The key is not making it easier to remove adminship.  This proposal gets us closer to the real problem, but fails to fully perceive it as does the common call to separate the functions of adminship. The real solution to the current (and relatively long-standing) problems with RfA and adminship in general is the marriage of the "technical" side of adminship with a "political" side, which is rarely acknowledged.  Successful reform will involve separating these two aspects, rather than the more common idea to separate some technical pieces from others.  The proposal below is a bit lenghty, but it's the product of years of thought, and I encourage you to read it.  If you don't have the time, well then, the take away point is that we should create a distinction between those administrators trusted to intervene in highly-controversial areas and those not so trusted. The technical bits of adminship are, indeed, no big deal.  With a large community of administrators and an alert body of stewards, the possible danger of obvious abuse of the administrator privileges is nearly zero.  As an illustration, in the heat of the recent dust-up on commons, an administrator there "went rogue" and vandalized the main page.  His edits were reverted in less than a minute: http://commons.wikimedia.org/w/index.php?title=Main_Page&action=history…did=38894141. Even in an absolute worst-case scenario of administrator abuse (for example, vandalizing the main page and then deleting a large number of pages with just less than 5,000 revisions in an attempt to lock the servers, especially abusive shenanigans in the MediaWiki namespace, or inserting malicious code into monobooks), the damage done would be reversed in under 10 minutes. Given this, it is highly improbable that any vandal/banned user would attempt to gain administrator status solely for the purpose of carrying out some such abuse.  The danger comes from a compromised account or a higly disaffected administrator, and neither of these possibilities can be headed off by any level of standards at RfA, however high. Why, then, has adminship become a big deal?  Because in addition to the purely technical functions of adminship, administrators also have a political function.  Administrators are often compared to janitors, but the metaphor is highly flawed.  Janitors empty the wastebins, but they don't decide what should go in them.  Many of the functions of adminship do not carry a significant political component: blocking obvious vandals, most instances of speedy deletion, fixing cut and paste moves, deleting old userpages, straightforward AfD closures, etc. are simple instances where a trusted user is needed to perform a technical function. On the other hand, there are cases were administrator functions become highly charged and political - in closing controversial AfDs, blocking in many 3RR situations, and above all, in cases where some sort of intervention is necessary against well-established users who have engaged in some sort of unacceptable conduct.  In these cases, the role of the administrator is fraught and ambiguous.  He is faced with highly political choices about how to judge consensus, what course of action to take, etc.  It is customary for relatively new and inexperienced administrators to stay out of these situations and leave the decision up to an administrator who has more experience and, for that matter, for political weight within the Wikipedia system. The problem, though, is that there is no formal guidance of any kind as to who should actually make such decisions.  From a policy perspective, an administrator sysopped last week has the same standing as someone with years of service.  More importantly, a long-standing administrator with a reputation for more questionable judgment has exactly the same standing as a long-standing administrator with a reputation for impeccable judgement. There is no drawn by the community, except in the various most informal way, to separate administrators who should intervene in highly controversial situations from those who should not. It is intervention in the highly controversial cases that causes problems and allegations of abuse.  Our concern is, or at least should be, primarily in who is making highly controversial administrator judgements and on what basis, not who is carrying out F5 speedy deletions or blocking obvious vandals.  Concern over these highly controversial judgements, because there is no line separating those administrators who engage in them from those who do not, is what has driven steadily escalating standards at RfA.  We are less concerned that a newly-appointed admin will prematurely block a vandal without any warnings tomorrow, than that he will, in 12 months, block a well-established user for the wrong reasons after a heated debate at ANI. In other words, the problem is that RfA is being asked to make a judgment that should not be made at RfA. What we need, then, is not a way to desysop more easily, but rather a way to delineate highly-charged and controversial administrator actions, and the administrators qualified to perform them, from uncontroversial administrator actions, and the administrators qualified to perform them.  I will not presume to provide a full criteria for what separates controversial from uncontroversial administrator actions, but I would suggest something along the lines of the following.  Controversial: Arbitration enforcement actions, blocks of established users for any reason other than suspicion of account compromise, close of AfDs where the consensus is not clear (this of course becomes itself a murky distinction, but could be well enough set apart), reversal of the actions of another administrator except when those actions are plainly abusive.  Non-controversial: All others. As for deciding which administrators are qualified to make decisions in the most controversial areas, I would suggest that we already have a group of people, the bureaucrats, in whose judgement the community has expressed particularly high confidence.  I would propose that the bureaucrats become the group who are expected to undertake the controversial administrator actions; this would almost certainly entail some expansion of the current bureaucrat pool, but personally I like the idea of tying the controversial administrator actions to the ability to promote administrators - it underlines their seriousness, and at present, the bureaucrats do not have many functions.  If, however ,the community is unwilling to combine the two groups, another group, say "sub-bureaucrats" could be created, but I must emphasize the importance of a bright-line distinction between those administrators trusted to perform highly controversial tasks and those not trusted to do so.  Obviously, the ordinary administrators would still have the technical ability to intervene in the highly controversial areas, but doing so would obviously entail serious consequences or desysopping. This brings up a final point: the issue of administrators with insufficient knowledge to appropriately follow policy on, for example, speedy deletion. I firmly believe that if we separate the political and non-political aspects of adminship, this becomes less of an issue.  While an administrator taking the wrong course in a controversial area is akin to a janitor, who is empowered to decide what to throw out, deciding to throw away your important papers because he doubts their importance, the mistakes of lack of policy knowledge and inexperience are more like a janitor who, because he doesnt' know any better, throws away the recycling and attempts to recycle the rubbish.  The second category of mistake is more easily rectified.  The old idea, of some sort of mentoring for new administrators, does nothing about the political aspects of adminship (making controversial decisions) which is why it has failed in the past, but it is a perfect solution to the problem of inexperience/ignorance.  New administrators who do not have a full grasp of the speedy deletion policy, or the blocking policy for vandals, or the criteria for granting autoreviewer status would be encouraged, perhaps through a formal process, to get up to speed on those areas by a more experienced mentor.  If we carry through this proposal, there is every reason to believe that the crowd at RfA would be much more willing to promote more candidates and the process would become much less grueling. Our shortage of people to perform technical tasks could be easily reduced, if not eliminated. This proposal is not process creep or the introduction of unneeded bureaucracy.  It is also not an answer in search of a problem.  There is clear acknowledgement that we have a problem, and this solution is a minimalist one.  As I have proposed it, it simply takes advantage of an existing process (RfB) and group of users (bureaucrats) and would require only minimal amendments to policy, setting aside those areas of administrator conduct that are highly controversial and requiring that only bureaucrats act in those areas. _______________________________________________ WikiEN-l mailing list WikiEN-l(a)lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l

I think this is something to untangle. We need to get to the bottom of the community's fears about "overpowerful" admins, by talking through and delineating what a single admin can expect to face in awkward situations.

I've never been in favour of restricting admin discretion, which is really what is being proposed.

We can't anticipate the challenges the site will face (even though it may appear that there is little innovation from vandals and trolls).

I do think admins can be held to account for their use of discretion. Right now it seems that a piece of the puzzle is missing: admins don't know clearly how they stand in relation to the actions of other admins.