--On Wednesday, 28 March, 2001 10:10 -0800 Dave Crocker
<[EMAIL PROTECTED]> wrote:
> Continued reliance on invitations and hosts ensure several
> problems.
>...
> If we are serious about trying to optimize the meeting in
> terms of cost, reliability and convenience, we need to choose
> a standard set of extremely convenient (and less expensive)
> locations and then keep using them.
We have, so far, found almost no inexpensive (for some weighting
of meeting-setup and attendee costs) sites outside the US. And
your thoughts about doing away with invitation/host structures
entirely, regardless of their other merits, would tend to
further increase costs, especially if we use more than a few
sites. If any international meetings are needed, we rapidly
head into a tradeoff situation not very different, IMO, from the
one we have albeit with a slightly different constraint space.
> Re-use reduces learning curve and that reduces problems (and
> cost).
Total aggregate cost, including aggravation costs, certainly.
Costs as seen by either the meeting organizational process or
the attendees, maybe.
>> the plan has been to do this statistically. I.e., when 2/3 of
>> the active participants are from outside the USA, I assume we
>
> As I believe Randy Bush pointed out, the flaw in this analytic
> methodology is that a meeting in the US is an unequal barrier
> to participation from outside the US. I'm not "voting" for
> changing the current proportion of US/non-US meetings, but do
> feel compelled to note the danger in using history as the
> basis for deciding the future.
When actual costs are considered, the barrier is peculiar, and
neither your oversimplification nor Randy's really work (mine
isn't any better). At many times of year, it is cheaper for me
to get to London, or Paris, or Frankfurt than to San Jose.
Until intra-European fares started to come down, and maybe
still, it could be cheaper to get from Stockholm to the US than
from Stockholm to many parts of western and central Europe.
As someone else pointed out, the percentage of US participants
at a non-US meeting probably gives a better clue about how we
should be distributing things than the percentages of non-US
people at a US meeting. I suggest that, although one could
argue some bias, the statistics get even more interesting if one
discounts "visitor" effects, e.g., counts only people who are on
their second or later IETF meeting or people who come from more
than a few hundred km from the meeting site. (I don't have
anything against visitors, but, if we wanted to optimize IETF
meetings for drop-in, or wander-in, traffic, it is fairly clear
that we would hold all meetings in San Jose or its close
vicinity.) And, by any of those weighted or unweighted
statistics, we can't justify more than one meeting a year
outside the US without appeal to social arguments rather than
where the participants are based.
> On the average, IETF decisions are best made when they focus
> on the primary concerns of a situation and not on the
> ever-present mass of other issues.
I agree completely. But I think that suggests that the present
formulae are probably just about right, at least for the
present. As I note above, your concerns are lowering dependence
on hosts and sponsorship would, IMO, tend to reinforce, rather
than change, those formulae. E.g., despite the cold and some
accessibility factors that have been raised on this list lately,
I've favored going back to Minneapolis which, with our third
appearance there, will definitely fall into a category of
convenient and inexpensive locations which we know how to make
work.
john
