The Routing Over Large Clouds Mailing List Archive by date

[Grenville Armitage <gja@thumper.bellcore.com>]

Tim,

>>(Do we need to solicit opinions from the broader IETF community about
>>the appropriate definition of the term "version?")

Is that what you want? Another colossal waste of time as keyboard-pedants
jostle for post-christmas fame and fortune arguing over the difference
between a 'version' of a protocol and a simply different protocol?
Please, tell me it ain't so!

	[..]
>>On the other hand, I believe that NHRP and MARS should be be allowed to
>>evolve independently.

This is unfortunate, because I consider you therefore to be completely
and utterly wrong. They ought, if anything, converge as time rolls
on.

And I think - nay, I know - you've not fully read the spec if you
think the evolution of either NHRP or MARS will be impeded by the
current fixedheader and 'version' field. Why? Because the upgrade
path to incremental modification of functionality is primarily intended
to be through the TLVs. Re-read the intro to section 10 of ipmc-10.
(And have a look at draft-kandlur-ipatm-mars-directvc-00.txt for an
example of exactly this sort of approach to evolution.)

Its fairly obvious you've simply got a hangup about a 20 octet
fixed header. Why?

gja (truly bemused)