I'm CC-ing gentoo-project on this, because some Council members and
perhaps others seemed to indicate a preference for moving the discussion
there.  This response turns out to be much longer than I had intended
because I always try to follow my reasoning to see where it goes,
whether I like the conclusions or not.

On Thu, 2008-07-10 at 21:54 -0700, Donnie Berkholz wrote:
> On 12:26 Thu 10 Jul     , Ferris McCormick wrote:
> > On Wed, 2008-07-09 at 22:49 -0700, Donnie Berkholz wrote:
> > 2.  But for both devrel and userrel, the Code of Conduct loses almost
> > all its impact unless response is immediate --- CoC's intent, I think,
> > is to help keep the mailing lists and #gentoo-dev channel on track
> > pretty much in real time.  I know this was the original idea behind it,
> > and this was one reason we felt we needed people outside devrel to help
> > enforce it (devrel is not set up for immediate responses);
> 
> The concepts of poisonous people and repeat offenders are explicitly 
> mentioned numerous times in the 20070308 council meeting. Here are some 
> examples:
> 
> <wolf31o2|mobile> kloeri: banning people from the lists, not 
> necessarily... but reducing the requirements on devrel to suspend 
> "repeat offenders" might just make them think about their actions before 
> doing them, and that could decrease the flames a bit
> 
> <kloeri> there's some devs that are persistently poisoning the project 
> that I want to deal with but that's not just related to mailinglists
> 
> <wolf31o2|mobile> christel: agreed... I think we need to be a bit more 
> strict on our developers... after all, in the flames involving users, 
> developers are just as much at fault as the users... perhaps if the devs 
> didn't respond in kind, the flames would subside much quicker, etc
> 
> <kloeri> I don't want to ban anybody but I do want to be much harder on 
> devs poisoning things consistently and I'm going to file at least one 
> devrel bug in that regard
> 
> <kloeri> I don't think we can force people to follow netiquette in 
> general but we can do more to smack devs up when they're constantly 
> being a pain in the ass
> 
> 
> On the topic of userrel's power to ban people from lists, which is a 
> long-term action:
> 
> <robbat2> on the side of devrel not having 'teeth' - kloeri mentioned 
> before that infra previously wasn't very responsive to requests to do 
> things (he cited a userrel request to remove user from the ML)
> 
> <christel> i have a question, if we are to start enforcing etiquette 
> policy, i think we may want to ensure we have one which also cover users
> 
> > 4.  That is, we (devrel, userrel, averyone else perhaps) should use Code
> > of Conduct to stop elaborate flame wars before they can burn out of
> > control.  Whether a flame war ever merits a bug will vary from situation
> > to situation, but generally if we have a flame war and wish to impose
> > some sort of sanctions because of it, we really need to be hitting
> > several people or none with warnings or brief "vacations."
> 
> I agree that we should attempt to take short-term actions in response to 
> immediate offenses.
> 
> > 5.  I am not sure where the current Code of Conduct document is, but
> > I'll volunteer to help update it to bring it into line with how we wish
> > to use it and to help clarify who has what authority under it, and that
> > sort of thing.  I have come to support it, and I'd like to help make it
> > more effectively used in the rather narrow context for which it was
> > designed before we consider extending its reach.
> 
> On the topic of trying to write down every possible way to go about 
> this, I also agree with them:
> 
> <g2boojum> christel: I actually think you want it to be more vague than 
> specific.  "Don't be a jerk."  Please don't define "jerk", or you get a 
> five-page treatise on why the bahavior doesn't really fit the 
> definition.
> 
> <seemant> we really need to be careful in adopting document upon 
> document upon document
> 

I note two things.  (1) As I read this, no one here is arguing for
anything like a permanent ban; (2) The main thrust in this appears to
address *poisonous developers* except for christel who advocates
including non-developer users as well.  And I don't see anything
suggesting that wolf31o2/kloeri/christel/seemant are discussing
permanent action, although I don't have the complete context.

So, I don't think I have any argument with
wolf31o2/kloeri/christel/seemant here, but I think what you cited
*supports* my view.  Let me quote kloeri again, because he seems to be
the strictest among them:

<kloeri>:   there's some devs that are persistently poisoning the
 project that I want to deal with but that's not just related to
mailinglists

<kloeri>: I don't want to ban anybody but I do want to be much harder
 on devs poisoning things consistently and I'm going to file at least
 one devrel bug in that regard
 ....

In case you are misunderstanding me (well, from you other response I
know you are;  I'll try to address that once again below):  I don't mind
being more vague than specific (I don't ask for a definition of "Don't
be a jerk", really).  My concern goes more to who determines "jerk-ness"
and what we do about it.  And it bothers me a lot that a small number of
people believe themselves qualified to make that decision in secret.
Code of Conduct seems to require that the people applying it are
actively working with the "jerks" involved.  Even kloeri said he was
opening a bug on some developer for all to see.

As for seemant's "document on document" --- I agree.  But I do insist
that our policy documents reflect what we can do and why.  Nothing in
our current documentation that I can see indicates that we should have a
group of people rooting through our archives in order to put together a
case for imposing a permanent ban on someone, and if you are really
suggesting any such thing, *something* needs to be updated in order to
put the community on notice.

There's a difference between "document on document" and a real change in
policy/procedure without and document changes at all.

Please resist the urge to dismiss my description as ad hominum.  When I
say "root through the archives" I realize that that is not a neutral
description of what people have in mind.  But I think it is completely
accurate.

====================================================================
Now, I'm going to change the topic slightly and explain what I think the
context of Jorge's proposals is.  I ask him to set me straight if I'm
getting it wrong.

As I understand it, these proposals fit into the context of the Code of
Conduct, and no matter what you say, I am certain that the Code of
Conduct was put in place to address problems as they occur in order cut
off and prevent brush fires.  In this context, his permanent ban
proposals would be the final sanction after quite a long run of working
with someone through the Code of Conduct itself.  And I have never seen
anything suggesting nor anyone proposing that the Code of Conduct has a
long reach into the past to apply to someone now.  Code of Conduct
addresses current conduct; it does not address past conduct except in
the context of what is going on now.  I ask Roy or Jorge please to
correct me on this.

So, if we were to add Jorge's proposals to the Code of Conduct, they
would fall into that context, and would never come into play at all
unless triggered by some sequence of Code of Conduct violations starting
at the time they were adopted.  Personally, I would probably not support
that, but you might be able to talk me around.

Now, as I have said, I think providing for moderating the -dev mailing
list fits much more neatly into the Code of Conduct, and I had thought
we would have this in place by now (although the push for that seems to
have died --- it's sort of funny that right now I'm the one pushing it).
As I see it, this would give us the option of shunting all posts from
someone to a group of moderators who would either pass the posts or
bounce them with an explanation.  We already do this on at least one of
our mailing lists (gentoo-dev-announce?) so it is nothing new and it
works well in the Code of Conduct context.  It also solves the problem
of future posts from "poisonous people."

Donnie suggests elsewhere that moderation is not the answer because the
number of "poisonous people" is small and the group tends to be the
same, but I don't understand the point.  All that says is that the list
of people being moderated would be pretty static.  To that I answer (1)
So what?  Does it matter that the moderators have a static list rather
than a dynamic one? (2) That's invalid anyway because we've never tried
it.  We don't know how anyone would react to constant requests to modify
a post.  So to reject a solution to a perceived problem because "we know
it won't work" even though it fits nicely into the Code of Conduct and
instead put in place a policy of pre-emptively banning so-called trouble
makers strikes me as ill-conceived and premature.

Oh, yes.  I'm going to stand by my statement that imposing permanent
bans in secret without involving the parties involved is cowardly.  If
we are willing to take such extreme measures against people, we should
be willing to face them to discuss the problems and to negotiate less
extreme alternatives.  It's easy to write an email to someone saying
"You're banned from all things Gentoo";  it's somewhat harder to talk to
that person about it.

I'm also going to stand by my statements that the intent behind the Code
of Conduct has always been to provide a way to react quickly to problems
as they unfold.  This is not explicit in the Code of Conduct, but I
think it is a fair inference from the discussions leading up to it.  I
ask Roy or Jorge to correct me if I am wrong.  Or we could just ask
Christel; she wrote it and she knows what she had in mind.  (Actually, I
guess I'll ask Christel if no one else does.)

Regards,
Ferris
-- 
Ferris McCormick (P44646, MI) <fmccor@g.o>
Developer, Gentoo Linux (Devrel, Sparc, Userrel, Trustees)