The feedback was from a subset of the WG. For some time my
IETF meeting minutes have included a discussion of the percentage of
attendees who were, by their own admission, new to the WG, and who did
not profess to be thoroughly familiar with the topic being discussed.
This percentage was always a MAJORITY of the attendees. Thus the
results of the poll taken at the mtteing is not necessarily indicative
of the sense of the WG.
Drawing conclusions like this from the data you have presented is entirely
nonsensical. It is in fact quite possible for a meeting involving absolutely
nothing but long-time members of the WG to produce results at odds with the
working group as a whole -- all you need is for the attendees to be from a
non-representative subgroup. (This outcome occurred at the Atlanta meeting of
the IETF-822 WG, as a matter of fact.) Similarly, it is equally possible for a
meeting with nothing but entirely new members to produce results that do
reflect the wishes of the group. (Something very much along these lines
happened at the IETF-822 WG meeting in Santa Fe.)
You state that a majority of the people in attendance at previous meetings have
been newcomers. So what? The IETF has grown dramatically in size over time, and
there are always a large number of new faces at meetings these days. All you
have demonstrated with this statistic is that the composition of the PEM WG
meetings is not materially different from other WG meetings. This does not
translate into any sort of conclusion that there was only a nonrepresentative
subset present -- the two matters are orthogonal.
Besides, the composition of earlier meetings of the PEM WG is irrelevant here
anyway -- the composition of the most recent meeting may have been (and in fact
in my opinion was) quite different.
The only relevant issues here are whether or not the WG had an adequate
understanding of the specifications to make an informed decision, and whether
or not a representative sample of the membership was in fact present.
The former was assessed by asking how many people had read the specifications.
As the minutes show, roughly 25% said they had read the current draft and 50%
said they had read either the current or previous draft (the two are very
similar). These numbers are way above average for IETF WG activities, as I'm
sure you know -- I was surprised by the show of hands at the time.
As for assessing whether or not there was adequate representation, well, that's
why we have last call.
Hence the referendum. If there was a real
sense of urgency in getting this spec out the door, I would have
expected the authors to have revised the previous draft, in response
to comments documented in the previous IETF meeting minutes
(distributed to the mailing list one day after that meeting) and in my
more detailed comments (which were delayed by a week or so), sooner
than the traditional 1.5 weeks before the next IETF meeting.
I have already described why this pattern has developed in this WG. It in no
way translates to any lack of a sense of urgency in getting these
documents progressed.
Ned