Wikipedia:When unbundling admin power works
Over time, as the system of Wikipedia administrators has become more and more dysfunctional, various abilities that were one admin-only have been unbundled from the admin toolset and provided to a broader class of editors in good standing.
In organizational management in the real world, there are three kinds of administrative work:
- #1. specialist – fiddly work others are not competent to perform but upon which their ability to work depends, directly or indirectly (examples: systems & network administrators, attorneys, flight-line crew);
- #2. clerical – rote functionary and support work that needs to be out of the way of those doing the central work of the organization, or inefficiency results (examples: secretaries, bookkeepers, quartermasters);
- #3. managerial – top-down decision-making to:
- #3A. direct labor to the organizational goals (examples: project leaders, HR, division managers, organizational officers), or
- #3B. manage human foibles (examples: labor dispute adjudicators, and again division managers and organizational officers).
The Wikipedia editorial community itself already forks #3A and #3B sharply, with "project" (content-work) leadership being a totally informal meritocracy of who is getting the work done sanely, i.e. the most active and intelligent editors, often the effective leaders of wikiprojects, and it operates rather like leadership in hunter-gatherer societies, by suasion and assent based on results, not by post-agricultural-style hierarchy and vested authority. The "human resources department" function of #3A is also distributed, with the whole community deciding who to "hire" as an admin or 'crat, and who to "fire" as an editor (via WP:ANI (though WP:ARBCOM and WP:AE have taken and formalized some that distributed community authority).
For #3B – "bossing people around" – Wikipedia is very hierarchical and formal, with WP:OFFICE lawyers and other WMF officers, then stewards, 'crats, and admins, plus "bodies" like ArbCom and MedCom, plus a division of "rules" into policies, guidelines, essays, and processes and procedures. For better or worse, when it comes to being able to tell people what to do we're very, very processy. Unbundling in this area is not likely to be successful any time soon, only in areas #1 and #2 (see template-editor, file-mover, page-mover, account-creator, etc.) Where the unbundlings we already have do cross slightly into #3B's "laying down the law" territory at all, there are strict rules regarding their application (and they tend to become tighter over time); this is even true of non-admin closure, a largely informal variety of unbundling in this area without permission bits being involved.
Unbundling works well for #1 and #2 because there's no intrinsic reason to not approach these from the same egalitarian "are you going to get it done, correctly?" viewpoint as we treat topical, content-development de facto leadership. The impediments are entirely incidental (e.g., cleaning up the CfD backlog requires some presently admin-only powers, simply because they have not yet been delegated like template-editor and page-mover have).
The principal impediment to further unbundling is the failure to recognize that behavioral constraints, rather than technical limitations, are what make the power to move or delete something not particularly dangerous. Because we started with "admin or not admin" with no middle ground, the community suffers under the illusion that the ability to delete or move something at all will always be used to delete or move anything, out of process and willy-nilly, simply because it's technically possible. This actually makes no sense at all, and WP's present operation proves it doesn't. File-movers, template-editors, and now page-movers do not go around wreaking havoc, but stay within what they're allowed to do (on pain of losing the bit and possibly other sanctions). More broadly and more obviously, WP's entire operation is one in which the power to edit could be used to blank or make senseless or misleading any random article on the system, yet this does not actually happen systemically at all (individual abuses are usually detected quickly and dealt with, and done by newly arrived and usually anonymous outsiders, or by banned bad apples returning as sockpuppets). There is no actual reason to believe that unbundled admin abilities, limited by rules regarding their use, would be used inappropriately more than some tiny fraction of 1% of the time. Our community values and enforcement processes will actually ensure that they aren't.
Every single unbundling to date has been opposed by doomsayers predicting disruptive chaos in form of editors abusing the user-rights, and collapse of adminship as suddenly irrelevant. It's just not happening. All of our unbundlings have proven quite effective at reducing backlogs of tedious administrative work, and have raised no consequential problems of any kind.
Making some particular technical- or rote-work function no longer the pointlessly exclusive realm of admins does not magically make admins obsolete, since their primary present function (whether we like to admit it or not) is dispute mediation and disruption lock-down, the only thing out of administrative work types #1, #2, #3A, and #3B that may intrinsically require a higher level of maturity and trust, and thus naturally lend itself to an admin class of type #3B.
The most frequent reason for doomed unbundling proposals is overreach, into area #3B. The community is never going to unbundle those functions, because they are the ones – the only ones – that actually sometimes require better-than-average, and not just editorial, judgment. WMF has also made it clear that they're never going to get rid of the administrative class, for legal reasons. Part of being a US-based nonprofit is having some kind of election, and a meaningful process for determining competence for the authority to take actions that affect people.