Wikipedia talk:Requests for adminship/Defining the problems

Define the problems first?
I think from comments by User:Durin and others that we need to define exactly what problems we have with RFA now. I think that the best way to do that is probably by subsections under this, with commentary under each subsection, so that as we progress we can see the whole list of problems clearly. - Stephanie Daugherty (Triona) - Talk - Comment - 21:18, 24 September 2006 (UTC)

(I've started with a couple observations, PLEASE ADD MORE) - Stephanie Daugherty (Triona) - Talk - Comment - 21:18, 24 September 2006 (UTC)

"Adminship shouldn't be a big deal"
Somewhere along the line, the "adminship is no big deal" text was removed from RFA. It really isn't a big deal, as admin actions are reversable. The process we have now resembles senate confirmation hearings. Its too complex, too political, and fails to focus on the main issue, which should be trust. - Stephanie Daugherty (Triona) - Talk - Comment - 21:18, 24 September 2006 (UTC)


 * That is, unfortunately, not the case (as a number of quite recent situations have amply demonstrated). While blocks may be reversible in the technical sense (although there are some subtleties even in this regard), the personal effects of being blocked on the subject of the block are often not.  A great many of our problems seem to stem from certain administrators believing that blocking someone is "no big deal" and acting accordingly. Kirill Lokshin 21:51, 24 September 2006 (UTC)


 * In what way do you think people are getting rejected that does not relate to trust? Some people don't trust users with low edit counts, others don't trust users with civility problems, maturity issues for others and yet others lack of knowledge re: copyright etc. All of these relate to a lack of trust. The one criteria that might be questionable is 1FA but even those users invoke it as they don't trust a candidates editing knowledge enough for them to be an admin. Qualities in a editor that cause you trust some one are not necessarily reflected by others. Are there any good examples of RfA candidates being blackballed for a political reason? David D. (Talk) 03:32, 26 September 2006 (UTC)


 * Experience and trust are two seperate things. Editors and admins with lack of experience make mistakes. Mistakes are no big deal, we revert them, discuss them, and move on. Trust is an issue of whether or not we expect someone to use the tools maliciously or abusively. There is a relationship between the two, but an editor with less experience may be very much trustworthy. The worst an admin can do accidently isn't very much - the worst an admin can do maliciously, well that's another story. - Stephanie Daugherty (Triona) - Talk - Comment - 04:45, 26 September 2006 (UTC)


 * Further... PrincipleOfFirstTrust and WP:AGF should be required reading here. Trust and good faith should be the default in a Wiki. - Stephanie Daugherty (Triona) - Talk - Comment - 04:47, 26 September 2006 (UTC)


 * I would still like to see an example of an RfA that you think was unfair due to politics. What kind of potential admins are not gettng through the process? Also, AGF and trust could mean very different things to various people.  You're of the opinon that anyone that can edit and has shown trustworthy behaviour should be an admin, others have a higher bar.  There is nothing that makes that higher bar unreasonable. David D. (Talk) 05:28, 26 September 2006 (UTC)
 * Perhaps it's wrong for me to use my self as an example, but: Requests for adminship/AzaToth → A z a  Toth 15:33, 26 September 2006 (UTC)
 * Not wrong, i'll take a look at it. David D. (Talk) 15:40, 26 September 2006 (UTC)
 * I'd also like to point out Ambuj.Saxena's RfA. &mdash; Werdna talk criticism 06:47, 27 September 2006 (UTC)

The lines between rejection, no consensus, and acceptance are too blurry
I think this was an intentional part of the design, since this was originally not supposed to be a vote, however, with RFA having long ago turned into a de-facto voting process by its participants, this now causes heated debates whenever someone doesn't agree with the closing 'crat's interpetation of the discussion. - Stephanie Daugherty (Triona) - Talk - Comment - 21:18, 24 September 2006 (UTC)
 * I believe that the line should be left blurred. One of this great things about Wikipedia (in all aspects) is that rules are not set in stone. Wikipedia is not mechanical (i.e. if the bar is set at 80% for consensus and the editor gets 79.9999% approval, the RfA fails), and should be allowed a decent range for individual interpretation. --physicq210 21:23, 24 September 2006 (UTC)
 * Agreed. What's the problem with heated debates?  If you get annoyed, discuss it.  Ral315 (talk) 05:24, 25 September 2006 (UTC)
 * it is intentionally left to the bureaucrats to make an informed good faith decision. this has worked very well in the past. The guidelines to be observed by the bureaucrats is that 80% is sufficient for a consensus. I believe it happens more often that a 75% result is taken as a "successful" RFA than a 85% as a "failed" one, althogh both would in principle be arguable. Bureaucrats should be very careful not to indulge in opaque or partisan decisions. It wouldn't hurt to impose solid limits of 70%-90% on their good judgement: it would be very hard to justifiy failure in the face of a >90% result, or "success" in the face of a <70% majority. This did not need pointing out in the past, and in the sprit of "don't fix it unless it is broken", things worked fine. But now we have a precedent of bureaucrats considering a 61% majority a "success", so this needs to be addressed. Either RFAs are simple majority votes, in which case 61% is obviously good enough, or we stick with the "consensus" system, in which case it obviously isn't. Seeing that we are not in desperate needs of more admins, I think it would be a terrible mistake to move from "ca. 80% consensus" to "simple majority". And if bureaucrats begin to make too much of their power to interpret results, it will sadly be necessary to impose on it a fixed limit (such as the 70-90% range I mentioned). dab (&#5839;) 08:24, 25 September 2006 (UTC)
 * I hate to be pedantic - consensus isn't about a defined value, its about general agreement on the action to take. Simple majority would mean 50% or greater. "Supermajority" is usally defined proceedurally as either 2/3 (66%) or more depending on the voting body. - Stephanie Daugherty (Triona) - Talk - Comment - 08:49, 25 September 2006 (UTC)

Admin accountability
I believe that the RfA process is under many pressures partly because there is no similar organized way to hold admins accountable for their actions or desysopp them if necessary. If we answer this latter question, we will see more people ready to say "adminship is no big deal." The consensus issue will also become less sensitive, and I'm sure that stuff like the Carnildo affair won't happen. To get the ball rollin':


 * 1) Blocks are serious, last resort measures. It can be useful to desysopp an admin who has been blocked (justifiably) by a fellow admin, but for just like a week or so to allow him/her to cool off - this is different than the "cool off" mandated by blocks themselves. The second block should result in a 30-day "cool off," and the third block should lead to desysopping.
 * 2) WP:ANI should also become a place where admins can receive advice and suggestions on how to perform potentially sensitive tasks. There are many admins who aren't abusive but simply make a mistake - they need immediate guidance of more experienced admins.
 * 3) Maybe a bureaucrat's committee can be created (subordinate to ArbCom) to look into cases concerning proposal no. 1 and cases of serious admin abuse. Rama's arrow  10:23, 25 September 2006 (UTC)
 * 4) Perhaps admins should be encouraged/required to hold an editor review each year to receive suggestions and criticism.
 * 5) More and more admins should be encouraged/required to help at Admin Coaching. This way, they get to improve themselves and learn how to help others.
 * 6) Promote greater teamwork and coordination between admins. If you create a WikiProject Admintasks or a similar device to inform/coordinate admins on backlogs and serious issues, we will see greater efficiency as well as a culture in which admins don't feel compelled to make sensitive decisions themselves, in their own interpretation of policy. Rama's arrow  10:26, 25 September 2006 (UTC)


 * That said, this process should not revolve around the question of desysopping. Rather how admins can improve and learn. The same for RfA - every productive editor is a prospective admin, so the process should encourage learning instead of creating a contest, vote-like atmosphere. Rama's arrow  10:39, 25 September 2006 (UTC)


 * I think (1) and (3) are unnecessary, Arbcom is sufficient for such things. (2) sounds good. (4) has been discussed before and won't work - there are over 1000 admins, assuming each review takes a week, that means there will be around 20 going on at any given time. It works for smaller wikipedias but en is just too big. I'm open to pursuasion on the last 2, but I'm not sure they would have any great effect. --Tango 11:28, 25 September 2006 (UTC)


 * Actually (5) and (6) are the more-difficult-to-implement points. (1) is necessary - there needs to be a line in the sand even for the most experienced admins. I think (1) is quite liberal in permitting as many as 2 blockings. I feel admins should be the last people here to have to be blocked. Rama's arrow  11:34, 25 September 2006 (UTC)


 * How would being desysopped help you cool off? Usually it's arguments and name-calling that cause people to heat up. This wouldn't address that problem. Christopher Parham (talk) 18:16, 25 September 2006 (UTC)


 * We need an active role for bureaucrats in this - it is important that they act in a system of checks and balances. Rama's arrow  11:37, 25 September 2006 (UTC)


 * I don't see a need for fixed penalties - if an admin is getting blocked, then it's not routine and shouldn't be treated as routine. I think ArbCom should be left to decide punishments on a case by case basis as they do now.


 * Bureaucrats were selected to assess community opinion, not to be any kind of decision making body. ArbCom were selected for that purpose, so that's the body we should use for it. --Tango 12:03, 25 September 2006 (UTC)


 * My dear sir, reform has to start firmly. We need change, and we need to do what's necessary. Now I don't really care if we enact (1) or not, but the reform will be effective only if there is a line admins musn't cross at any cost. We can't expect the problem to rectify itself. Rama's arrow  12:32, 25 September 2006 (UTC)


 * There is also a set of suggestions proposed by Dab at WP:VPP. Rama's arrow  12:32, 25 September 2006 (UTC)


 * Pasted from WP:VPP:

dab (&#5839;) 12:12, 25 September 2006 (UTC)
 * we need a working WP:RFDA. We may not have needed it in the past, but we do now.
 * CAT:AOR should be understood as a matter of course. Admins too enamoured with their position of power, or too deluded with their own importance to subscribe to this shouldn't keep their buttons.
 * IRC is not part of WP. An editor's merits and admin-worthiness are exclusively a function of their on-wiki work and behaviour.
 * admins are WP editors trusted with some minimal good judgement. Tendencies to re-cast them as some sort of separate caste need to be firmly countered. Admins who have ceased to see themselves as members of the community of editors should hand in their mop.
 * CAT:AOR is meaningless. By its present wording, an admin "open to recall" may decide suffrage in whatever way he likes, and if six people "with suffrage" want his recall he can still do whatever he likes, e.g. open an RFC on himself (but RFCs have no power to enforce anything) or send the matter to RFAr (which is doubly meaningless since (1) you cannot open an RFAr on yourself, and (2) it doesn't take six people to open an arbitration case).  &gt; R a d i a n t &lt;  18:26, 25 September 2006 (UTC)


 * The core of CAT:AOR is that its members "are willing to stand for re-confirmation of adminship if six editors in good standing request it." The peripheral aspects&mdash;the precise definition of "good standing" and venue for reconfirmation&mdash;are flexible, but members are expected to interpret them in good faith.  Proposals for rigid, mandatory recall have a poor track record; CAT:AOR has survived and flourished because it is voluntary and flexible.  As it grows and as we learn from experience what works and what doesn't, it may be desirable to set more specific criteria.  But even now, good-faith membership in CAT:AOR is an important and meaningful commitment, and I commend any admin who undertakes it. Tim Smith 04:40, 26 September 2006 (UTC)


 * Any RFDA proposal that makes desysopping a popularity contest is inherently flawed. Administrators will often be expected to perform actions that make them unpopular. I strongly agree that administrators are expected to be part of the community and function well within it, but to expect that every good-faith administrative action will be agreed with by the community is unreasonable.
 * There are currently several areas of policy where community consensus doesn't necessarily apply. WP:V, WP:NOR, WP:NPOV, WP:BLP, WP:NPA, Right to vanish, Privacy policy, WP:C are among these policies - actions under them, if technically correct, aren't necessarily subject to public opinion, and there are cases where a position and course of action that violates these policies will be defended by popular support. Part of an administrator's job *IS* to enforce these rules, regardless of whether or not the community wants them to apply to a particular case. There are only a handful of administrators willing to get their hands dirty in such matters, but it is essential for the integrity of an encyclopedia that we have, and follow such rules, and administrators should not be subject to deadminship simply for following policy in good faith, regardless of how contentious the particular application is.
 * The current WP:RFAr mechinisms work, as does emergency action by a steward in the case of mass-disruption or security matters. I see no need to change this, although I do see a strong need for involving administrators in, and making administrators subject to dispute resolution. - Stephanie Daugherty (Triona) - Talk - Comment - 04:06, 26 September 2006 (UTC)


 * Tim, I get that and it's a good idea in theory, but to me it seems that in practice it doesn't really work that way. — Preceding unsigned comment added by Radiant! (talk • contribs) 08:39, 26 September 2006 UTC

RFA is often a grudge match
Rather than being a test of whether or not someone can be trusted with the tools, its often a place where anyone with a grudge comes to air it, and people seem to hold grudges for a long time. - Stephanie Daugherty (Triona) - Talk - Comment - 19:47, 2 October 2006 (UTC)

Pile-ons
One problem facing RfA at the moment is that of a pile-on; there are too many votes saying the same thing. Votes which don't bring up new points may be useful in gauging the proportion of RfA regulars who approve of a candidate, but they aren't very usefull in determining consensus. This is especially true of support votes; normally there isn't anything to say but 'per nom' or just 'Support', but the proportion of support to oppose votes is what matters in deciding the decision. --ais523 15:00, 4 October 2006 (UTC)

Spurious reasoning
Occasionally, there are RfA votes based on reasoning much of the community would disagree with (the parody 200MWTE condition somewhere in the WT:RFA archives explains this better). It isn't clear how much discretion is allowed in ignoring these, or even whether it's fair to ignore them. This is especially true with spurious-vote pile-ons (for instance, inexperienced users opposing a candidate due to the candidate enforcing policy they didn't like); whilst rare, RfA would be better if this sort of thing didn't skew the outcome of votes. (One possible solution would be to move to a less vote-based system.) --ais523 15:00, 4 October 2006 (UTC)

Insufficient research by participants
Many participants do little more than a cursory glance of the RfA before making their opinions known. This results in ill-informed decisions based on spurious, non-existent, or contextless evidence. Moving away from voting is a solution to this problem. &mdash; Werdna talk criticism 15:07, 4 October 2006 (UTC)
 * I'm the opposite in that regard; I often do quite a bit of research and then can't make up my mind and don't vote. You have made a valid point, though; this is probably the reason why the fixed percentage that is designated to mean 'consensus' isn't lower. There are still too many users who have a fixed list of standards based on edit counts or suchlike, and just use various tools to find out whether the candidate satisfies them. (This is why recently, when I've use my counter on an adminship candidate, I've been generating diffs (encourages research and highlights areas) rather than just counting (even a detailed count isn't enough, and counts stifle further research).) I agree that moving away from voting will solve this problem more or less altogether. --ais523 15:16, 4 October 2006 (UTC)

Lack of hard standards
The current adminship process is confused by the lack of firm standards for qualification. - Stephanie Daugherty (Triona) - Talk - Comment - 08:57, 5 October 2006 (UTC)

Drift in standards
The unofficial "standards" of qualification drift increasingly torwards tougher and tougher requirements, with little rationale, and tougher standards tend to get adopted by participants over time without understanding if they are needed or why they are needed - Stephanie Daugherty (Triona) - Talk - Comment - 08:57, 5 October 2006 (UTC)

Removal of administrators
The process for removing administrators, while established, isn't well understood by the community, and is seen as an extraordinary measure. The percieved difficulty in removing an administrator leads   many participants in the process to be overly critical, since adminship is rarely taken away. - Stephanie Daugherty (Triona) - Talk - Comment - 09:06, 5 October 2006 (UTC)