Wikipedia talk:Adminship renewal

Bureaucracy

While I think this proposal is OK "in theory," (meaning: I think the idea of reviewing admin privileges is a good one) I worry about the additional administrative load (not just on bureaucrats, but on the community as a whole) from dealing with hundreds of adminship renewals each year. Does anyone have an idea about ways to limit the negative impact of this? Nandesuka 13:22, 14 May 2006 (UTC)[reply]

The bureaucracts can do that - we can elect more if there is a need. It's worth the trouble in exchange for a lot less of problems caused by admin abuse, loss of trust, etc. - Mailer Diablo 14:56, 14 May 2006 (UTC)[reply]

This would mean approximately 800 renewals a year, perhaps growing to a thousand or more. That's twenty a week. The burden isn't just on bureaucrats, but the whole community. I barely find the time to vote on a few RfArs a month. I'd hate to have to spend more time just maintaining the status quo. -Will Beback 05:19, 17 May 2006 (UTC)[reply]

Support in principle

I believe a process similar to this, and perhaps actually this one, is a good idea. Even though I disagree with the way Ardenn has been using recent AfD's to make a point his underlying argument is valid. I have come across at least two current admins who have been borderline abusing their role. Not enough to take to Arbcom, but I suspect they would behave quite differently if they knew they would be held to account at the end of the year. Good systems tend towards openess and accountability. Currently Wikipedia is open, but lacks accountability.

I need more time to think about this specific proposal, but I certainly support it in principle. Gw e rnol 13:27, 14 May 2006 (UTC)[reply]

Thanks! Glad to hear your comments. - Mailer Diablo 13:39, 14 May 2006 (UTC)[reply]
Don't you mean RFAs, not AFDs? Stifle (talk) 21:38, 14 May 2006 (UTC)[reply]
- Yes I do :-) For some reason I often get those Wikronyms transposed. Thanks for catching that. Gw e rnol 21:40, 14 May 2006 (UTC)[reply]

problems

Been suggested before the problems are:

The numbers involved. We have 909 admins at the moment. that means 2.4 a day or 17.4 per week. that is a heck of a lot of votes.
- If there is sufficient interest, I think getting people to vote would not be a problem. Wikipedians are more than happy to endorse sysops in which they are satisifed with their actions. - Mailer Diablo 13:59, 14 May 2006 (UTC)[reply]
  - It isn't a question of interest. It is a question of the work involved.Geni 14:05, 14 May 2006 (UTC)[reply]
There are certian admin tasks which are going to make you unpopular. We would rather admins didn't have to wounder if they were popular enough to carry out a certian action
- Which is why it is 50%. - Mailer Diablo 13:38, 14 May 2006 (UTC)[reply]
  - Have you any idea how many people you can annoy by inforceing the 3RR. And then there is the whole copyright thing which tends to upset people.Geni 13:45, 14 May 2006 (UTC)[reply]
    - You mean Carnildo - he actually made the 50% cut (Assuming no neutral, and given he's just desysopped), something to think about. - Mailer Diablo 13:53, 14 May 2006 (UTC)[reply]
      - No there are other cases. The problem is either you set the barrier too pass so low that it becomes a meanless resource burning rubber staping thing or you set it high enough to start knocking out admins simply for doing unpopular jobs.Geni 14:05, 14 May 2006 (UTC)[reply]
        Considering that there can be two other extremes - Consensus to retain sysophood, or consensus to remove sysophood. - Mailer Diablo 14:23, 14 May 2006 (UTC)[reply]
what is my renewal date?Geni 13:28, 14 May 2006 (UTC)[reply]
- Theorically, 15 December 2006. - Mailer Diablo 13:38, 14 May 2006 (UTC)[reply]
  - except I was last readmined on the 10th March 2006.Geni 13:45, 14 May 2006 (UTC)[reply]
    - Oh, Danny. :P It'll still be December. - Mailer Diablo 13:51, 14 May 2006 (UTC)[reply]
      - well what about the 1 March 2005 date? The break there was slightly longer.Geni 14:05, 14 May 2006 (UTC)[reply]
        Ed Poor, lol! Still the same, unless it was re-sysopping by re-application. - Mailer Diablo 14:23, 14 May 2006 (UTC)[reply]

More ideas

Geni raises some excellent points. There are interesting (if imperfect) real-life parallels in countries like the USA where sheriffs are voted into office. Here are some ideas that may be worth considering:

At renewal time, any user can comment, but only admins vote on renewal. The advantage is that all editors get to give input on an admin, but trolls and vandals can't overwhelm the process. However, this is counter to the way Wikipedia works now (at least in RfA) and could introduce valid concerns about cabals/cliques. A completely open renewal process could disuade admins from making the important, controversial decisions as Geni says.

Automatically de-sysop admins after a year. Admins would be encourage to re-apply, and naturally their record would be taken into account. The advantage of this over renewal is it reduces the administrative burden as inactive admins wouldn't re-apply so the number of cases to review would be smaller.

Only renew adminship once, after a year - essentially a probationary period. Alternatively renew on an elongating scale - the first renewal after 1 year, the second 2 years after that, the third 3 years later etc. This would mean early in your term as admin you get reviewed more often, but more established admins get longer terms.

Intermediate levels of adminship. I've seen this discussed on RfAs and other places before. Have "junior" and "senior" janitors with different tool access. Over time people get "promoted" throughb the levels with review at each stage. This is another variant of the "probation". I've always viewed this as introducing unneccessary complexity - most editors don;t really grasp what an admin is (nor should they) let alone the notions of gradations between admins. But maybe this is the way to go?

More as I think about this, its an important topic and I hope we can get more experienced users and admins to comment on it and generate a real consensus. Gw e rnol 17:00, 14 May 2006 (UTC)[reply]

A solution looking for a problem

The rational section has three reasons for this proposal. (1) Admins have become inactive. So what? I don't see why an inactive admin is a problem. Their account is no more at risk of ursurping than any other and their not doing any harm by not doing anything. (2) Used tools improperly. No doubt this has happened, but why isn't arbcom and jimbo sufficient to remove those sysops who do this? (3) Have lost trust. Insofar as this is different from (2) I don't understand why its relevant. Merely because people think I might do something against community consensus doesn't seem reason to remove administrator status. This illustrates what I think is the fundamental problem with this proposal. It forces people to decide what they think about a person and state it publically. This will tend to divide the community for little to no gain. --best, kevin [kzollman][talk] 17:50, 14 May 2006 (UTC)[reply]

Conversely, what if the consensus of the community is that an admin cannot be trusted with continued access admin tools (regardless of whether he has abused them to a level warranting the ArbCom's intervention)? If an admin's use of his tools is (perceived as being) more harmful than beneficial, there's no recourse, at present, for removing them; the only cases where people have been desysopped have involved egregious abuse of a "WTF was he thinking?" kind. Kirill Lok s h in 18:02, 14 May 2006 (UTC)[reply]

Kirill put it much better than I did below. Arbcom is for massive breaches of trust. There is no mechanism to deal with admins who abuse the tools in small ways all the time. There are admins doing this today and its hurting us. Gw e rnol 18:15, 14 May 2006 (UTC)[reply]

Kzollman, let me disagree, at least with point 2. I think there is a real problem here. I have encountered two admins (and one in particular) who are breaching what I consider to be the role of their office. No major incidents that would be taken to Arbcom, but a string of low-level infractions that they know they can get away with. More importantly, all editors on Wikipedia and particularly admins and b'crats should be accountable. WP does a great job at being open, but a less good job of being accountable. As the number of admins rises it becomes more important that we have open accountability practices.

So I believe that in principle we need accountability because it is healthy for any community. It also removes an argument from the nutjobs (let me clearly state I do not mean Arden) who scream cabal and clique at us and with some justification. With due respect to Jimbo, any title-for-life is a bad thing even if we are WP:NOT a democracy. I also believe that in practice we currently have admins who abuse their powers, know exactly where the line is that will take them to Arbcom and a de-sysopping and operate just under that line. This eats away at the confidence that editors have in all admins and unless it stops, it will undermine Wikipedia over time. Gw e rnol 18:05, 14 May 2006 (UTC)[reply]

No doubt there are people that live on the fringes of legitmate sysops. But, they will eventually step over the line or the sum total of their behavior will result in an arbcom action. Given how people react to admins like this in RfCs and such, I suspect they would get the necessary 50% required here by this proposal anyway. In addition, the voting process will be incredibly divisive with little benefit. Also, there are unpopular admins that have done nothing wrong. These admins will be put through a tortuous re-adminship process every year for no benefit. My inclination is this proposal will result in exactly the same people being admins as our current system, but cause a lot more conflict and departures.

This highlights the sort of awkward middle ground this proposal has to occupy. You want to make it easy enough to reconfirm that admins aren't affraid to do their job, but hard enough so that we can get rid of people that won't be removed by arbcom. The history of RfCs suggest to me that there is no such middle ground. --best, kevin [kzollman][talk] 18:16, 14 May 2006 (UTC)[reply]

On point (1) - From Chinese Wikipedia : 处于不活动状态的用户拥有管理员权限，是系统的安全隐患；实际职守的管理员比例降低，也给维基社群带来虚假的安全感。i.e. Sysops with a low edit count/inactivity is a recepie for disaster; this gives an inflated count of the real number of sysops, and gives a fake sense of security to the Wikipedian community. For the other two points, Kirill and gwernol has said my piece. - Mailer Diablo 19:04, 14 May 2006 (UTC)[reply]

I agree with kzollman and oppose this proposal. With over 900 admins, you're talking of three new renewal votes per day. Bearing in mind that bureaucrats can't remove sysop flags, only grant them, this is also going to result in a massive work overhead for the stewards too. Admins who are egregiously bad will be dealt with by ArbCom. The bar of RFA is consistently rising so I do not expect there to be an overload of admins.

I also agree with the converns that Geni raised above. Admins who do unpopular tasks (not wishing to pick on anyone, but William_M._Connolley (talk · contribs · blocks · protections · deletions · page moves · rights · RfA) and JzG (talk · contribs · blocks · protections · deletions · page moves · rights · RfA) might be possible targets) could end up getting victimized as a result of that.

However, some of Gwermol's counterproposals might possibly be worth considering. If we were to limit voting to admins, but allow all users to comment, then the goal (removing admins who abuse their admin tools, but not enough to get ArbCom whacking them over the head with a clue-by-four) would be achieved, without rendering people liable to vindictive hatred from users. Stifle (talk) 21:37, 14 May 2006 (UTC)[reply]

- Admin-only voting would only serve to further perputrate the perceived cabal-ism, which defeats the main benefit of greater accountability. Suffrage limits may be better IMO. - Mailer Diablo 02:18, 15 May 2006 (UTC)[reply]

Even the die-hard conspiracists don't claim the cabal includes 900 members. -Will Beback 05:23, 17 May 2006 (UTC)[reply]

Exceptions

I've broadened the exception from Jimbo to any official representative of the Foundation; I see no concievable way that someone like Danny could have their adminship revoked without Foundation approval. Kirill Lok s h in 19:43, 14 May 2006 (UTC)[reply]

Support... ish

As a Computer Scientist / Systems Admin with a healthy bit of paranoia under my belt, I can see a need for rights (espically adminship rights) need to expire and be renewed. But I understand the concerns that this could generate alot of unnecessary work. But some software modifications, this may be easy to accomplish (albiet posibly processor intensive).

With assignment to user groups, there could be an last renewed and expiration date (2 additional columns) added to the table, (we could define a constant such as 00/00/00 to be no expiration in case of Jimmy and the Foundation). Then 7 days before the account's rights' expire we could have a bot (cron job?) analyze tables and look for criterion that could lead to a question if the adminship should be renewed (i.e. lack of recent editing, complaints against the particular admin since last renewal). If no such indicators exist, we could automatically renew the adminship, otherwise we could automatically create a proposal for the NON-renewal of administrative rights (closing on the day administrative rights expire) and automatically notify all members who voted or commented on the last RfA or Renewal vote. If the renewal is not overturned then renew, otherwise expire and let the user be a normal user again.

Also IMHO if it takes 80% to make the user an admin, I think it should take 80% to force expiration. --Charlie^{(@CIRL | talk)} 02:23, 15 May 2006 (UTC)[reply]

70-80% to become an admin. I doubt any admin could get down to only 30% support without long since haveing been removed by arbcom.Geni 06:13, 15 May 2006 (UTC)[reply]

The need to expire was recently debunked as a dangerous myth. Kim Bruning 06:47, 15 May 2006 (UTC)[reply]

Hi Kim, Could you point me somewhere on this debunking you refer to? From a systems administration standpoint, the more users that are granted unnecessary administrative privileges the more attack vectors a cybercriminal can compromise a system. Therefore it's considered best practice to require expirations. I'd also like to make the analogy to EMT certification as well. EMT certification is a statement that a person still has the requisite knowlege and ability to perform all tasks in the line of their duty (saving lives), they are required to recertify and renew over time to make sure they're still fit for duty. Regards, --Charlie^{(@CIRL | talk)} 06:56, 15 May 2006 (UTC)[reply]

Ooh, it was recently linked either on slashdot or newsforge. I'll see what I can find. In the mean time, in summary:

On one particular very old system, expiring passwords was the correct stratagy. People just cargo-cult copied that stratagy without addditional thinking to other systems, so it's not nescesarily correct. A deeper look shows that it might be wiser to have strong passwords set once, and have people memorise them once. This both allows for better passwords, and eliminates the famous "password on post-it stuck to monitor" problem ;-) Kim Bruning 07:25, 15 May 2006 (UTC)[reply]

This might be it: http://www.smat.us/sanity/expharmful.html Kim Bruning 07:28, 15 May 2006 (UTC)[reply]

I'll agree that too frequent password changing is harmful (as you put it, post-it on the monitor problem), however I don't believe those arguments apply to rights expiration. (I'm not advocating and I don't believe this proposal is currently advocating disabling administrator accounts, just "desysopping"). A rather extreme example of rights expiration would be that US Security Clearances require renewal on a regular basis to verify that person still requires the rights they have been granted. A more moderate (and possibly applicable) example is if someone gets up and leaves a job all electronic access rights will be revoked and the account will be disabled because having the account open is a security hole and a potential attack vector to the company (at least on the resources the particular user had access to. Likewise here, Administrators have a considerable amount of power (notably, the ability to delete content and block users) and I don't believe that we can assume the existance of strong passwords (it would be an interesting poll to see how many people know what it means to be have a strong password nevermind have a strong password) nor can we assume security of account passwords (have you noticed that the login form is not SSL enabled by default?). Therefore at least if an administrator inexplicably vanishes I feel rights should expire.

As far as other renewals go, I view them as somewhat akin to annual reviews in a corporate environment. If the user is doing things harshly wrong they should have their rights partially revoked (as these rights are power). If they've only been scolded a few times, it's a good time for reflection and feedback on how to do better and improve themselves as they continue to be an asset to the project. --Charlie^{(@CIRL | talk)} 07:59, 15 May 2006 (UTC)[reply]

I see where you're coming from. I'm still an old skool "adminship is no big deal" kind of person, and think of it as a grossly misnamed "advanced editing licence". As long as any particular user shows that he shan't likely blow up the wiki, I would prefer giving that user the admin bit.

Expiration and review of 95% of our userbase (95% admins is my ideal objective), <understatement>might be somewhat impractical.</understatement> Kim Bruning 08:28, 15 May 2006 (UTC)[reply]

Impraticality is very true if this were to be undertaken completely manually. With a database and software change though, it might be able to make this more feasible. At least for vanished admins anyways...(a simple check of when their last edit in any namespace would be able to tell if they had vanished, a bureaucrat could then verify that the admin has in fact left wikipedia (or is simply taking an extended break). Part of me is thinking that almost all admins wind up with illegitimate complaints against them so if there was a way to easily tag & index legitimate complaints against the admins in the database (perhaps a way for the ArbCom members to do this?), then I think it would be easy to cut out a vast portion of the review work. Charlie^{(@CIRL | talk)} 14:34, 15 May 2006 (UTC)[reply]

Marked as rejected

Because this proposal pops up about once every (two) months or so. We've also had an experiment on Requests for Adminship which seriously Did Not Work. So erm, snowballed it. Isn't there a perpetual proposals page you can go to look, to save yourself investing your time on things like this? Kim Bruning 06:38, 15 May 2006 (UTC)[reply]

See also: Wikipedia:Admin_accountability_poll (reconfirmation was accepted by 20, rejected by 60) Polls are generally evil though. Kim Bruning 06:41, 15 May 2006 (UTC)[reply]

Wikipedia:Village pump (perennial_proposals), does not appear to have this entry yet. If you like, you could revertounsnowball me, and let people work out this page. Of course, if it fails yet again (very likely), please do add this topic to the perennial proposals. Kim Bruning 06:46, 15 May 2006 (UTC)[reply]

In that case, I shall work towards being the cabal of the cabal! Nobody can stop me! Muhahahaha! X) (Seriously, oh well, but I'd love to see ideas to improve admin accountability within the community besides just ArbCom) - Mailer Diablo 08:43, 15 May 2006 (UTC)[reply]
I'm reverting it. I saw this on the pump yesterday, and snowball isn't anything. --badlydrawnjeff(WP:MEMES?) 02:51, 16 May 2006 (UTC)[reply]
- Whatever. Do you think this proposal will succeed or not? Kim Bruning 10:13, 16 May 2006 (UTC)[reply]
  - I think it could with some serious discussion. It's apparent it's becoming more and more necessary to have more accountability. --badlydrawnjeff (WP:MEMES?) 11:02, 16 May 2006 (UTC)[reply]
    - Accountability would be fine. Unfortunately , adminship renewal does not, in fact, test for accountability. (As we have found once every other month up till now :-) ) Kim Bruning 11:09, 16 May 2006 (UTC)[reply]
      - Well, there's no way to fix that if we reject the proposal without discussion. d;-) --badlydrawnjeff (WP:MEMES?) 11:20, 16 May 2006 (UTC)[reply]
        How about instead of making the same proposal over and over, try to make a different proposal. A novel concept, I'm sure! ;-) Kim Bruning 12:11, 16 May 2006 (UTC)[reply]

Re:improving admin accountability, I brought up another idea on WP:RFA last weekend. The motivation is pretty much the same. Outside of ArbCom, options are limited when trying to make admins more accountable. Temp de-admining people that have transgressed on their admin tools might help them respect the community trust a little more in problem cases. Rx StrangeLove 04:21, 16 May 2006 (UTC)[reply]

Tail eating

We're starting to get to the point where proposals are being rejected because, um, they are rejected. Yes, this talk has happened before, and yes, I think that this idea wasn't the one I'm looking for, but we have to be careful about stopping talking/thinking/working on something just because we haven't gotten anywhere before. Beware inertia! - brenneman ^{L} 09:34, 16 May 2006 (UTC)[reply]

I'm fine with that. But the ideas behind this particular proposal were tested on RFA a couple of months ago. This proposal comes along every other month. Every other month after long debate, everyone agrees it's a bad idea. Perhaps it's wiser to spend that energy on a different idea? Kim Bruning 10:16, 16 May 2006 (UTC)[reply]

Ideas evolve. They don't pop up like anything. Which is why even if this is rejected, the discussion as a result is still useful in the quest for greater admin accountability. - Mailer Diablo 18:26, 16 May 2006 (UTC)[reply]

Saying the same thing over and over == evolution? Kim Bruning 18:35, 16 May 2006 (UTC)[reply]

Fuck accountability. Limit the damage first. Let's stop having beauty contests in which we empower editors on the basis of their popularity. Let's rotate adminship at random. You get a month or two and then you have to give it up. Power ceases to be entrenched and the faces change, so they are not always opposing efforts towards change in concert. Allow a system of deadminship. X number of legitimate (meaning grounded in policy rather than something you just don't like) complaints from editors who would be eligible for adminship and you lose it (or use some sort of approval thing if we want to be more positive). Because it's only temporary, it won't be so painful. Make X big enough and it's only the really bad who suffer from it. This proposal would allow those people who spend a lot of their time politicking or fighting with problem users to spend more on working on content. Yeah, I know. Guaranteed to be rejected. Too many people who like having the power and don't see it as "no big deal" any more. Grace Note 01:07, 17 May 2006 (UTC)[reply]

Personally, I like the concept of 1 year and then retire. <innocent look> . The nice thing is, we don't even need to modify policy to implement this. <very innocent look> Kim Bruning 11:24, 17 May 2006 (UTC)[reply]

And if they refuse to retire, will the Sandmen come after them? ;-) Kirill Lokshin 11:39, 17 May 2006 (UTC)[reply]

Ooooh, fun! How about Rick Deckard? :-) Kim Bruning 12:02, 17 May 2006 (UTC)[reply]

If it pops up every now and again, then perhaps there's some point in that proposal? Where is the poll/discussion/vote, I have troubles finding that page. Miacek and his crime-fighting dog (t) 18:16, 1 August 2010 (UTC)[reply]

Narrow it down a little?

Since the accountability (voting is evil ) poll suggests that direct community de-admining won't fly that leaves 3 other groups: admins themselves, bureaucrats and ArbCom. And since overall admin approval/respect rating is down around politicians and tax collectors that leaves bureaucrats and ArbCom. So...how about a request for renewal page. Anyone can list an admin they feel might have lost the community's trust and the bureaucrats decide if the case is valid or not and send the admin in question to RFA if they so decide. Or ArbCom can do it. The point being that a very small percentage of admins need to be worried about so a process that automatically includes everyone is a huge waste. If people think accountability is an issue then it needs to be narrowed down a bit. The accountability (voting is evil ) poll does indicate that admins are not held in the highest regard...whether that means taking some bits away is another matter. Rx StrangeLove 04:16, 17 May 2006 (UTC)[reply]

I don't think allowing witchburning would be particularly useful. There are already enough centres of conflict on this wiki and creating another one would not really be a great idea. I'm not sure either that asking the arbcom to condemn the witches would be a good solution. One of the gripes ordinary editors express is that the arbcom takes admins' sides in disputes without actually considering the merits of the case. I think there's some truth in that, and a succession of exonerated bad boys and girls, which is likely what we'd see, would not help mute that particular gripe. -- Grace Note.

Aye, that's a good point. About not needing more conflict, I meant. Well, actually the points about ArbCom being erratic and the problems with admins feeling bulletproof if they passed are good ones, too, but the main one is about the conflict. If adminship were not a big deal, then getting defrocked would be not a big deal, so perhaps some behaviors would be moderated... but that's hard cycle to kick off. I mentioned it on my RfC, but I'll bang on about it here too: Get six guys with 500 edits to say "Tag you're it" and I'll run again. Call me ever-hopeful, but perhaps some voluntary accountability (assuming I'm not alone in it) could trickle upwards?
brenneman ^{L} 12:23, 17 May 2006 (UTC)[reply]

I think that'd be the best outcome but I have my doubts about it becoming common enough to do any good. It'd be great to install this as a core value...but how? Rx StrangeLove 20:43, 17 May 2006 (UTC)[reply]

And since overall admin approval/respect rating is down around politicians and tax collectors that leaves bureaucrats and ArbCom. This is an oft used belief that seems to be floating around. Does anyone have any evidence (other than a vague "in my experience...") to back this up? Judging community attitudes is very hard, especially hard when one doesn't use methods designed to prevent observer bias. --best, kevin [kzollman][talk] 18:17, 17 May 2006 (UTC)[reply]

Well, I certainly exaggerated that point. I don't think it matters because there's no way editors at large are going to let the admin group make decisions as to who needs to have another RFA. But to answer your question the accountability poll linked above is pretty split on the admin job performance question with a fairly low level of participation.

The main point I think is to insulate the process from short term controversies that bring out the pitch forks and torches. The assumptions I'm making are that an automatic renewal process won't fly and that neither the community at large or the admin group will be trusted deciding who needs to be re-examined. To me, that leaves a small group of trusted editors deciding who needs a renewal RFA. I suppose another assumption I'm making is that enough admins volunteering to go back to RFA to make it effective isn't likely.

So I guess if you grant those assumptions the possibilities are narrowed down a bit: Not all admins need to go back to RFA and a process insulated from public opinion needs to be found that makes decisions as to who does go back. The first thing is to drive a stake into the ground that says that admins need to be more accountable and then decide what to hang on it. Rx StrangeLove 20:38, 17 May 2006 (UTC)[reply]

Term limit versus term of office

This proposed policy intends to introduce a term limit

This may seem like a quibble, but I can't find anything in this proposal about term limits. The process seems to concern itself with imposing terms of office. I've seen many comments on project pages about term limits for admins, but is anyone actually proposing this somewhere? Bovlb (talk) 18:37, 28 January 2008 (UTC)[reply]

Changed. Bovlb (talk) 04:35, 20 February 2008 (UTC)[reply]