Re: Design Teams (was v 1.2, IETF material)

[Bob Stewart <[e-mail]>]

This is a messy issue.  I'm living through part of the mess right now.  SNMPv2
is being developed from a proposal submitted by a design team.  I'm having to
constantly fight the perception that the well-known proposers are simply
getting their way, even though we are spending considerable mailing list and
meeting time trying to determine an open consensus.  I don't know if I'll win
the perception battle.  I'm not even convinced such victory is possible.

The design team submitted a comprehensive, well-considered proposal, including
four implementations.  We waited for other proposals, and got none, so theirs
became the base.  The clear, strong, consistent community consensus has been
that we should complete our work quickly, based on their proposal.  This has
not kept us from protests that we are sacrificing speed for quality and that
the authors are getting their way because of an old boy conspiracy.  In fact,
we've made a lot of changes, but no substantial changes have been handed out
simply for political peace, partly because I object strongly to such
compromise of quality.

They didn't document all their discussions, including everything they
considered and why they did what they did.  I don't believe you can ever do
that well enough.  I've often heard doing that discussed but never saw it
tried.  Time pressure and the heat of progress cause such good intentions to
slip away, and new people will always believe their idea is enough different
that the archives don't count anyway.

I don't see any new way to fix this without sacrificing the quality of work
done by a small, focussed, experienced group to the madness of design by
committee.

The only solution I see is what I'm doing: everything I can to be open and
fair while maintaining consensus-based quality and getting done in the time
demanded.  I can only hope history will be kind.

	Bob