Currently they have more abilities than bureaucrats.
They can grant or
revoke any level of access. Currently bureaucrats can only grant
bureaucrat or admin status.
I think the best way to go about this is to simply have two types of
bureaucrats, local bureaucrats who can add and remove all user privileges
and have access to other high level maintenance functions, and interwiki
(super-)bureaucrats who can do this on any wiki. I think this additional
"steward" level is needlessly confusing.
Additional features may be implemented depending on
Ultimately I would like to see stewards capable of configuring the power
structure on each wiki individually
We are entering very dangerous territory here. Your original proposal was
to decentralize power, and that is all well and good. But this idea, to
configure power structures on a local level, has a lot of room for long-
Let's say I am a highly influential and generally good-natured user who is
very active on, say, the Nepali Wikipedia. I'm also known and trusted on
the English Wikipedia, so I am quickly given the "keys" to this new power
structure configuration module for ne.wikipedia.org
Now let's say I happen to believe that it's a good idea to let sysops ban
any user for repeated POV edits, given prior warnings. I'm sure there are
many otherwise quite reasonable people who would believe that this is a
good idea. Because of my influence and my level of control, it would be
very easy for me to implement this power structure on the Nepali
As the Nepali Wikipedia grows, this principle becomes quite established.
"POV users" are routinely banned arbitrarily. Because of my influence, I
have managed to become the "last instance" of appeals -- there is no
voting before banning, but I, the "steward", have obtained a position that
effectively allows me to veto or approve any ban. Because I, and nobody
else, can desysop users, I can easily enforce this principle.
At some point in 2008, some very eloquent English-speaking user of the
Nepali Wikipedia complains about this unfair power structure, appeals to
Jimbo, the mailing lists etc. However, he is quickly shouted down by
people who fear for the autonomy of "their" Wikimedia wikis. They will
defend almost any rule on the Nepali Wikipedia as being "the business of
the Nepalese" and argue that Jimbo et al. should stay out of it, because
they are afraid that if they do not do so, their own idiosyncratic
policies will come under scrutiny.
"Individual projects should have this level of autonomy!" they will say.
And: "These rules have been in place for years, there's no reason to
change them now." There is a high level of tension between the English and
the non-English spheres, so Jimbo et al. will stay out of it to keep the
peace. This will set a precedent and other complainants will be similarly
Now everyone in this scenario acted with good intentions. However, one
single user has effectively managed to obtain control over what is and
isn't NPOV in a certain language Wikipedia. There might be more stewards
later on, but that merely means a change from a monarchy into an
oligarchy. Their decisions may be wrong or right, that doesn't matter, as
I think we agree this is not a scenario we want to happen.
I don't like the idea of implementing power structuers on the local level.
I see no need for it and a lot of ways for things to go wrong. Yes,
there's a difference between a 20-person wiki and a 20000-person wiki. But
these differences can be accounted for by developing a multi-stage power
structure that can be generalized for all wikis, e.g.:
< 10 users : no sysops or bureaucrats; use Meta to ask super
-bureaucrats to set permissions accordingly
< 100 users : at least one bureaucrat, local "Requests for adminship"
page, sysops can ban users if there is general consensus
to do so, bureaucrat can remove sysop status if there
< 1000 users : "Requests for bureaucrat status" page, appeals process,
1000+ users : quickpoll-style system for dealing with emergency
I find it highly improbable that the cultural differences are so great
that we can not deal with the problems - vandalism, sysop status abuse
etc. - on a general, language-independent level. If you approach this
problem from a purely rational perspective I believe you will have to
admit that this is quite clearly so.
In fact, I believe that many non-English wikipedians will embrace your
proposal not for rational reasons, but purely because they want to grab
every little bit of autonomy for "their" wiki that they can get.
On each reasonably large wiki there are at least a couple of users who are
subconsciously motivated by a real drive for power. They will rationalize
this by saying that "We don't want to be controlled by the English
Wikipedia", "We have different problems that need different solutions",
but in reality they are already planning on what to do with these new
privileges, and not all outcomes will be desirable. I have already
observed that on some wikis, individual users have worked hard to
establish a status of authority for themselves, and a tool like the one
you are describing is exactly what they need to solidify that status.
It is also an excellent tool for astroturfers, who can dedicate a lot of
time to becoming influential and powerful "stewards" on up-and-coming
Keep in mind that Wikimedia is just a couple of years old, but it will be
here for a long time to come. Unless we screw up in *major* ways, and
someone creates a successful fork, the *minor* mistakes we make today will
have *huge* consequences many years in the future that are very hard to
foresee now, and very hard to repair later. The different languages
effectively act as barriers, and it is very difficult to spot dangerous
developments through these barriers. I am strongly opposed to effectively
granting absolute autonomy to a small set of individuals in defining power
structures that will likely be in place for years, if not decades.