One final message from me on this topic, then I'm done ...
Date: Mon, 17 May 2010 08:10:01 +0200
From: Eliot Lear <lear(_at_)cisco(_dot_)com>
Message-ID: <4BF0DDB9(_dot_)60003(_at_)cisco(_dot_)com>
| but I do accept that they have the authority to make such a statement,
| if rough consensus could have been shown.
I didn't ever say that the authority wasn't there, in fact, if I recall
correctly, in my original message I think I explicitly agreed that the
IESG has the ability to issue such statements when circumstances warrant.
What I said was that they shouldn't here. There are two reasons.
One is the conflict of interest, the nomcom is a topic, where (just possibly
extreme emergency excepted - which is really hard to imagine here) the
IESG should always ensure the full IETF procedure is carried out.
I said enough about this before so I won't say more here. [Aside: nomcom
procedure as I recall it has always been that if there's an area of
ambiguity or doubt, the nomcom chair makes a ruling, if there was an urgent
problem here that needed fixing, that would be the way to do it ...]
Second, is that there is no need for any kind of statement here, from anyone.
If the IESG simply did nothing, everything would continue working just the
way it has in the past, nothing would break, or fail to work, in any
way at all, nothing that needs a quick fix anyway, and there is no ambiguity
that needs clearing up.
I know there are people who believe that the requirements for nomcom
membership should change, or should be interpreted differently than
they have been, or now that it is practically possible in some cases to
interpret things differently than it has been in the past, we should do
do - and all of that is fine, I don't agree with all of that, but they're
all legitimate viewpoints - but they all represent changes to the process,
and should be achieved via the working group route, and not via the
back door method of an IESG statement changing the rules.
I have seen the argument that RFC3777 always meant "attended for the whole
IETF meeting" and not just "attended" as it says. I'll explain more below
why I think this is an incorrect interpretation, but for now let's just
accept that is what 3777 means, and that (because of that) the IESG's
statement is just a clarification, because attendance on a day pass cannot
mean "attended for all 5 days" (I'll explain below why that's incorrect too,
but let's ignore that problem for now).
I would note that in the revised IESG statement, even the IESG agree that
what they are doing is changing the rules - the only justification for allowing
day pass attendees from the past 2 meetings to volunteer, and not those
for future meetings, is that the rules previously allowed them, but they are
not to allow them any more. If 3777 really meant "attended the whole meeting"
then the original IESG statement would have been the way to go - people
who attended one of the last two meetings on a day pass, clearly did not
attend the whole meeting, and if 3777 does require attendance at the whole
meeting, they should not be eligible. By making the change made in the
revised statement, the IESG is making it clear that they are changing the
eligibility rules for the nomcom - that's something only a WG should do.
But if "attended the whole meeting" is what 3777 really means, then the
IESG statement doesn't go far enough (not the current version, nor the
original), we know from Rus Housley's message ...
housley(_at_)vigilsec(_dot_)com said (Sat, 15 May 2010 11:15:03 -0400):
| Attendance was determined by the presence of a check-in date. That is, the
| person's registration packet was picked up at the registration desk.
that the secretariat have the ability to determine who was actually present,
and what's more, when they were first there (the check-in date mentioned.)
Even if they didn't have that now, we could easily ask them to collect
that information for the future, as (just as been done in the modified
IESG statement), if we are actually making a change, no matter what change
is made, or how - via WG or IESG statement, we don't want to apply
retroactively, only to the future.
With this information, we can (or could) disqualify from counting towards
nomcom eligibility anyone who hasn't collected their badge by (say) 11:00
(11am) on the Monday morning - if the intent is that the volunteer must have
attended all 5 days, then certainly anyone who wasn't there Monday morning
(maybe pick some other time instead of 11:00 - that's the kind of issue a
working group could hash out, of course) was not there for the full 5 days,
and we would know that just as certainly as we do for someone who attended
via a day pass.
Now we have excluded everyone who arrived too late to have garnered enough
IETF experience to be a nomcom volunteer, we just need some way to detect
all of those who leave too early - this one the secretariat would currently
be unable to handle I suspect, but we could ask them to collect badges of
people departing on Friday, say sometime after 10:00 (or 11:00) (Whevever the
Friday sessions end these days), anyone departing before that could take
their badge with them (in case they return) or hand it in (where it would
be ignored for this purpose) - only those people who handed in a badge after
the termination deadline would be showing evidence of having remained until
the end of the IETF and so have (possibly) attended for all 5 days.
We'd still need to deal with people who are at the meeting at the start,
and again at the end, but miss much in the middle, if we are requiring a
complete 5 days of attendance. This would be trickier (maybe the RFID
stuff could be managed to watch for people mingling in the halls, or the
bars, as some messages seem to imply is important in order to become a
nomcom volunteer) - or this could simply be skipped as unlikely to make
enough difference to matter.
Just in case anyone is confused and about to argue that all of this is
too much, and unnecessary, I reiterate, I agree - all of that assumes that
being a nomcom volunteer (the requirements in rfc3777) requires attendance
at the whole of the IETF meeting, and not just attendance at some part of
it - a view that I do not hold.
But there's one more (logical) problem with the IESG's proposed statement
I believe, in that it assumes, as best I can tell without justification,
that someone who uses a day pass only attends for one day. I cannot find
anything that justifies this opinion. I have been looking for the
announcement of the day pass experiment, and cannot find one anywhere
(not in my saved e-mail archives, and not on the IETF (or IAOC) web pages,
I admit to not attempting to search the archived IETF mailing lists, my
saved e-mail contains only what I considered important when the message
was received - I do have the IETF meeting #76 registration announcement
that said about day passes being available as an experiment for the first
time - but that's not very precise about just what the experiment is, or
the rules - it's about what I would expect for where it appears (ie: it
is OK in context) but I was hoping I'd find something more comprehensive
elsewhere, but couldn't.
Anyway, I cannot find anything that says that someone cannot attend 2
(or more) days of an IETF meeting using day passes for each day.
That is, you simply turn up Monday, obtain a day pass, then if you feel
inclined, turn up Tuesday, and obtain another day pass, and again
Wednesday ...
Now (especially at the rates set for the forthcoming meeting) this might
not make much financial sense, in that attending every day on a day pass
would cost much more than simply paying for the full meeting (for the
forthcoming meeting, if I recall correctly, a day pass is > 1/2 the early
bird full registration rate - for IETF76 however, someone could have attended
3 days on day passes and paid less than the early bird rate, or 4 days on
day passes and just about broken even on the full fee registration (so it
was not even financially silly to use a day pass more than once - and even
now I can imagine scenarios where it still makes sense to do it that way,
even if the eventual cost might be higher than paying for the full meeting).
I challenge anyone to explain why someone who attends 5 days and pays the
full fee IETF rate is more qualified to serve on the nomcom that someone who
attends the same 5 days and instead (for whatever reason) pays for 5 day
passes instead of a full fee registration.
But the IESG's proposed policy is to rule the 2nd type of attendee (of future
IETF meetings) ineligible where the former kind is eligible - which is really
a perverse variation of the intent expressed by some of the people who are
supporting the IESG's proposed policy, That is, someone who can demonstrate
that they did attend all 5 days of the meeting is to be excluded, in favour
of someone else who actually only attended 30 minutes, but paid for the full
meeting. Bizarre!
One last topic for this message then you will hear (read) no more from me
on this topic (just possibly excluding the possibility that someone asks me
a direct question not already answered here or in an earlier message, or
just perhaps, if the IESG try a third variation on their statement).
That is, did 3777 ever really intend to require that nomcom volunteers should
have attended all 5 days of the IETF meeting in order to qualify?
The implication being that it was simply too hard to test this previously,
we simply assumed that anyone who paid for it all would attend it all, but
now with (use of only one day of) the day pass type of registration, we can
easily identify some people who did not attend all 5 days, and so we should
exclude them.
I disagree with the assumption - when 3777 (and 2027, 2282 and 2727 before it)
were being discussed, no-one ever considered that it was necessary to have
been at the entire meeting in order to have "attended". 2027 says
(4) members of the IETF community must have attended at least 2 of the
last 3 IETF meetings in order to volunteer.
That's where "attended" originated, later RFCs have qualified things a
little, and changed 2 of 3 into 3 of 5, but "attended" remains as written
then, without any real thought being given to what it means (nor to changing
what it means).
2027 was published in Oct 1996, and discussed in the period before that.
I can assure people who may not have been attending IETF meetings in that
period, that it was *VERY* common for attendees to leave on Thursday
night, and not stay for the Friday morning sessions - my guess just based
upon observing activity around the hotel checkout desk, and number of
people still around on the Friday, is that perhaps 75-80% of all of the
attendees at meetings around and before this period would not have been
present for the Friday.
That is, if "attendance at the whole IETF meeting" is what 2027 meant by
"attended" then there would have been very few eligible volunteers. I
have no doubt that we'd have several people who have served on nomcoms or
at least volunteered to the point of being in the final list (passed the
attendance check) who will admit that in some of their qualifying meetings,
if not all, they weren't actually present for the Friday morning session.
This was not a secret during 1995 and 1996 - everyone in the poised, poised95,
and poisson working groups would have been aware that people come and go
during the IETF meeting, and that only a very small number of people actually
attend the whole thing (whatever that means - if you're in the hotel, but
in some non-IETF private meeting for a day, is that "attending" the IETF
meeting, or not? What's more, even then it was common for people to attend
for just one day - one of the standard issues to defuse was to appease people
who were attending for just one day only to discover that the WG meeting they
wanted to attend had moved at the last minute to a different day.
All this was common knowledge (well, some infrequent attendees might not
have known about the possibility of schedule changes, but those attending
the poised family of working groups, and working on the drafts that became
2027 and successors certainly knew it.)
Had the working group ever intended to require attendance on all 5 days,
and had that survived the absurdity test, then I am quite sure that the
text would have said so. It doesn't, it just says "attended" with no
qualification at all on what that means. Nor does it say "to have attended
a potential volunteer must have been participating in the meeting for
at least 3 of the meeting days" or anything else like that.
Maybe (subject to practicality) something like that might be a good idea,
but we're not here now to discuss what might or might not be a good idea
to require, but what 3777 does require, and whether the IESG statement is
a change of the rules (which really should go to a working group - perhaps
some other options should be considered, not just this one take it or leave
it) or whether the statement is just a clarification of what the rule always
said.
I submit that it is clear that it is not just a clarification, all 3777
(and all of the RFCs that preceded it) ever required was simple attendance,
with no specification at all of what that means, or what counts as having
attended. 3777 doesn't even say how the secretariat should check who
attended - we all assume that they must use the paid attendee lists, and that
seems like a logical implementation method, but 3777 does not require that.
If we can assume (or agree to) all of this, then I further submit that there
is absolutely nothing that needs to be done now - the secretariat is just as
able to tell who paid for, and picked up a badge (of whatever type) and
thus who attended (I doubt anyone who did not pay but attended anyway is going
to be drawing attention to their actions by demanding to be a nomcom volunteer,
but if that ever happens, and they can demonstrate that they did attend they
should be allowed to volunteer - this is so unlikely I think we should just
ignore the possibility until it happens).
Whether the attendance is via a single day pass, or 5 one day passes, or
a full, early-bird, or student, registration should make no difference.
None of that has ever mattered before, there's no reason for us to believe
that it should matter now.
It would be different if we had a report from a nomcom chair that the
nomcom believed that some of its members had insufficient IETF experience,
and that the qualification rules need to be altered. I don't believe that
we have that, but if we do, that would be all the justification needed to
immediately charter a WG to look into the issue, and produce a proposal
for the IETF to consider (with no attempt from me to pre-judge what the
results might be, but I'd be astounded if a change that was just "no day
pass attendee volunteers" was the result - especially since no past nomcom
could possibly have had day pass attendees, day passes didn't exist before
the last nomcom was formed, so they could not possibly have caused whatever
problem might have already been observed.)
Absent a problem, I see no compelling need to do anything at all right now.
Perhaps there is a desire to make some changes, which might grow strong enough
to warrant action (in the form of formation of a WG), but I see nothing
immediately that should require that, or any, action.
Imagining what future problems might arise, and seeking to counteract them
before we know what they are, especially without a thorough examination of
all of the available options to correct whatever it is that needs correcting
is just counter-productive.
Please, IESG, just withdraw this statement, and do nothing at all - nothing
needs to be done (and some subset of IETF attendees believing that something
needs to be done does not make it so.)
kre
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf