At 01:58 PM 5/30/2010, Thomas Dalton wrote:
>On 30 May 2010 11:43, David Gerard <dgerard(a)gmail.com> wrote:
> > Indeed. The first - and, I would have thought, jawdroppingly obvious -
> > result would be that no-one at all would go near such work in any
> > circumstances.
>
>Exactly. The big problem with community desysoppings is that any admin
>doing their job properly will have enemies. The longer you do the job,
>the more enemies you will have. Whenever you block someone, you annoy
>the blockee. Whenever you delete an article, you annoy the creator.
>Whenever you protect an article, you annoy the person whose version
>you didn't protect on. If you let those people be in charge of the
>desysopping process, we won't have any good admins left doing even
>slightly controversial work (which, as I've explained, is pretty much
>all admin work).
These are the arguments that have maintained the dysfunction. But:
(1) most legitimate admin work is not controversial to any degree
that would affect an admin's status in the active community, which is
what counts. Blocking an IP vandal isn't going to harm that, and it
will only help it. If the IP vandal then registers an account and
goes after the admin, sure. But, then, as to proposals that those who
supported an RfA might retract that, or cause adminiship to be
suspended pending examination, are concerned, this would be useless.
Legitimate administration is indeed like janitorial work. Can we
imagine a good janitor getting into an argument with other employees
of a school or office as to what should be thrown away? Adminship was
supposed to be "no big deal." When an administrator is asserting
personal power over an editor, something has gone awry. Police have
no power to punish, they may arrest on probable cause, but they then
step aside and let the community make decisions on sanctions or
release. A police officer who has become personally involved and
insists on pursuing an individual might well be removed or ordered to
work in other areas.
"Whenever you delete an article, you annoy the creator." Well, it
might seem that way. But admins aren't supposed to be deleting
articles in the presence of the creator's objection, unless there is
a critical issue, and, by the rules of adminstrative recusal, they
should only do this once, personally, absent true fire-alarm
emergency. It better be good! For anything further, they'd go to the
community and not use tools to gain an advantage. And I've seen
admins violate this, causing a lot of unnecessary disruption because,
indeed, the editor then gets seriously pissed off. That's as to
speedy deletion. As to regular deletion, an admin is assessing
arguments and consensus at an AfD, and, if doing this well, doesn't
delete unless there is consensus for it, or, alternatively, the
arguments are clear and evidenced. And if the creator objects, the
admin politely considers the objection, and, if the admin can't
reverse, suggests DRV and is done. Seriously done. Probably not a
good idea to even argue for deletion at the review, the admin's
reasons should have been given with the original closure. Being
reversed should be no shame.
(2) good recusal policy requires an admin to stand aside and not
pursue an individual editor. An example of how this could work was
what happened when Iridescent blocked me in 2008. It was indef, but
she wrote, "indef as in indefinite, not as in infinite," or something
like that. And then she made no attempts at all to *keep* me blocked.
She presented her reason, and that was that. It was then between me
and the community, not me and her. As a result, I had no sense of
serious opposition to or from her, and no enmity. I still think she
made a mistake, but administrators are volunteers and will make
mistakes. Am I unusual? Maybe. But if an editor is, say, blocked for
a day by an administrator who then leaves unblock template
instructions and even wishes the editor well, and does it all
politely and correctly, it's going to be very visible if this editor
then embarks on a crusade against the admin -- unless the admin truly
was involved and shouldn't have touched the block button. Sure, it
happens. And it's very visible if anyone looks! Indeed, this editor
is likely to stay blocked or to be seen as seriously biased against
the administrator and possibly as genuinely dangerous to the project.
"I was blocked by a horrible monster" is very much not a way to get
unblocked, it rarely works.
(3) "community desysopping," per se, is a really Bad Idea. It should
be and must be much easier, and community discussions tend to be very
much a popularity contest, and waste huge amounts of editor labor.
Rather, some kind of administrative recall, as an easy process that
could result in *suspension* of administrative privileges, and even
without some presumption of actual misbehavior, merely in undoing,
temporarily, what was done with the RfA, makes much more sense.
Involving those who approved the adminship in the first place, and
who supported it, seems like a possibility that could be quite
efficient and quite clearly fair. I'm not detailing a process here,
but it would presumably be appealable. Suppose you approve my
adminship and I then find it necessary to block you. Under these
conditions, I'm probably not biased! But suppose it pisses you off.
If you can convince some number of other supporters to ask for
suspension, that might be automatic. I.e., there would be, perhaps, a
consent and request, in advance, part of the original RfA, that a
bureaucrat remove privileges under stated conditions, as verified by
the bureaucrat. This removal could then be undone through some
process, which might simply be a new RfA, but without the presumption
of a supermajority being needed. Indeed, I'd think a majority for
unsuspending should be enough (really, it would be the judgment of a
neutral bureaucrat, because of the possibility of pile-on from a
faction). And beyond that, if something was awry (such that pile-on
of a faction offended that the administrator was enforcing overall
policy), the matter could go to ArbComm on request, which might look
at the behavior of all parties. I've opined that ArbComm should
effectively suspend admin privileges for any admin if they accept the
case on a showing of probable cause of abuse; it might well do this
by issuing an injunction against use of tools in some area, not by
actual removal. And, as well, the "removal" I'm suggesting by a
bureaucrat might simply start with the admin abstaining from tool use
as instructed by the "recalling" editors, according to an original
promise, and it would only become an actual request to a bureaucrat
and then actual unsetting of the bit if the promise was violated.
Note that any administrator should probably recuse from use of tools
in an area when reasonably requested by a few editors, at least
pending discussion, I get into this more below. If I recuse on such
request, say on request by some process involving those who granted
me admnisthip in the first place, I may still be able to serve the
project in almost all the ways I'd be using admin tools anyway. It's
only when an admin uses tools, consistently in some area, and having
become involved in some way, personally, that there is a problem.
Often an abusive admin in one area still does good work in another,
if they can stay away from controversial use.
Would people approve of an admin just to gain an ability to torpedo
the admin later? I doubt it. It would be too easy and too visible to
shoot down, and the situation of an admin blocking someone who had
supported the admin gaining tool access would tend to look like "It
must have been necessary!" rather than the reverse. Frivolous
interference, or interference that has the effect of harming the
project, with an administrator especially, should be a sanctionable
offense. On the other hand, making a complaint that is considered
reasonable when reviewed should not be sanctioned, and that it
sometimes is, in effect, is chilling. Personally, I was appalled when
I filed an RfAr over administrative abuse, which was effectively
confirmed by ArbComm, it really was abuse, and the sysop lost his
bit, but ArbComm allowed the case to be massively broadened into a
"Whatever Abd Ever Did That Could Look Bad" mess. If I'd done so
much, there should have been an RfC on my behavior, and then a case
if conflict remained, a separate case, where I'm the topic, not
complicated by administrative abuse.
Indeed, that sysop mentioned had been causing problems with his tool
use and general editorial behavior for years, and I saw
administrators back off from confronting it because it was so
"expensive" because of the faction (a small but active minority)
backing him. It's still going on, but it now looks more like an
end-game, because some highly privileged and connected administrators
finally figured it out and how factional support was allowing it to continue.
Calling better process "getting rid of admins' is not a fair
statement of what decent proposals would look like. The structure
should make it easy to *restrain* administrative abuse, which would
start with much less drastic process than removal, it would start
with normal dispute resolution, at least at a low level. If a dispute
over tool usage continued, the actual usage might be examined, as
usual. In addition, my view, continuing to use tools where a user,
with anything even remotely reasoanble, objects, is not a good
practice, it should only be done in emergencies. Normally, with
respect to a registered user, an admin should recuse, practically, at
the drop of a hat. When I've suggested this, it's been claimed that
this would result in vast wikilawyering, but that objection is
clearly preposterous. If I block you and you scream that I'm biased
and should recuse, I'd respond. "Of course. I regret that this has
distressed you. I'm recusing. Bye." Actually, I'd do even better than
that, I'd provide a biolerplate set of instructions on how to appeal
an unblock. Naturally, when I blocked, I should already have provided
the block reason and the important evidence, or, if I hadn't, I'd
provide that. And then drop the whole matter. Unless I thought the
project would benefit from the unblock, I wouldn't unblock. But I'd
step aside from objecting to an unblock by any other administrator.
And then, in the future, if I saw a threat to the project from this
editor, I'd go to a noticeboard like anyone else, but I'd disclose
the prior request for recusal and acknowledge the claim of bias.
There are administrators who detest recusal policy, and they've been
very explicit about it, and if ArbComm were awake, it would order
their admin privileges suspended until it assured them that they "got
it." Instead, they practically have to dismember an unfortunate
editor right in front of ArbComm for it to be noticed. As long as
they avoid that, they're cool! They object that an editor could then
avoid being blocked by requesting recusal from admin after admin.
Given that recusal doesn't unblock, and even if they get an unblock,
are they going to claim that the unblocking admin was biased? That's
going to look really, really bad. I think within three such requests
or so, they would almost certainly be looking at an indef block with
no admin willing to unblock. I.e., a defacto ban, only with minimal
fuss, and they'd have only ArbComm to appeal to, and ArbComm is now
denying even some reasonable appeals, as far as I've seen. They don't
want the hassle.
Adminship should be "no big deal," as was claimed at the beginning.
And thus it being suspended, in part or even in toto, should be "no
big deal." Rather, some administrators very much think it's a big
deal, clearly. And this "big deal" concept, enforced by them when
they vote in RfAs, keeps people from volunteering and being accepted,
far more than some idea that allegations misbehavior might result in
a relatively harmless suspension. Someone who could not accept that
probably has the wrong idea in the first place about adminship and
thinks it is of some personal advantage. It isn't. It's an
opportunity to do some boring, relatively unrewarding work. But some
think of it as an opportunity to exert more power than regular
editors. What is wrong with this picture?