ietf
[Top] [All Lists]

Re: Proposal for a revised procedure-making process for the IETF

2001-10-13 20:20:02
    Date:        Sat, 13 Oct 2001 11:20:08 -0400 (EDT)
    From:        James M Galvin <galvin(_at_)acm(_dot_)org>
    Message-ID:  
<Pine(_dot_)BSF(_dot_)4(_dot_)21(_dot_)0110131109430(_dot_)18244-100000(_at_)two(_dot_)elistx(_dot_)com>

  | Robert, it has never worked this way while I've been Chair.  Neither
  | POISSON nor any working group gets to do what it wants to do when it
  | wants to do it.  All topics have always needed IESG approval.

From where exactly do you draw this conclusion.   I have just reviewed
rfc2418 (bcp25), and I can find nothing that says that.  It certainly isn't
in 2026.

The IESG needs to agree to the charter, that's clear, and the AD needs to
agree to any milestone updates, etc - but for work that falls within the
charter, the WG decides just what it wants to do.

If you, and other WG chairs, are running to the ADs every time someone
raises  new topic, then no wonder the IESG are way over worked - it should
be for the WG chair to decide whether something is within the charter or
not - if it is, then fine, it gets discussed.  If not, the WG may decide to
disagree with the chair - and if required, ask the AD to replace the chair
(as can happen any time the WG doesn't like what the chair is doing).  If it
is agreed by the WG that the issue is currently outside its charter, then
the WG can rewrite its charter, and submit that to he IESG for approval.
That's where the IESG should get involved.

  | It's just that with POISSON (at least while I've been Chair) the approval
  | was just handled by the Chairs on behalf of the working group.

The poisson charter doesn't seem to have been revised in a long time, so
it doesn't seem to me that anything needed to have received IESG/AD approval.
(Aside: revising the charter might be a reasonable thing to do - my impression
has been that that's been deliberately avoided, precisely because of what I
mentioned in my last message - a desire to have the WG not start on anything 
new).

Of course, it is just fine, and completely proper, for the WG chair to
keep the AD informed as to what is happening in the WG (though the AD is
likely to be on the mailing list anyway...).  But that's a different issue.

So, what really constrains what the WG can do is what the charter says it
can do, and for poisson, that includes (as well as stuff that has been
done, but as the charter hasn't been revised, old stuff is still there) ...

  - Discuss an update of the Nominations Committee document RFC-2282, if
    and when the current NomCom has suggestions for changes.

  Furthermore, the POISSON WG will review documents that are related to
  the IETF standards process (but that will not be produced by the POISSON
  WG itself) when available. Such documents may include: [list omitted]

  Last but not least, Poisson will serve as a generic platform where the
  IAB and IESG can discuss policy questions if there is the need for
  consensus polling.

Anything that falls in any of those categories is valid for discussion
without anyone asking the IESG if it is to be allowed.   Eg: the ietf
list charter, and the code of conduct docs pretty clearly fit under the
2nd of those - if you asked for IESG approval before discussing them, then
there was no need.   Nor should there be, we need a way for people to
propose changes to the process in a way that they can be discussed, and
if the community agrees, proceeded with, without discussion being able to be
stifled by the IESG (or anyone else).   That's important to us remaining an
open body.

  | When there's a working group task it will have its own list.

If that's because a new WG has been formed, that's fine.  And if we wanted
to form a new WG to consider (for example) exactly (only) a revised 2026,
then that would be (or could be) exactly what should happen.

But for one WG, there has to be ONE mailing list that is the WG list.
Obviously there can be other lists established from time to time to move
out discussions, but there has to be one place which is where the whole
WG make their collective decisions.   Ie: if another list were to be created,
and reach some kind of consensus as to a doc to publish (or something) that
can't be considered as a WG position until it has been verified on the one
WG mail list.   That's the one that section 2.2 of 2418 says needs to exist.

"a mailing list", "the mailing list" (etc) (no mention there anywhere of
there being multiple mailing lists - and for very good reason).

  | Otherwise, why does it matter whether discussion of
  | topics come and go on the ietf list or the poised list?

Because the ietf list is the best list in the IETF to be unsubscribed from.
It very rarely has any content that anyone cares about, and very often has
lots of content which is of concern to almost no-one.   On the other hand
the poised list targets exactly the people who care about IETF process issues,
and doesn't bother all those who don't care very much, and mostly contains
only discussions relevant to IETF process issues - it isn't a painful list
to remain subscribed to at all.

kre

ps: you will note that I am moving this discussion back to he poised list
(the list of the poisson WG).   While involving the IETF community in general
in this discussion is a good thing, I think there's been plenty on the ietf
list now for anyone who reads it to know there's a debate in progress on
in the poisson WG as to its future.   On the other hand, to discuss the
future of a WG, any WG, in the absence of that group, is not at all the
proper thing to do.   Please keep future discussion on the poisson list
(whether or not you decide to copy it to the ietf list as well).

I have dropped the ietf-process list though - I think it is time to simply
kill that one.  It was a failure for its purpose.   And it continues to
insist that I subscribe to it using the address that's in the envelope
sender header (Return-Path) of messages I send if they're to avoid sitting
in a queue waiting for a moderator to approve them - which is completely
bogus ... there's no way I am ever going to subscribe my envelope sender
address to anything (that's for bounce mail only), and nor am I about to
change that when I send messages to that one insane list.



<Prev in Thread] Current Thread [Next in Thread>