--On Sunday, 06 June, 2004 16:43 -0700 Dave Crocker
<dhc2@xxxxxxxxxxxx> wrote:
> John,
>
> JCK> the outlines of this were
> JCK> presented in plenary in Seoul and the general consensus
> was that JCK> people should move ahead with this process,
>
> 1. Attendance by the on-going IETF pool of participants was
> _very_ poor. Any attempt to claim that positive feedback at
> that event is somehow representative of the larger community
> should raise some deeper questions about IETF process. (And by
> the way, claiming that that room showed strong support does
> not match my own perception or the perception of quite a few
> other attendees.)
I was there the first part of the week. I pretty much know who
was there and who wasn't. Oddly, I've expressed the concern
that a particular plenary, or process WG session, was not
reflective of community consensus before (and that was why I
tried to choose my language carefully). But my recollection is
that on the previous occasions, you've taken the other position,
i.e., that the plenary is our best single location mechanism.
On that, and on the classic problem of people having mixed
perceptions based possibly on their biases coming in, see below
(and the latter are really old problems, in the IETF and
elsewhere, but clearly one should be careful about drawing
conclusions).
> 2. An "outline" is very different from a detailed
> specification and I carefully stated that Harald's note
> appears to indicate that things have moved beyond
> consideration and into actual implementations. If that is not
> what is happening, then Harald should explain what actual is
> happening.
I didn't hear what was actually presented there, since I didn't
hear it, and selected the word "outline" to be conservative.
But the I-D strikes me as reasonably specific and it was posted
before the meeting and, I assume, pointed out there. People in
the IETF are, if I recall, expected to read documents on
subjects they find relevant. And even to comment on them if
they don't like them.
> 3. Presentation and discussion at a face-to-face IETF meeting
> is never an IETF decision, remember? What happened to the
> explicit confirmation with the _real_ IETF plenary, namely the
> online community?
Oh, I remember. See above and below.
> JCK> Nothing has appeared on
> JCK> the IETF list, or in any other place that I'm following,
> that JCK> would convince me that there is community consensus
> against JCK> their proceeding, and
> draft-daigle-adminrest-00.txt has been JCK> posted since
> February and hasn't seemed to draw much negative JCK>
> attention either.
>
> 1. So this is one of those "proceed if there is no massive
> objection" kinds of decisions? Shouldn't a fundamental change
> in IETF administrative structure carry an expectation of
> rather more proactive, broad-based support, rather than merely
> waiting for some sort of opposition constituency to assert its
> ugly head?
Well, maybe. Or, more specifically, "absolutely yes, that is
how it should be... only the IETF pet pig hasn't been getting up
to takeoff speed lately". I think it is a frightening
situation, but we are approving _standards_ which only a handful
of people in the relevant WG have read and understood, and the
modal number of comments on IETF Last Call about such documents
is zero.
The two documents I cited make a rather strong case that sitting
around and doing nothing in the hope that things will get
spontaneously better is a seriously bad strategy. That case
(or, if you prefer, problem statement) is independent of any of
the options about what one might do instead. So, yes, I'd much
prefer to see broad-based support for a particular option. And
I'd prefer to see other options placed on the table in the usual
form of I-Ds that make alternate suggestions. But no such
alternatives have emerged, nor has there been any sign of a
counterargument to the "sitting around waiting for things to get
spontaneously better almost certainly won't work" hypothesis.
And that makes this a lot, IMO, like our standards situation:
some people have put a lot of effort into studying the situation
and the options and have put forward a proposal. They have
asked for comments on the proposal and have gotten few if any of
them and there have been neither counterproposals nor coherent
arguments that they are trying to solve a problem that is best
left unsolved. In those situations with the standards process,
we move forward, rather than sitting around waiting until
massive support emerges. That low level of community review and
comment makes me pretty uncomfortable --I think it should make
all of us pretty uncomfortable-- but the alternatives are worse.
Whether the level of specification is adequate or not depends, I
think, on how far into detailed management (I'm tempted to say
"micromanagement" but it is too pejorative) the community wants
to go. I think the evidence is "not very far", a few of us who
get anxious about anything we haven't personally poked into
notwithstanding. In particular, this is not, IMO, a
"fundamental change in IETF administrative structure" (your
opinion may differ, of course). While it creates a platform
from which other, more fundamental, changes could be launched
were the community to decide to do so, all this one really does
is to move us into a situation in which all of the incoming
funds go into one pool so that the IETF leadership (I hope and
trust with adequate consultation with the community) can
allocate that pool accordingly, made sensible decisions about
meeting fees, etc. To facilitate that, it creates a one-person
staff who is really responsible to them.
That contrasts with the present situation in which one
organization is able to make its own decisions about what is and
is not important and in the best interests of the IETF and the
Internet, appoints its own staff, sets its own budget, and
adjusts the meeting fees to match. They do that in consultation
with the IETF leadership, but have no obligation to follow any
recommendations or conclusions from the consultative process (or
even to provide full details about cost and expenditure flows).
And there is also another organization that collects funds and
disburse them to yet other organizations, with more oversight
from the IETF but no mechanism for either drawing on the meeting
fees or underwriting IETF costs to keep those fees lower. Now,
that situation, or at least significant parts of it, actually
have some advantages that have served the IETF well over the
years (especially when there was more total money floating
around). But one thing I don't think any of us can do with a
straight face is to say that the proposed new system gives IETF
participants less insight into, or ability to control, things
than the status quo.
And my evidence that the typical IETF participant doesn't give
the proverbial back end of rat about this as long as things work
--and is willing to trust the leadership to sort things out so
they keep working-- is not the attendance at a plenary or two,
or the general silence on the mailing list, although both
reinforce the conclusion. Instead, it is the observation that
almost no one has taken the time, especially before RFC 3716 and
the I-Ds that led up to it were published, to actually find out
about and understand even the above minimal summary about how
things work today and have been working for the last decade or
more.
Do I wish we had more participation and involvement? Yes. Do I
wish we had better and more participant review of both standards
and these sorts of administrative procedure changes? Yes. Do I
want to see micromanagement, or management by mob, of the
external administrative model? Nope, but largely because I
don't think it would work well and, in particular, because that
I think that such a process would mostly turn IETF over to those
who could (and would be inclined to) spend a lot of time yelling
about things they didn't like, rather than to people selected
through the Nomcom process.
Do I think we can afford to do nothing --either about the
external administrative and budgeting model or about issuing
standards-- until all of those problems (and I do think they are
problems) solve themselves? Not unless we just want to fold our
tents and stop this IETF thing now.
I'll try to respond to the rest of your note at another time,
but I think most of it just misses the point.
FWIW, I would be a lot more nervous about the present situation
if the proposal were "hire a General Manager on a long term
contract with a specific mandate that was unknown to the
community". But that isn't the proposal. The proposal is, as
Harald has described (or, if you prefer, "outlined") it, to get
a consultant on board to take the next steps in putting the
framework together. I assume that will produce another report
and another opportunity for review.
Two more, nearly-out-of-context, comments...
* Yes, we could cut meeting costs considerably by holding all
meetings in the US and preferably, from a cost standpoint, in
the off-seasons of places that no one wants to go to then.
Minneapolis in the winter often comes up as an example in this
regard, but I can easily think of cheaper options, like Phoenix
in July. But we are supposed to be an international body and,
while people can participate more or less fully in our technical
work without setting foot in meetings, we have made rules for
ourselves that require meeting attendance as a condition for
taking on any sort of leadership or steering role. Independent
of the costs of US meetings for non-US participants (when that
was a key issue, we certainly could have debated different types
of weighting functions), the US has now created an entry visa
situation that makes it essentially impossible for network
engineers from a number of countries to attend. I didn't cause
that problem, nor did you or the IESG, but we need to deal with
it or give up almost any pretense of full participation of
people from many parts of the world. I think that is too high a
price.
* Like you, I think there are a lot of other ways to cut costs.
But, in the present environment, it is impossible to get
sufficiently precise budget figures to be able to make that
determination, so both of us are shooting in the dark. For the
record, when I was IAB Chair and a part of the formal budget
review process, I couldn't get those figures either. The
reforms suggested in these various documents appear to be
prerequisite to either getting those figures (nothing else has
worked and many things have been tried) or to putting at least
some of the cost centers out for competitive bid. Either one
would enable the type of transparency needed to make real
decisions about costs and cost control tradeoffs. So it seems
to me that you ought to be supportive, or at least tolerant, of
this effort if only on that basis alone.
john
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf