over 50% support should pass, but the amount of controversy would ultimately depend on the explanation provided by the bureaucrats. If they had strong reasons for promotion, then any controversy would likely be limited.

SuperMachine

22:42, 27 January 2007 (UTC)

It is indeed a good point to say that, whenever the RfA is controverted enough, and enough of a close call, even a concerted decision between two Bureaucrats would still be as open for questioning as it happens in the present system, and it could also create the false, but still negative, impression that a cabalistic system is in place. Notice something else: Wikipedia currently has 24 or 25 accounts with Bureaucrat access. Of those, about only 5 or 6 do regular RfA closing. As far as the Bureaucrats who work regularly here, we have a long-standing goal in the Bureaucracy where we strive so that it effectively makes no difference which one of us closes a RfA. It is even rarer than those controverted close calls that I, and I suspect other Bureaucrats as well when positions are reversed, would consider that if I had closed a RfA the result would have been different. We give no one reason to speculate something like "my RfA is real close, I hope Redux/Taxman/Essjay is the one to close it". It should make no difference.
As far as my suggestions about ArbCom, Jimbo, or OFFICE, those are a initial stage for what I do think would be important: that we arrive at an effective system in which any perceived abuse could be addressed efficiently. The first step would be determining what could be called abuse and what is only the natural result of a decision about a discussion where opinions where much more divided than they are in clear-cut RfAs, which will, of course, result in dissatisfaction that normally vents in the form of criticism of the closing Bureaucrat. It does seems to me that asking two Bureaucrats to concur in such close calls would only result in criticism of two Bureaucrats at once, not in the prevention of any criticism of Bureaucrat actions whenever a close call needs to be made. Redux 03:04, 28 January 2007 (UTC)

I think you're correct that no reason is given for people to say that... about the people you say it's not said about. That's why I say I'd be happy with either a statement that borderline promotions should involve discussion between BCs, or that there should be some sort of reconfirmation of the position. Alai 03:43, 28 January 2007 (UTC)

Two methods come to my mind:

1. Since bureaucrats "fight" over RFA to close them, they have spare time enough to analyze the RFA, reach a consensus between themselves and post it. Of course, now people would ask for the head of all the bureaucrats, that they need to resign or reconfirm.
2. This is not a vote, and everyone is entitled to their opinion. But some opinions make more sense than others, and the bureaucrats can review the RFA with a more critical eye. It is a matter of collecting the usually 8-10 or so most important opinions, and measure them against each other. I am not saying "dismiss" opinions, but if someone provides a good rationale for his opinion, and someone else supports it either with the "per XXX" or with a similar opinion, it is not two but instead one opinion supported by two users. Personally, an opinion like "has a low edit summary usage" weights less than "has been blocked recently due 3RR", or an opinion like "fails 1FA" weights less than "civility problems". I know this sounds arbitrary, but we are talking about exceptional cases and humans who can think logically and not bots.

I am dismissing others that could be done with bots ("80% or nothing", "assign some users more weight" based on previous RFA participation, etc), because they are the most debatable. However, these solutions would still put 'crats under fire. Until people understand that consensus is not voting, that they must give solid opinions, that consensus is not a magic numerical barrier, and that sometimes things won't end the way they appear, this kind of problems will always exist. We can create a "Request review", but what is the point? Bureaucrat will have to close it too, and since they are few, result is not likely to change. -- ReyBrujo 04:34, 28 January 2007 (UTC)

A controversial promotion would still be...controversial (obviously), even if two bureaucrats concurred, but it would hopefully prevent accusations of personal bias. Given how this whole discussion got started, I suppose a key question is whether the other 'crats would have agreed with Raul's decision to promote. I don't expect an answer to that (it wouldn't be fair to Ryulong), but past evidence suggests he probably wouldn't. At the very least, the reason given wouldn't be subject to a 'crat's personal interactions with the user. Trebor 13:30, 28 January 2007 (UTC)

• To echo Durin's usual answer to any and all proposals, what exactly is the problem that is fixed by this proposal? Do we promote many bad administrators with 70-80% support range? Kusma (討論) 14:32, 31 January 2007 (UTC)

  To give an answer to my own question, a problem is that we do not promote enough good admins in the 70-80% range. (Note that the last three involuntarily desysopped admins had more than 80% support in their RfAs). But is this the problem that this proposal tries to solve? Kusma (討論) 14:32, 31 January 2007 (UTC)

Moving forward with fewer numbers

Many people above seem to agree that it is better if bureaucrats consult with each other before taking unusual actions that are likely to be controversy inducing. However, many opposers to Durin's proposal (and some supporters) feel that including a fixed precentage (i.e. 75%) is a bad idea. So, I would like to suggest moving this initiative forward, but without the fixed threshold. A possibility that I suggested above is a declaration that:

If a bureaucrat intends to close an RFA in way that may be potentially controversial, then he should consult with other bureaucrats and identify at least one concurring bureaucrat at the time of the closing.

Other formulations are of course welcome, and feel free to make suggestions, but I think the core idea here is the feeling that difficult cases would be resolved more effectively, and with less wikidrama, if the result is presented as a consensus among bureaucrats rather than the actions of a single individual. Dragons flight 04:10, 29 January 2007 (UTC)

I think that something along these is sensible. The only problem I can personally see is the 'crystal ball' aspect of the "potentially controversial" clause. I might be inclined to phrase it more in terms of the closure being a 'close call' or the presence of 'unusual circumstances', or something along those lines (at the risk of being even more vague). Alai 04:30, 29 January 2007 (UTC)

I've already suggested that something similar be added to the original proposal, so I obviously support its adoption with or without the numerical component. —

David Levy

04:35, 29 January 2007 (UTC)

• Actually, since we trust bureaucrats to have judgment to determine when to promote or not, is there anything prohibiting them from making decisions en banc in borderline cases? Titoxd^(?!?) 07:49, 29 January 2007 (UTC)

• • I'm pretty sure there's not, and it seems to have happened a number of times in the past. Nor do I imagine anyone would object to this -- though that's not to say they might not object to whatever the decision might be, regardless of the particular process involved. The question is, can and will the BCs be enjoined or encouraged to do so on some consistent basis? I imagine one (or more?) of three things will happen: firstly, some fairly weak "expectation" wording will be added to a process page talking about this; secondly, some or all of the BCs may take note of this decision, and decide to consult in such cases (a number seem to have expressed intentions to do so, even before this latest debate), independent of such wording being formally agreed on and added; and/or lastly, one or more of the BCs may at some point decide to go ahead and make similar promotions without any consultations. Either of the first two is all well and good; if the third happens, we're likely to see a fresh iteration of discussions such as this, either mutedly or escalatedly depending on assortedly (such as how soon, who gets promoted, who does it, how 'discretionary' it is, and doubtless much else). Alai 03:01, 30 January 2007 (UTC)

• Unfortunately "potentially controversial" is an inherently backwards-looking judgment; I expect most of our current bureaucrats would not think their decisions are controversial until the WP:AN-storm starts... "where consensus is not clear" might be more in line with our current, admittedly muddy, definition of Wikipedia:Consensus. I agree that adding any numbers at all to the official document is silly, especially since the threshold has been instituted, violated, moved, penalized, promoted, drawn, quartered, and honored. -- nae'blis 17:33, 30 January 2007 (UTC)

• I would also support this, and perhaps keeping numbers out of the picture is a better idea. I think the numbers were introduced as a reaction to the problem; every controversial case has been where the support percentage was below 75. I expect (or at least I would hope) the 'crats can recognise when they're making a potentially controversial decision anyway, given their ability to read community consensus. Of course, given the rarity of these events, it's very tricky to tell whether this will work or not. Trebor 18:43, 30 January 2007 (UTC)

Many times it's been tossed around: "XXX promoted YYY without consensus, as there was only 74.9999 support!!".

PEople need to understand that consensus is not the same as "passing a numerical threshold". If they don't want to understand it, they will have a hard time on wikipedia, probably get burned out and leave. Consensus has always been about judgement, and not about numerical limits. New users (and new is < 6 months old): here's advice, stop thinking about discussions as votings. WHoever tells you that's the wikipedia way, it's misleading you, and on the long run will be harming your wiki experience. -- Drini 19:10, 27 January 2007 (UTC)

• Very much agreed. RfA is frequently (and I think more so relative to traffic than any other part of Wikipedia) the subject of disagreement and lack of understanding as to what consensus is. Determining consensus is no easy thing. It takes, I think, considerable experience on Wikipedia before a strong understanding of what consensus means within this context is understood. This is one of the reasons why it is so difficult to become a bureaucrat, and at times so difficult to remain one. What is different here at RfA as opposed to other areas of Wikipedia is that determinations are made on people, not on articles, templates, etc. Emotions can and do become trampled upon. --Durin 20:55, 27 January 2007 (UTC)

Agreed. And what about weak and strong supports? Are those all counted as 1.0?

Voice-of-All

18:36, 28 January 2007 (UTC)

• It's not a vote :) Besides, if such mention of "weak" and "strong" counted for more or less, everybody would mark as "ultra super duper incredibly maximum over-the-top" oppose/support, etc... --Durin 14:34, 29 January 2007 (UTC)

• Well, no, I weigh them in especially if they are backed with solid reasoning. Something meeting your example would get no extra weight, unless it was followed by something like, "but seriously, suppport due to X", and X was a solid reason. So in short, one of the things I do weigh in on close cases is the strength of the support and the strength of the reasoning behind them. It is illustrative in some cases that a large number of supports are weak and most opposes are strong, or vice versa. I'll say again and again, solid reasoning is much more valuable than a bare vote. - Taxman ^Talk 15:42, 29 January 2007 (UTC)

I would perfectly agree with you, if we are talking close cases, as you say (for me that means 75%-80% support, or slight deviations from it). If a bureaucrat would however say "This candidate had 72% support only, but the opposition was incoherent and unconvincing, so I promoted", that would violate the current community agreed process and should not be allowed. Oleg Alexandrov (talk) 16:27, 29 January 2007 (UTC)

Actually, if the opposition is unconvincing then the 'crat(s) should be promoting. I'm not sure what you're seeing as being violated by following that standard. We don't allow unconvincing deletion arguments to stand at AFD do we? -- nae'blis 17:29, 30 January 2007 (UTC)

The problem is there is no consensus as to what an "unconvincing" argument or a piece of "solid reasoning" actually is. The 'crats must have their own opinions, and they are probably pretty accurate, but unlike AfD there's no page to point at and go "candidate is good/bad because of this"; everyone has their own standards and way of judging whether a candidate is suitable. Trebor 17:34, 30 January 2007 (UTC)

That's the biggest issue; whether something is unconvincing is really in the eye of the beholder.

.V. _[Talk|Email

] 18:18, 30 January 2007 (UTC)

The trouble is that there's a lack of common standards at RFA (not as true at *fD), as well as the inevitable disagreement over the application of those standards (just as there is at *fD). I think that it's not an unreasonable step for a bureaucrat's "discretion" to be be extended in the direction of discounting or at least "downeighting" opposition (or indeed support) based on admin standards that clearly do not have community support. Thus to make the distinction of fact explicit: I'm suggesting that if an opposing comment makes the "finding of fact" that a candidate is uncivil, say, then the BC take that fully into account, even if they don't agree with it; whereas if an opposing comment's "finding of fact" is that the candidate has very few image uploads, then the BC should have discretion to consider that opposition as not being in line with the generally expressed community consensus on the matter of standards, as against consensus on the particular case. (Alter examples to reflect actual community consensus... if any.) This could be significant where someone has a number of "misc. opposes" for various reasons that are collectively "incoherent" (as against individually well-argued, or not.) Alai 19:57, 30 January 2007 (UTC)

What about arguments that have partial community support? For instance an admins activity in article-writing is viewed by some as vital, so they can understand editing disputes, and by others as an irrelevance. Trebor 20:35, 30 January 2007 (UTC)

That's a good point. I was thinking in terms of criteria for which there was an apparent "broad consensus against", i.e. only a small minority of people expressing or applying such a standard. Thus "not enough article-writing" I'd see as something there's not a consensus on either way, whereas extreme versions of it, like "1FA" I'd say there was in effect, a consensus against not being a sensible standard. Alai 21:42, 30 January 2007 (UTC)

Maybe I wasn't clear. By "unconvincing", I'm talking about things like "candidate is incivil", without a diff and with several people objecting to the characterization by showing diffs of clear civility in the face of adversity. Or "No projectspace edits", when a simple edit count will show this to be incorrect. 1FA is a valid reason for those who ascribe to it. -- nae'blis 19:48, 31 January 2007 (UTC)

(reset indent) Yeah, I was just wondering how the 'crats ended up weighing this sort of thing, in the absence of any established guidelines. I mean it rarely matters, as there is usually a firm consensus to promote or not; but in the cases where there isn't, how do they judge it? Trebor 21:50, 30 January 2007 (UTC)

Apologies if this has been brought up before, I'm not a regular contributor to the Rfa process unless a user I have interacted with regularly is nominated; however, after reading a few of the reasons for opposition in several discussions given by some users I'm curious as to their "validity" if you will, for example, the following statement (paraphrased) from Qwghlm's ongoing Rfa (my first nomination):

Oppose • I am not entirely convinced you need the tools...No person "has a right" or "deserves" to be an administrator- it is not a status symbol nor some milestone to be achieved...Try again in a few months when you have more experience with images...

Now, one thing about the Administration process which stuck in my mind was the "no big deal" quote from Jimbo. I read this as "every good faith editor who isn't going to cause damage by being given the tools has the potential to be a sysop". The opposition I have read on several occasions, in some form or another as doesn't need the tools just doesn't sit well with me - what harm comes from giving someone the tools if they don't use them? I can see that if you gave sysop privileges to a user who had no idea of process then there could be some mess to clean up, and opposes for that reason are fine. But I can't see why people have issues giving administrator privileges to users who are clearly well versed in Wikipedia process - unless there is a strong argument against having too many administrators (if there is, what is it?).

Sorry for the rant, but I wanted to see what the opposing views were. Cheers. QmunkE 21:22, 27 January 2007 (UTC)

The issue isn't, literally, that someone doesn't need the tools. (At least, not to me.) The issue is that if you have not participated in areas concerning the tools, there is no way to evaluate if you will use them well. -Amark moo! 21:29, 27 January 2007 (UTC)

Amarkov makes an excellent point, if someone who has been around for 3+ months and isn't showing any activity in the areas where an admin will work, it's reasonable to assume 2 things, 1 - the user isn't overly interested in administrative tasks and 2 - the user probably doesn't have the necessary experience. With the huge backlogs and workload on existing administrators, it's unreasonable to promote editors who will only use the admin tools when they come across a problem, we're needing admins who will seek out and address problems through the use of WP:AN, WP:ANI, WP:AIV, WP:XFD, WP:COPYVIO etc. It's unfair to the project and to existing administrators to promote editors who won't make regular use of the tools; promoting editors who won't use the tools could have the side affect of creating !voter apathy whereby those who do need the tools can't get the tools thanks to there being "too many admins". I'd say those who don't need the tools, no matter how well known or how good, shouldn't waste the projects time with RfA, and leave RfA for people who do need the bit, for whatever reason. Admins on Wikipedia should be downtrodden janitors, not the arrogant and vengeful idiots some are at present, and until our admins actually stop with the arrogant crap, we'll never be able to sort out RfA and promote the right editors. --

Heligoland

22:01, 27 January 2007 (UTC)

Admins can do a lot of damage. It's a limited amount, but that limit is still quite high. Therefore, promoting anyone is a risk (it doesn't matter how great the risk is or isn't, there is definitely a risk), so there needs to be a reason to take that risk. If someone isn't going to use the tools, then there is no reason to take the risk. This is actually in keeping with the "no big deal" concept - adminship should be given out as it is needed. It shouldn't matter who the person is (beyond the obvious trust issues), what matters is what they'll do with the tools. Adminship isn't a status symbol, it's a toolkit. If you don't need to use the tools, there is no point in you having the toolkit - you don't gain anything by it. --Tango 23:45, 27 January 2007 (UTC)

I was going to make the same sort of point as Tango. Entrusting people with the more destructive tools is a step of faith by the community. They can never be certain that the admin won't misuse them (even if it would only be under extreme circumstances) or even that someone else might access their account and cause damage. This faith is worth putting in when the access will help the encyclopaedia, but an unnecessary risk if it will not (or if it would help only minimally). The points made further up are important too; if you haven't participated in administrative areas, then it's hard to judge if a candidate understands how to. Trebor 23:52, 27 January 2007 (UTC)

It strikes me as too extreme to say that admin tools should only be available to those that need them. I don't need the tools, but if I didn't have them there would be one less person available to take a whack at

CAT:CSD or do a blocked move or whatever. I think this is the most common mode of admin-tool use and you get a skewed view of 'what admins do' by reading the posts of the most common AN participants or whathaveyou. That said, Amarkov is right; 'doesn't need the tools' is often shorthand for 'has no experience in relevant areas and it's hard to tell if the candidate knows how to use the tools properly'. Opabinia regalis

02:47, 28 January 2007 (UTC)

By "need", I guess we mean "need in order to do what they intend to do to improve wikipedia". If all someone intends to do is write articles, then they don't "need" the admin tools. --Tango 16:14, 29 January 2007 (UTC)

My interpretation of this issue is that there are some administrators who don't seem to do administrative tasks. I've noticed admins who don't seem to use the tools at all. I don't know if they've forgotten about it or if they just like the clout... but this is a big issue in my eyes.

.V. _[Talk|Email

] 18:23, 30 January 2007 (UTC)

Tango: until they need to delete one, that is ;) Since the question was asked, it seems helpful to point out that - extravagant answers to question 1 aside - "need" is a somewhat subjective assessment. V, I'm not sure what's a 'big issue' here; the fact that some admins don't regularly take admin actions? Why? Opabinia regalis 01:48, 31 January 2007 (UTC)

Two reasons. For one, having the tools and not using them creates a bit of an administrative dead weight. Optimal administration systems work when the admins are constantly being involved in their administrative field. Otherwise, it makes for a stale system. You can see this in places like Categories for Deletion... there's always a massive backlog there. Secondly, it pads the admin count so that aspiring admins have a harder time getting in. Obviously, not every admin can be sitting on Wikipedia doing admin actions all day. It's just that there should be some degree of regularity, especially with the various admin backlog issues that WP has.

.V. _[Talk|Email

] 17:06, 31 January 2007 (UTC)

If this idea works, it may even be possible to relax the rules on who can make the initial assessment (to all admins, or maybe even all registered users). If there is a consensus, it doesn't matter who speaks up first, there still won't be any objections, and a crat can come come along after 24 hours and carry out the decision. --Tango 00:01, 28 January 2007 (UTC)

RfA is a consensus gathering exercise. No, it's not. If you look at Wikipedia:Consensus, you'll see that "consensus" basically applies to what appears in articles - and that makes sense because when there is disagreement, one side or the other ultimately stops fighting about it (with rare cases going to ArbComm). As stated at that policy, RfAs are decided by a supermajority. -- John Broughton ^(♫♫) 00:20, 28 January 2007 (UTC)

There's a bunch of people working on that page right now, the contents are in flux, so don't quote it. En.wikipedia is run by consensus in different forms and flavors throughout the entire system. Requests for Adminship is not run by supermajority. That and it's a tradition that anything actually going supermajority gets killed outright. So be careful saying that too loudly, someone might hear you. --Kim Bruning 06:19, 28 January 2007 (UTC)

RfA may be decided by de facto supermajority, but it's definitely meant to be consensus, or at least "rough consensus". --Tango 14:44, 28 January 2007 (UTC)

Not even today. See Wikipedia talk:Consensus. --Kim Bruning 15:29, 28 January 2007 (UTC)

A little under a year ago there was an RFA which was accidentally closed early as a failed nomination with 62% support... unfortunately, that was a mathematical error and the RFA actually had 73% support. It was suggested that the bureaucrats should take another look at it as, while 62% support would be dismissed out of hand, 73% support might be sufficient to promote if thorough review found that (as some thought) the opposition seemed weak/questionable and consensus generally in favor.

This was greeted with scorn and statements that 73% support was so far below the thresh-hold of acceptability that no bureaucrat would ever consider promoting at that level. Yet, here we are ten months later and the 'unthinkable' has been done... a couple of times.

I don't have a problem with bureaucrats having wide discretion in judging RFA consensus. I don't have a problem with the 'discretionary' promotions the bureaucrats have made. However, I DO have a problem with such discretion being applied inconsistently. If 69.4% is sufficient to suggest a possible community consensus (hey, it's more than two-thirds) then it should always be enough for the bureaucrats to take a careful look and decide whether the person is qualified... not just when they personally support the applicant. I'd be alot more comfortable with a bureaucrat making a discretionary promotion because, 'I found the "he does not have enough MediaWiki talk namespace edits" oppose votes unconvincing', than because, 'I know this guy and he is cool'. I especially don't want to ever again see people browbeaten for suggesting that 73% support is significantly different enough from 62% support to warrant another look. --CBD 20:38, 28 January 2007 (UTC)

While I admit to not having seen signs of inconsistancy personally, I wholly agree with the statement above. I also note Redux's comments much higher up about how the bureaucrats work very hard to be consistant among themselves. I would tend to agree and believe this. I don't see it as cabalism, I see it as:

• 1.) People who would seem to be well-versed in Wikipedia policy/guidelines, or at least those which would be commonly referenced at or about RfA.
• 2.) People which the community has voted whom they have chosen to trust with the use of such tools, and the discernment of when and how to use such tools.

In other words, when a group of individuals follow the same rules, they often make similar decisions when looking over the same information. That's partially why I have no problem suggesting that we require 2 bureaucrats to close RfAs. They would seem to me to be rather consistant, and so 2 or 20 should still come to the same consensus. This solution just has the added benefit of following

WP:BITE, since newbies, or those who may have personal reasons to be concerned will have the added comfort knowing that there was truly a consensus regarding such action. - jc37

20:48, 28 January 2007 (UTC)

The current tally for Support/Oppose/Neutral is frequently wrong, especially for the Oppose count. And I don't see any edit problems in how folks inserted their #Vote. Thoughts? Keesiewonder ^talk 11:25, 29 January 2007 (UTC)

Ah - Maybe this needs to be updated manually?! Silly me. Keesiewonder ^talk 11:29, 29 January 2007 (UTC)

Yep. Cue brand new (probably unnecessary and repetitive) discussion about whether we even need the tally in the first place. – Chacor 11:30, 29 January 2007 (UTC)

So true. Sorry to contribute redundancy. Too bad all the computers and software involved with WP cannot keep accurate tallies for us. Keesiewonder ^talk 11:38, 29 January 2007 (UTC)

There are automatically updated (although with a time delay) RfA-vote counters at

WP:BN. The ones on the RfA page are updated manually (normally someone will update them pretty quickly if the previous voter didn't, although sometimes they stay wrong for a while). --ais523 13:04, 29 January 2007 (UTC

)

On the occasions that I add my recommendation, I don't update the counter as I made a recommendation and didn't vote so I don't correct the tally. MLA 14:59, 29 January 2007 (UTC)

You can usually tell who is going to run for adminship fairly soon by spotting the people who update the RFA scores every time a vote comes in, solely to get their Wikipedia: space edit count up.

►

10:05, 30 January 2007 (UTC)

Those edits should be marked as minor and they still need to have more than just Wiki-namespace edits. Gizza^{Chat ©} 10:44, 30 January 2007 (UTC)

It would be rather simple to write a bot to automatically update the tallies every few minutes, but that would be a very silly use for a bot. Oleg Alexandrov (talk) 16:10, 30 January 2007 (UTC)

That's an incredibly pessimistic view. --

talk

) 16:33, 30 January 2007 (UTC)

I, too, disagree with Proto. I've been in the habit of making such tally updates for months, merely so it is easy to tell how someone is doing at a glance - inaccurate bellwether though they be. The fact that I just accepted a nomination is really unrelated, and I don't think I could care less about the effect on my Wikipedia namespace edits. Picaroon 00:20, 31 January 2007 (UTC)

This is a very naive question, and I post it only to make a point: if, as we all know, a !vote is only a recommendation, and is meant only to provide some guidance for the closing bureaucrat, why do we need numbers recorded at all?--

Because for all practical purposes RfA really is a vote. In general a candidate with at least 80% favorable votes is made an admin, a candidate with less than 70% favorable is not, and those in the range 70-80% can be admin'd at bureaucrat discretion. Yes, there are exceptions, but most of the time it's straight up or down. Raymond Arritt 00:39, 31 January 2007 (UTC)

Please see Wikipedia:Bots/Requests for approval/Cydebot 3 for an ongoing disusccion of a bot account that may require admin rights. — xaosflux ^Talk 13:06, 31 January 2007 (UTC)

It may be worth some thought on creating some different standards for adminbots going through the RfA procedure, as a way to increase the likelihood of approval for useful adminbots. I have written some ideas I had at User:NoSeptember/adminbots. It would be nice to have a smoothly operating adminbot RfA system, as such requests are likely to become more and more frequent, and the time to do this is before a particular bot is undergoing an RfA. NoSeptember 14:14, 31 January 2007 (UTC)

This does need to be addressed more formally but for now lets kick it around on User:NoSeptember/adminbots, I've added some thoughts there already. Gnangarra 15:06, 31 January 2007 (UTC)

• One of the things that admin bot requests trip up on and routinely fail over is the lack of acceptance of a bot having admin rights. I think if we have a discussion and consensus decision on whether bots should have admin rights or no we'd help clear this up. --Durin 17:06, 31 January 2007 (UTC)

  There are plenty of people who are not firmly on either side of that question. So the issue of what conditions should be imposed on adminbots to satisfy people's legitimate concerns will determine how they decide the question of whether we should have them or not. NoSeptember 17:27, 31 January 2007 (UTC)

This is in reference to this edit.
I'm, of course,

• Agreed. Also see Wikipedia:Bureaucrats'_noticeboard#Wikipedia:Requests_for_bureaucratship.2FDeskana. --Durin 13:43, 31 January 2007 (UTC)

Actually, in a previous discussion I believe consensus was to allow anyone who wants to to close. That wasn't my idea at the time, but now I am convinced that this was actually the correct line to follow. If someone closes incorrectly, it can always be discussed here. This would also be the best way to judge if someone is suited for bureaucrat (or for admin, or etc for that matter). If they do closings (mostly) correctly on requests for adminship or articles for deletion or what have you, then that contributes to their suitability for being a bureaucrat or admin respectively. --Kim Bruning 15:28, 31 January 2007 (UTC)

But if someone messes up and closes a close call RfA, we will have a Big Mess and a lot of unnecessary drama. I think we should be more interested in ensuring a fair and proper close for the candidates of RfAs than of measuring the closing talents of potential future bureaucrats.
There is no need for the impatience to close RfAs past their 7 day scheduled deadline. In the pre-bureaucrat days, RfAs often sat unclosed for weeks (until a developer got around to it). See this for a history of closes in 2002/2003. NoSeptember 16:08, 31 January 2007 (UTC)

What previous discussion? I don't remember any such consensus and find it unlikely. --Tango 16:36, 31 January 2007 (UTC)

It should also be noted that this very project page says "Only bureaucrats may close or de-list a nomination as a definitive promotion or non-promotion." I have noticed a lot of snowballing of RfAs lately by non-bureaucrats; this may not be a big deal, but closing any non-snowball one should definitely be reserved for bureaucrats, or, if there is consensus against this, the instructions on the project page should be changed. Heimstern Läufer 16:41, 31 January 2007 (UTC)

The only consensus in existence for "anyone to close" refers to the situations I mentioned in my post above. There certainly is no consensus for anyone to close any RfA, especially since that wouldn't make much sense, either logically (that would negate the need for Bureaucrats as a separate access level, with specific duties, as it exists presently) or technically, in the cases where a promotion would be in order. There has been discussion on whether or not any user should be able to close snowballing RfAs, that is, early delisting. Ordinarily, the only situations where non-Bureaucrats should close a RfA early is if it is snowballing exactly because it fits the situaton n.1 from my post above (e.g.: user with 15 edits posts a RfA, either because the user misunderstood the guidelines for requesting adminship or because it is an attempt of disruption/trolling — specifically when the user requested adminship because s/he misunderstood the minimal expectancies for requesting the tools, we discussed the utility of having the misguided RfA closed asap in order to avoid loosing a potentially good new contributor because of excessive criticism brought on by the premature RfA). Redux 18:11, 31 January 2007 (UTC)

Thanks for the clarification; this matter has confused me for a while. Heimstern Läufer 00:12, 1 February 2007 (UTC)

Just take a look at recent sizes of

As to the issue of people improperly removing the CSD tag: I suggest that we look at old versions (24 hours old I would suggest) of your bot's output page and check on all of the non-red links. Some will be protected deleted pages, some will be improved articles, but some will be articles that should be re-added to CSD. Your bot output keeping track of what was historically in the category makes this possible. Thanks. NoSeptember 15:35, 31 January 2007 (UTC)

Indeed, this was the main reason I started making and updating those lists. Unfortunately, they still can't catch people who remove the speedy tag before the next update of the list. --Cyde Weys 15:51, 31 January 2007 (UTC)

Could you do a test on that? Do an output every minute for a brief period (10 minutes during prime time) and then see how many additional articles you picked up versus what you would have caught with the usual 5 minute schedule. NoSeptember 15:57, 31 January 2007 (UTC)

We need more admins who fuck? *chuckle* Sorry, couldn't resist. I don't think it's fair to attack people who oppose due to lack of mainspace contributions, although if you specifically mean those who ask for great articles I'd understand.

WP:ENC, so it's only fair to expect some work on the actual encyclopedia. – Chacor

15:20, 31 January 2007 (UTC)

I think there is a bit of a disconnect when it comes to that. There are many editors who write great articles and there are many editors who devote there resources to counteracting vandalism (whether rc patrol ore new page patrol). Perhaps these editors are not masters of the english language but still want to contribute to the encylopedia in some way. I personally think it is a bit offensive to oppose based on an editors massive amount of work reverting vandlism but little article contributions. I would rather an editor work where they are comfterable than feel forced to add uncited content in articles just to get contributions up. When it comes down to it, some of the rc patrollers spend so much time on wikipedia protecting the article created by the main contributors and when it comes to RFA, it often seems like they are slapped in the face for doing it. -- Chrislk02 (Chris Kreider) 15:25, 31 January 2007 (UTC)

Alright alright...I'll get back to clearing CSD backlogs. :-P Nishkid64 15:32, 31 January 2007 (UTC)

If every admin could clear 1 item of CSD backlog? :-) --Kim Bruning 15:39, 31 January 2007 (UTC)

At this rate it'd have to be every active admin clearing seven items of CSD backlog every day, with an exponential growth curve in the number per admin per day over time. --Cyde Weys 15:41, 31 January 2007 (UTC)

I fear that too many pages are being put on the CSD backlog. I'm seeing what in my opinion are reckless deletes not based on any wikipedia policiy. Mathiastck 14:41, 1 February 2007 (UTC)

/me puts in an order for one of those shirts. sounds nice, fashionable even. :D

Bubba hotep

15:40, 31 January 2007 (UTC)

I'll be nominating someone for adminship within the next 48 hours; perhaps that will help; and I guess I need to put in a couple of shifts myself at CSD soon. It was much easier to kibitz on ANI and heckle. :) Newyorkbrad 15:48, 31 January 2007 (UTC)

<bad edit conflict> Slightly off-topic... but worth saying here, I think... speaking as a

• A big help at
  WP:AIV has been the creation of User:HBC AIV helperbot. Makes maintenance of WP:AIV far, far easier. It even updates it when an admin blocks someone without knowing they were listed at WP:AIV. Nice piece of work. --Durin
  16:14, 31 January 2007 (UTC)

<ironic edit conflict>I second that! Great bot. Also

Bubba hotep

16:21, 31 January 2007 (UTC)

Cyde's bot has been keeping track in edit summaries of how many items are listed at CAT:CSD since 25 January. It's now just 6 days later. The number of admins we have has changed only slightly over the last 6 days. There's not enough of a pattern here to say, affirmatively, that we must have more admins to deal with this and that there is an exponential consistent pattern. Looking at what data there is, it looks like the numbers vary greatly through a range...not necessarily exponential growth. Also, since making note of the problem here, the list has shrunk >100 items. Maybe it's just a statistical bump; not enough data to know. --Durin 16:19, 31 January 2007 (UTC)

I think you may be misunderstanding my point in listing the bot-updated page. I'm not using it to say that anything has gotten appreciably worse in just a week, but rather, to illustrate how persistently high the number of pages in

CAT:CSD is. --Cyde Weys

17:21, 31 January 2007 (UTC)

• I understand what you're saying. I simply disagree. 6 days of traffic analysis does not a pattern make. DF's graph shows considerably more data, and shows a slow rise of 33% over the last six months, not an exponential rise. The spike we saw of late has now been reduced to 104 entries. I think the recommendation to have the bot notify WP:AN when the list gets over 500 is a great idea. --Durin 17:40, 31 January 2007 (UTC)

since making note of the problem... - making note of it is obviously the cause of this decline. Maybe the bot should be programmed to notify WP:AN whenever the count goes over 500 ;). NoSeptember 16:29, 31 January 2007 (UTC)

• Not a bad idea. I did some quick stats on this and found that the 2nd standard deviation is at 441 entries on CAT:CSD. There's been three spikes above that number over the last few days. There certainly isn't a pattern of exponential growth though. Looks more like some regular CAT:CSD cleaners just went off and did something else temporarily. The list of active CAT:CSD entries is now down 40% from when Cyde made mention of the problem here (and it was only here that he made mention of it). I suspect if instead the bot made notice as you suggest that the backlog would be cleared in short fashion. --Durin 16:34, 31 January 2007 (UTC) PS: And now it's down >50% from Cyde's mention here.

I agree with the sentiment behind Cyde's comments. I feel many of the RFA "requirements" that candidates are forced to meet in order to pass an RFA aren't helpful. I've been saying this for six months, however, and it doesn't seem likely to change now. Firsfron of Ronchester 16:42, 31 January 2007 (UTC)

I've been tracking CSD for just over 6 months. 500 is unusual. I've added a chart at the top of the thread. Dragons flight 16:49, 31 January 2007 (UTC)

Time of day is an issue. A surge of junk article creations occur late evening (USA time), they get tagged quickly, but it takes a while for the admins to delete them. NoSeptember 17:04, 31 January 2007 (UTC)

Actually, averaged by hour, the peak time of day (6 AM EST) is only ~25 more CSDs than the min time of day (7 PM EST). The fluctuations from other sources are much larger than the daily cycle. Dragons flight 18:32, 31 January 2007 (UTC)

• Based on the chart now posted by Dragon's Flight in this section (thanks DF!), there does appear to be about 33% growth in CSD requests over the last 200 days. We've had a 15% increase in the number of admins over the same time period. So, there is some basis for concluding we need more admins. But, the growth is not exponential. --Durin 16:51, 31 January 2007 (UTC)
• Just want to also make it clear that I agree with the sentiment that some of the reasons for opposition are invalid. Also, having more admins is not a bad thing, but a good thing. We certainly don't need proof that we need more admins in order to say we need more admins. --Durin 16:55, 31 January 2007 (UTC)
  • I think many of the oppose voters have criteria based on the right idea, but implemented incorrectly. It is a good idea to oppose people who are not experienced with Wikipedia policies. However, "200 Wikipedia space edits" or "20000 Wikipedia space edits" neither proves nor disproves experience. Similarly, administrators are janitors who need to support and have respect for article writers that do not care about maintenance work. However, you don't need to be a good article writer yourself to be sensitive to the problems they face and to be able to help them write a great encyclopedia. And good article writers might be really bad at janitorial tasks. So let us please try to get more good admins, not people who fit some random set of edit count / article writing / image criteria. Kusma (討論) 17:06, 31 January 2007 (UTC)
    • I suppose we need more admins who are willing to tackle CSD, so yes, big plus for any RfA answering question one with "I am dying to plod through CSD". Candidates who are more along the lines of "well, it's time I moved up; I won't be using my buttons much unless maybe when I freak out over some dispute" are not really relevant.
      dab (𒁳)
      17:11, 31 January 2007 (UTC)
      • There are plenty of viable candidates who are not going to use the tools much. That does not mean they would make bad admins. If we have particular, consistent backlogs it'd be nice to have candidates who want to address those but a candidate who does not specifically want to address those should not be thought of less because they do not want to tackle the now-current backlogs. --Durin 17:15, 31 January 2007 (UTC)
        • indeed. I am just saying the "OMG CSD crisis" argument doesn't add anything in favour of those.
          dab (𒁳)
          17:50, 31 January 2007 (UTC)
          • I believe that is an inaccurate characterization of what other people in this thread have been saying. --Cyde Weys 19:06, 31 January 2007 (UTC)
• True, but more accureately we need more current admins to check on CSD regularly. Often I'll be the only one there deleting things at a slower pace then they're coming in (2 per minute or so). Granted, this is coming from a guy who had people dislike his speedy stance, but really it's not about needing more admins, just about getting them more active.--Wizardman 17:51, 31 January 2007 (UTC)

What about the fucking admins who were fucking promoted in the last six fucking months? Why aren't we fucking asking them to do the fucking work? In fact, if someone will fucking give me a fucking list of their fucking Usernames, I'll leave fucking messages on their fucking talk pages fucking asking them to fucking help. ;-) Anchoress 18:13, 31 January 2007 (UTC)

Here's your list. Get cracking! --Cyde Weys 18:17, 31 January 2007 (UTC)

No probs. Anchoress 18:22, 31 January 2007 (UTC)

As a non-admin who wouldn't mind getting digging into the CSD backlog if he had the tools, I can say that the appproach some people take when weighing in at RfA do make the whole thing pretty intimidating. At present I'm reluctant to think too hard about making an RfA with how hard some people are on candidates. Frankly, I kind of think we could rethink how we talk on RfA: for example, is it really necessary to write Strongest possible oppose for candidates you feel shouldn't be given adminship? I really think all this does is make the process more intimidating for users considering it. I'm not looking for sympathy here (heck, if I never get the tools, that's OK by me), I just think it's worth noting that the currently grueling nature of RfA does make it hard to approach and that it might be worth hearing from a non-admin about such things. Heimstern Läufer 18:28, 31 January 2007 (UTC)

true. if we had a working de-admining process, people wouldn't need to be so concerned about promoting the wrong guy, and it consequently RfA wouldn't need to be so intimidating for bona fide candidates. Alternatively, we could introduce "deletors" that have delete buttons but no block buttons. I would be most happy to give out delete buttons by the dozen as long as I don't have to consider if the person in question will know how to block users with the required decorum. User blocks as a rule are far more emotional than article deletions (at least in my experience), and we don't have a backlog of users in need of blocking.

dab (𒁳)

18:52, 31 January 2007 (UTC)

Some education of editors may be in order, though. How many of those speedy requests are without merit? I probably remove 5 speedy tags a day when I have the time to look due to them not meeting the criteria. The last thing we need are more "fucking admins" who can't read the criteria, given the amount of poor speedies in some aspects that have been discussed at

CSD's talk page. --badlydrawnjeff talk

18:55, 31 January 2007 (UTC)

This f'ing admin checks

80% (or more) of the work being done by 20% (or less) of the hands. -- nae'blis

19:57, 31 January 2007 (UTC)

I don't think he's arguing that people literally say we don't need more admins. Rather, they put that idea into practice by opposing RfA candidacies by applying standards that are either unreasonably strict or not relevant to adminship. Christopher Parham (talk) 20:54, 31 January 2007 (UTC)

• As of 21:50 UTC, the backlog is down to 10 items on
  CAT:CSD. Thus, in just under six hours we've basically completely wiped out the backlog. I think if this bot notified WP:AN, we'd see similar results. What do you say Cyde? --Durin
  21:52, 31 January 2007 (UTC)
  • I don't think so. It might work the first few times. After that, everyone would just start ignoring the messages. Having more admins is a much more sustainable solution. --Cyde Weys 16:11, 1 February 2007 (UTC)

• Looking at the "Images / Proposed deletion" having in some cases over a week's worth of old speedies not done says something about the real CSD backlog, IMO. Having them in subcategories doesn't mean there's no backlog currently. feydey 00:59, 1 February 2007 (UTC)

• There are $BIGNUM admins, assume that 20% do most of the admin actions, guess that a further 20% are posted missing. That still leaves a whole lot of people that could, with a bit of encouragement, be mobilised to spend some time on speedy deletion, prods, and the like. The quickest, and easiest, way to get more admins now, if that's really necessary, is to persuade inactive admins to become active, or semi-active. Likewise, anyone who was voluntarily demopped can be remopped just by asking, unless the rules have changed yet again. There's no need to leap off into uncharted territory, or to unleash evil hordes of adminbots, until the simple things have been tried. Angus McLellan (Talk) 22:06, 31 January 2007 (UTC)

We don't need more admins, we need less lazy ones, which is about everybody. —Pilotguy (radio check) 03:34, 1 February 2007 (UTC)

Being an admin doesn't mean you must use your tools. Rather, you are given the tools because you are trusted to not misuse them. People contribute where they find it fun to, that's the reality of a volunteer project. Oleg Alexandrov (talk) 16:05, 1 February 2007 (UTC)

Well said. We should promote potential part-time admins when they come up on RfA. Just because someone can't dedicate full time to this website doesn't mean they can't contribute as a good admin. Indeed, they are less likely to burn out than many very active admins are. Slow and steady sometimes does win the race ;). Simple rule: If someone is not going to misuse the tools, promote them, regardless of how active they are. NoSeptember 16:17, 1 February 2007 (UTC)

Precisely.--MONGO 05:41, 2 February 2007 (UTC)

It's a volunteer project though. There can't be any pressure on anyone to do anything because there's no payment involved. This characterization of "lazy" is unfair. --Cyde Weys 16:10, 1 February 2007 (UTC)

Matthew, last time i looked wikipedia had a voulantry userbase, you cant force pepole to become "non lazy admins" as you put it, users and therefore admins are not an expendeble commdity like hard drive space is, they are our most valued recource. Why should we remove admin powers from less active admins, it will do nothing but create a buercratic headache and resntment among the community because admin x fell 1 deletion short of there quota.

Uh, has anyone thought to dive into Category:Wikipedia administrator hopefuls? Looks like are 671 "fucking admin hopefuls" right there... EVula // talk // ☯ // 05:58, 2 February 2007 (UTC)

Every admin reading this who uses the pyWikipediaBot framework, please update to the latest version from CVS. I have coded a new bot named csd.py. It is the fastest semi-automated way to process speedy deletions that does not require the development of strong AI. Please tell me what you think. I've been using it to clear out CSD backlogs and it goes really, really, fast. --Cyde Weys 17:44, 31 January 2007 (UTC)

Oh yes, coming soon, better deletion summaries (based on the particular speedy deletion tag used). --Cyde Weys 17:48, 31 January 2007 (UTC)

great. we could also have a bot replacing unmet CSD with prod after some time, and deleting unchallenged prod articles after a week. Come on, a prod tag that stays on an article for a week? How bad can it be to have a bot delete those? Maybe require that the article is younger than a certain number of days so very smart vandals don't start prod-ing valid articles with little visibility.

dab (𒁳)

17:49, 31 January 2007 (UTC)

Very bad. I've seen many week-old prods that shouldn't have been deleted. Just because no one looks at a prod over the course of a week doesn't make it valid — remember, we do have over 1.5 million articles! We always need human eyes on it at the last step right before deletion. --Cyde Weys 17:51, 31 January 2007 (UTC)

ideally, yes. not below a certain content:crap ratio. csd/prod is added by a human, and if nobody challenges that, hey, it's a consensus. The human could just as well have blanked/redirected the article, and it would never have shown up on any queue. plus, articles can always be undeleted. This goes for new articles. Remember, Jimbo also blocked anons from creating new articles, a move that would have been unthinkable in, say, 2004. I agree that an article aged more than one month or so should not be deleted by a bot.

dab (𒁳)

17:56, 31 January 2007 (UTC)

One potentially untrusted person adding a prod cannot count as a consensus. If a human has blanked or redirected an article, that is simple vandalism, and can be fixed by anyone. But if it ends up being deleted as a result of vandalism — well, I'm sure you can see how this is a completely different situation. --Cyde Weys 18:02, 31 January 2007 (UTC)

Keep in mind that I've seen some prod tags on articles a couple years old that had maybe 1 edit every six months all by inactive users. Assuming they're going to magically come back to remove the tag is crazy, so I agree that it would need to be looked at manually.Wizardman 18:00, 31 January 2007 (UTC)

Automated PROD deletions is a bad idea. The policy behind PROD is that two people (the prodder and the admin) wanting an article gone, and nobody wanting it kept, is sufficient consensus to delete - if a bot replaces the admin, you end up with one person being enough to delete an article. --Tango 18:20, 31 January 2007 (UTC)

What of all the unautomated backlogs listed here? NoSeptember 18:25, 31 January 2007 (UTC)

I strongly agree that we mustn't automate prod deletions. Part of why we have admins is so human reasoning is required in the final to delete or not delete, especially in the case of a prod, where there has been no discussion. Heimstern Läufer 18:30, 31 January 2007 (UTC)

alright, alright, I sidetracked this. I agree this would be some sort of last resort solution, in case we really got buried in CSDs. As it is, I agree the cost/benefit ratio of such an approach is not attractive. Let's still try to make CDS handling as automated as possible. We want an admin to be able to deal with about 10 CSDs a minute (take 5 seconds to decide, click a button, next)

dab (𒁳)

18:38, 31 January 2007 (UTC)

csd.py does exactly this. Although if you're a fast reader/assessor (and have a fast net connection), you can go even faster ... Cyde Weys 18:40, 31 January 2007 (UTC)

so, dealing with the backlog of 500 CSDs would be a matter of 45 minutes? Surely we have enough admins to do that!

dab (𒁳)

18:47, 31 January 2007 (UTC)

Yeah, but it needs admins who use the pyWikipediaBot framework. Unfortunately, we don't have nearly as many of them as we should (which is what makes all of these opposers of Werdna's RFA even more ludicrous). --Cyde Weys 19:05, 31 January 2007 (UTC)

I don't think I agree that the final keep/delete decision on a page marked for deletion can be made in as little as 5 seconds. This is the final step before deletion and the admin hitting the button needs to be sure this is the right step to take. Yes, there are articles that can be fully reviewed in 5 seconds but I wouldn't say they are the majority of what's marked for deletion at any given time. Newyorkbrad 19:12, 31 January 2007 (UTC)

I agree, 5 seconds isn't enough time for non-obvious deletions. I wouldn't trust anyone to make a 5 second deletion on a regular basis. Many can be for sure, but we shouldn't be trying to make speed a core value in "speedy" deletions.

RxS

20:55, 31 January 2007 (UTC)

you make it sound as if "deletion" meant "wiping from the face of the earth". While all it is is setting a flag in the database. Honest mistakes can be made, no problem, it's a question of healthy balance between "establish watertight notability or we'll nuke it" and having pages of debate on every garage band. Anyway, I think we've established that the CSD backlog is no real problem. As soon as it starts to loom, start complaining about it and it will be gone within the hour :)

dab (𒁳)

22:11, 31 January 2007 (UTC)

Not at all, but try telling a new user that the flag got changed on that new article they created that they can't find anymore. I'm not concerned about the lost content so much because you're right, one way or another if it belongs we'll get it back somehow. I'm always more concerned about newbies who don't understand why their good faith contribution was deleted. If we did a better job with new users, deletions wouldn't be such a big deal.

RxS

22:41, 31 January 2007 (UTC)

Surely you don't think this is an acceptable "solution" (or a solution at all, for that matter). The backlog can balloon to hundreds of entries in just an hour or two. We can't constantly be yelling at everyone to deal with CSD. Hell, the only reason it worked here was because a lot of the people at

WT:RFA haven't done it in a loooong time, so they see this message and either figure, "Hey, I can help", or "Hey, let's shut Cyde up" (depending on how sinister they are, of course) and work on some CSDs. But there's the problem of diminishing returns. If I come back here tomorrow saying there's a CSD problem, and then the next day, and the next, eventually, nobody is going to be doing anything special about CSD based off the messages here. We need a sustainable solution. Yelling about the CSD backlog when it frequently gets too big just isn't sustainable. --Cyde Weys

22:29, 31 January 2007 (UTC)

Except that deleting a page is removing it from the face of the Earth. You can't count on the archive table holding deleted edits permanently.[1] Titoxd^(?!?) 01:06, 1 February 2007 (UTC)

Aah! (satisfied sigh) Thank-you Titoxd for quoting that quote from Brion. Can we please, please stop saying that deletion is not a big deal. If it is spotted and reveresed quickly, then there is no problem. But the deletion archives are (a) so MASSIVE and (b) prone to permanent deletion at some stage, so we must maintain high standards at the point of deletion. Carcharoth 13:56, 2 February 2007 (UTC)

So Cyde, where is this new bot you've created? -

I'm wondering if it's possible to automate the process of leaving notices on the pages of users who created articles that get deleted, particularly the blatant nonsense articles. I have a feeling that some users are being ignored who could benefit by being given notice that their content was deleted, and why.

IronDuke

01:15, 1 February 2007 (UTC)

"Every admin reading this who uses the pyWikipediaBot framework, please update to the latest version from CVS." Seems pretty self-explanatory. Ohh, and the bot has been renamed to speedy_delete.py. --Cyde Weys 01:19, 1 February 2007 (UTC)

I'm not an admin (and I don't want to sound dumb), but where's the CVS?

Yuser31415 (Editor review two!)

01:31, 1 February 2007 (UTC)

Here --Cyde Weys 01:45, 1 February 2007 (UTC)

Thanks .

Yuser31415 (Editor review two!)

02:00, 1 February 2007 (UTC)

Nevertheless, I'm quite impressed with pywikipedia bot, so I'll head over to the relevant page and get myself permission to run it, and consider enhancing it to suit my needs better.-gadfium 01:47, 1 February 2007 (UTC)

CVS is the inferior version control repository to RCS. :) (flamesuit on)

Just an observation on the section title: What I don't need is any fucking automation. Python is good for a lot of things but fucking isn't one of them. I prefer to use Python in totally different ways. Cyde's mileage may vary, but I don't thing I will be downloading any fucking modules regardless of supposedly how good the latest version is. How do we know if it's viruse free? The internet is cool but not that cool. To quote my Marine drill sergeant "This is my Python, this is my snake, this one is for the internet and this is for making stupid rhymes." --Tbeatty 01:49, 1 February 2007 (UTC)

Um, you can't download compiled code, only the source code. He doesn't have any virus embedded in the code I've looked at. (And a gigantic number of people review the code, besides me.) Dozens, perhaps hundreds of people run parts of the pywikipediabot framework. Any attempt at dropping in a virus would have been caught years ago. --Connel MacKenzie - wikt 01:08, 2 February 2007 (UTC)

It isn't that hard to automate four image backlog categories:

Oh, and aren't we getting a little off topic? MER-C 12:27, 1 February 2007 (UTC)

If the user is trusted enough to instruct a bot to delete an image, then they should just become an admin. --BigDT 12:39, 1 February 2007 (UTC)

I can't see why we need "fucking" need any more administrators, we just need more non-lazy ones who don't make it their goal to get the extra powers.. and then over time reduce there activity to nothingness. If the powers where say remove from the lazy/inactive I'm sure that would be a good kick-up-the-ass to get other admins doing some work and/or make a real need for administrators. thanks/Fenton, Matthew Lexic Dark 52278 Alpha 771 12:47, 1 February 2007 (UTC)

Accusing others of laziness isn't helping here, Matthew. You personally have never handled any CSDs. --Cyde Weys 16:14, 1 February 2007 (UTC)

• We definitely need more automation. Although
  CAT:MNT) and most of them keep getting bigger. Most of them are also simple drudgework, only lots 'n lots of it. More bots is t3h win. >Radiant<
  13:04, 1 February 2007 (UTC)
  • Almost all of the backlog you've just linked doesn't need administrator attention. Only a thousand or two out of over 100,000 items actually need the tools to deal with. --Cyde Weys 16:22, 1 February 2007 (UTC)
    • Indeed, hence my notice that we need more automation. >Radiant< 12:52, 2 February 2007 (UTC)
    • And many of the things in
      CAT:MNT are miscategorized and don't even need attention. The first two I clicked into were the category for soft redirects, and the category for stable version requests (which should not be happening for a rejected proposal). -Amark moo!
      00:35, 2 February 2007 (UTC)

...Or has the number of candidates who are "underqualified" increased recently? bibliomaniac15 05:12, 2 February 2007 (UTC)

I think the snow has melted due to global warming... MER-C 08:10, 2 February 2007 (UTC)

From my close to two years experience in following RfAs and RfArbs, I've got the impression that the current RfA process works extremely well in the sense that very few people who got promoted end up misusing the admin tools consistently, and those doing so have been very orderly de-adminned by the ArbCom.

Perhaps the current process works too well, by sometimes filtering out people who may not satisfy all the requirements du jour for being an admin but who still are mature enough to not misuse the tools (I don't have hard data for these statements though).

Anyway, I'd think that by letting bureaucrats promote over 70% we have more to gain than to lose. Bureaucrats would of course still be able to not promote people who are over 70% but whose opposers raise serious concerns.

So, what do people think of this proposed reform? Oleg Alexandrov (talk) 05:39, 2 February 2007 (UTC)

I agree, with emphasis on the opposes; in RfAs with large numbers, there are usually several big concerns raised, which are drowned in a sea of supports. — Deckiller 06:00, 2 February 2007 (UTC)

I think that the process for dealing with admin promotions (RFA) works better than the process for dealing with admin abuse (

talk

) 06:40, 2 February 2007 (UTC)

I strongly agree (I tried to bring this up in the past, but was shot down). Our current criteria are too high and lead to many good candidates failing at 70%+. The filtering process of RfA works very well (and in any case is independent of the support percentage: those recently forcibly desysopped all had more than 80%, sometimes more than 90% support on their RfAs). However, RfA is not only supposed to filter out bad candidates, it it supposed to promote all trustworthy candidates. It is not currently doing a very good job at promoting all of them, and lowering the promotion threshold is an easy way to increase the number of administrators. Looking at who failed with 70%+ recently, I don't think we'd be sacrificing too much quality. Kusma (討論) 06:44, 2 February 2007 (UTC)

I like the idea of lowering the threshold too. I think the standards that are frequently used and that a lot of RfAs fail because of (high edit count, percentages of edits in various namespaces...) are kind of irrelevant to the real question of whether the user's going to abuse the tools. These standards seem to me to have been going up for a while, so it's been getting harder to become an admin while our need for admins has also been increasing. In the spirit of no big deal-ness, I think it's important to give out the tools to qualified folks without too much difficulty, to prevent adminship from being treated as a status symbol. However, I think the points that Renesis13 brings up are valid and important, and I think it's important to work on making it easier to get folks de-adminned as well, and to figure out other means for dealing with these issues. Obviously that will take time, but I don't think that keeping the standards so high is a direct or effective enough way to deal with the problem. For example, someone could be on their best behavior until they're adminned and then let loose. So I suggest lowering the margin, and also working more directly on the problems of admin abuse. delldot | talk 07:31, 2 February 2007 (UTC)

Hmmmm, in early 2003, there were probably < 100k articles on English Wikipedia, and there certainly was no 3RR rule and probably a negligible amount of edit wars, and thus, in those days nobody will get on the wrong end of the stick because of a religious or ethnic dispute and get in trouble for it: Wikipedia:Requests for adminship/Jersey Devil 2, Wikipedia:Requests for adminship/Soman and Wikipedia:Requests for adminship/Dev920 in the last week all had a bit of damage due to being in content disputes. In early 2003, probably only the "important" articles were created, so don't get into trouble for being an (extreme) inclusionist or deletionist. Now you do, because different admins will close things differently....In the old days, good faith users probably never disagreed, because WP was empty and people were only adding new content, so everybody who wasn't a vandal became an admin in order to block them, but these days, not so. That's life I guess. Blnguyen (bananabucket) 08:19, 2 February 2007 (UTC)

Probably the only thing that couldn't be disputed/interpreted differently is deleting unsourced/unlicensed/orphaned fair use images (AfDs, etc, blocks, and protects can be debated) - I guess having a separate image deletion tool given out liberally cool speed things up, since the images are always 7+ days overdue now that Kimchi.sg is away. Having said that, how many people would want to delete them? Blnguyen (bananabucket) 08:24, 2 February 2007 (UTC)

I think 70% would be fine, so long as another stipulation for adminship was that candidates had to be editing seriously and usefully for a year before being considered. That way one would know the candidate better at the time of voting

Giano

08:53, 2 February 2007 (UTC)

Personally I think Renesis and I strongly oppose a lowering of standards.Administrators are expected to behave as model wikipedians at all times, and if there is a large percentage of people (30% can be quite a lot of people) expressing a view that the person would be unsuitible for the position, I see no reason why the candidate cannot remedy reasons for opposes and try to gain the required support. Viridae^Talk 09:28, 2 February 2007 (UTC)

Is 70% after a week of discussion considered to be consensus? Gizza^{Chat ©} 09:41, 2 February 2007 (UTC)

I don't think 70% is consensus, just as 80% is not consensus. It is just a supermajority. I don't think a (200/50/70) RfA can be said to have reached consensus, but we should promote the candidate anyway. Kusma (討論) 09:48, 2 February 2007 (UTC)

And personally I dont think 70% is a super enough of a majority. Viridae^Talk 09:50, 2 February 2007 (UTC)

I don't care what the exact number is. I think we currently do not promote enough administrators, and I think that is because we filter too strictly. Kusma (討論) 09:52, 2 February 2007 (UTC)

The current system discourages diversity and progress, because it favours people who have not pushed any boundaries. On the other hand, we should consider the possibilty that it doesn't matter that much who the administators are, as long as they behave themselves. Stephen B Streater 10:08, 2 February 2007 (UTC)

Define pushing boundries - I can't see why we should lower the % - if it aint broke why fix it. Werdna (as an example) right now has about 70% support - yet 80 people have opposed citing some serious issues. 80 opposers hardly seems consensus to me Glen 10:15, 2 February 2007 (UTC)

Indeed, but do you think it would magically become consensus if he had 240 or 320 supporters? Kusma (討論) 10:29, 2 February 2007 (UTC)

Note that I am arguing that the 80% mark is broken. The number of articles and editors is growing a lot faster than the number of admins, and we should fix that problem. Kusma (討論) 10:34, 2 February 2007 (UTC)

Have a look at my own RfA, and you'll see what I mean by pushing boundaries. In such a varied place as this, any visionary change will not achieve consensus until the man-on-the-street has caught up with the visionaries. That's not necessarily a bad thing, but a handful of fanatics can lose sight of the fact that Wikipedia was set up by visionaries, not plodders. By spreading FUD, they can derail any RfA which is based around a radical agenda, however positive or necessary to the long term success of the project it is. Stephen B Streater 18:40, 2 February 2007 (UTC)

(de-indent) Is it that 80% is too high a percentage, or is it that too many people (particularly non-admins) oppose for ludicrously nitpicky reasons (edit count, and the even more pernicious 'balance of edits across the namespace'), yet their oppose still carries as much weight as four well-reasoned and thought out supports? Bureaucrats aren't just bean counters, they were put in place for a reason, let them ignore any stupid opposes or supports, bringing it closer to how AFD works. This would also, however, require people to actually put some kind of explanation to their support vote.

►

10:41, 2 February 2007 (UTC)

I agree with this, but it seems pretty clear to me that the crats are bean-counting -- possibly because if they act otherwise they get twenty sections of complaints on WT:RFA. This seems like a simpler way to work around the problem that takes the pressure off the bureaucrats. Christopher Parham (talk) 21:28, 2 February 2007 (UTC)

Kusma if there was 300 supports there would be a mindblowing 90 opposes, which I doubt would make it through. A beuracrat cannot look at that and think it is a good idea to promote unless all the opposes are suspected socks disruptors etc. Viridae^Talk 10:43, 2 February 2007 (UTC)

It is part of the bureaucrat's job to consider whether apparently ignoring the concerns of 90 people in order to promote one person is going to be good or bad for Wikipedia. Kusma (討論) 11:11, 2 February 2007 (UTC)

Proto, generally supports sound pretty lame. "seems to be good, I see no reason not to support" there is no real argument or reason to give most of the time. Viridae^Talk 10:45, 2 February 2007 (UTC)

Many support or oppose votes are pile-ons "per X" or "per nom". You can't expect people to make a good argument for every vote. I do support large bureaucrat discretion, but I don't think they should explain which votes they discounted or why. Anyway, the 80% rule was made up some years ago, when RfA was a much less stressful process for the candidates and people did not pile on opposes for silly reasons. I think that adapting the cut-off percentage could be a good way to adapt the procedure to the changes in the way people vote. Kusma (討論) 10:48, 2 February 2007 (UTC)

Note I do also support lowering the margin for bureaucrat discression to 70-80%, from 75-80. Given that every admin who has been desysopped had more than 80% approval, there's no evidence that the borderline candidates are worse at administrative duties than those who passed with flying colours. In fact, given that the borderline candidates have never had a de-sysopping amongst their numbers, and those who got 90%+ have, you could even draw the conclusion that those who barely passed are on average 'better admins'.

►

10:53, 2 February 2007 (UTC)

What about

this one? NoSeptember

11:28, 2 February 2007 (UTC)

Would it be churlish to point out the 8th oppose vote? :)

►

12:12, 2 February 2007 (UTC)

Was that a sneaky way to get us to click on the link in your vote? :) NoSeptember 12:19, 2 February 2007 (UTC)

Did the concern of most of the oppose voters (cares too much about edit counts) have any relation to the desysopping? Kusma (討論) 11:55, 2 February 2007 (UTC)

Since we can't see into the future I doubt many issues raised in an RfA directly relate to the case that results in a desysopping. We've had people pass by 100% and then get desysopped, but that doesn't mean we can't trust those shady unanimously supported users ;). Though it may suggest that we have work to do on focusing our standards on things that really matter to what makes someone a good admin. NoSeptember 12:08, 2 February 2007 (UTC)

Or at least finding some consensus as to what isn't important to being a good admin. For instance 1FA was rejected as a reason to oppose, but some people still feel article writing is important. At the moment, everyone sets their own standards. Trebor 12:12, 2 February 2007 (UTC)

And these standards vary so wildly that it is getting very hard to get more than 80% support votes. Many people who passed RfA a year ago might not pass now (for example, people might oppose me for lack of image experience). Instead of changing everybody's standards, changing the promotion threshold is an easier way to get the job done (more admins). Kusma (討論) 12:32, 2 February 2007 (UTC)

Yeah to the first bit. I think this is most clearly evidenced where people who are obviously suitable to be adminned frequently turn down the first (or second, or third) nomination. Although users who have interacted with them know they would be good (or even great) admins, they may not have the "right" number of edits or they may not have been here for "long enough" or they may not contribute enough articles or they may not want to appear "too eager" for the bit, etc. RfA seems to be becoming increasingly hard to pass; I think the ethos should be "why shouldn't we give this user the tools" rather than "why should we". Trebor 12:51, 2 February 2007 (UTC)

• 70%? Hey, Bishzilla currently at 65%, give tools now! Geogre 11:31, 2 February 2007 (UTC)

LOL! Now this is how to vote-stack. Carcharoth 14:01, 2 February 2007 (UTC)

I'm not really in favour of changing this. The current ratio is 1:4 which shouldn't be insurmountable. I'd be hesitant to see the margin drop below 75%. That's Only 1:3. 70% is not much more than 1:2. But perhaps people should be encouraged to take this apparent need for admins into account when voting? I guess it doesn't really matter anyway. If we do drop the bar, I predict we'll see voters being more ready to vote oppose and less ready to vote support. Regards, Ben Aveling 12:59, 2 February 2007 (UTC)

And furthermore margin should be 0%, because I'm in favor of bureaucrats using brains (and people thus being forced to discuss more and better). --Kim Bruning 12:38, 2 February 2007 (UTC)_{(And give the bishzilla her broom and scales and death-ray already O:-) )}

As we expand most editors are going to become involved with more disputes their RFA will attract more oppose votes. I think that admin tools should become semi automatic. Any editor who's been here for a specific time and has a minimum number of edits can get the tools providing that two other admins support the request. Then all thats needed is quicker reversal process that can remove the editors tools if they are being abused. Gnangarra 12:53, 2 February 2007 (UTC)

• It is up to the crats whether they agree to lower or bar or not. It is not up to us. -
  Mailer Diablo
  13:07, 2 February 2007 (UTC)

  The power to lower the bar is with the crats. That doesn't mean that a community consensus to lower the bar won't influence them. Kusma (討論) 13:12, 2 February 2007 (UTC)

  Well, it's clear that community consensus to lower the bar is nowhere near 75%—probably not even 50%. Despite being an advocate in one of the borderline high-participation RFAs that didn't pass, I think if a candidate can't get at least 75% support then they shouldn't be made an admin and wouldn't support lowering the threshold. —Doug Bell ^talk 19:24, 2 February 2007 (UTC)

  I strongly disagree with that statement. The bureaucrats exist to perform a function; that does not mean they decide what the qualifications for adminship are. That is in the hands of the community. --

  talk

  ) 17:59, 3 February 2007 (UTC)