Let me give you an example of where this didn't work recently. At San
Diego, we had back-to-back meetings ...
There is another solution for real WG participants. Simply abandon the
meetings to what by someone's estimate is the overwhelming majority of
observers and other dead weights. Do not attend if are in the minority
who cares mostly about the protocols. Give the meetings to those who
want to educate, be educated, and generally rub shoulders with and as
super duper internet engineers. They won't notice your absence.
Even better, save a lot of time and effort by the IAB, IESG, etc. and
hire Comdex or InterOp to run the meetings. InterOp (or whatever it's
called these days) has long offered tutorials. Moreover, InterOp has
often hired IETF participants to do the teaching.
Instead, participate in protocol work by mail. As others have pointed
out, the mailing list consensus is the the only consensus that matters.
When something is decided in the IETF meetings, the odds are about 50%
that the decision is wrong, contrary to the mailing list consensus,
and will be reversed. The classic example of that syndrome was
the decision of the IAB in Japan to pick the ISO OSI suite for IPng.
Note that I am not being sarcastic.
Note also that I know that is no chance that the participants (not
to mention self-described observers) who are now complaining might
take my advice.
Finally note that complaints about hordes of the pointy haired,
marketeers, and other non-participants jamming the meetings are
nearly 15 years old. That ancient IETF that didn't have such
problems, or at least didn't have such complaints was either in
some other universe or didn't last more than a year or two.
Vernon Schryver vjs(_at_)rhyolite(_dot_)com