RE: comments on draft-ietf-sming-reqs-03.txt

[Bob Natale <Bob.Natale@AppliedSNMP.com>]

At 8/3/2001:01:09 PM, Durham, David wrote:

Hi Dave,

>...
>[Dave] Absolutely, simplifying the language is an important goal. Would not
>enabling reuse and removing some of the weirdness and arcane rules of the
>SMI help achieve this? 

In every massively successful technology (of which SNMP is a prime example
in the Internet protocol space). sub-optimal components )of which the SMI
is a prime example) ride along on the bandwagon.  The weaknesses inherent
in such sub-optimal components generally get compensated for in other parts
of the technology.  This happened long ago (investment-wise) wrt the overall
SNMP technology space.  There is really very little motivation to re-do the
SMI at this point.  As one who works with the technology every day, I can
still see many opportunities for improved products *based on the existing
components* (for the most part).  I am more interested in pursuing that
work )of which the EOS proposals are good examples, on the whole) than 
in going back and *significantly* changing something whose sub-optimality
has already been discounted (over many years of deployed work).  From what
I can glean, most of the current and prospective customer base feels
roughly the same as I do on this matter.

Personally, rather than getting into any acrimonious debates about this
set of work items at this time, I would rather see the community agree
to hand it off (back?) to the IRTF for additional research (including
cost/benefit and feature-interaction analysis) and experimentation.

I make the above comments with all due respect to the authors and those
others who have worked so hard on this effort.  And I apologize for
being one of those who (as Jeff perceived) really has not been regularly
tuned in to this channel.  Feel free to discount my input accordingly.

Cordially,

BobN