What's wrong with protocol competition?

[Dan Bernstein <[e-mail]>]

I like the simplicity and openness of the Malamud-Crocker proposal. I'm
glad they took the suggestion of making objective requirements for the
standards process. Unfortunately, there is a fundamental flaw in the TTF
model---one which exalts the Technical Board above the consensus of the
community.

In the Malamud-Crocker proposal there are Working Groups, Design Teams,
and problems. Everything is centered around the problems. By definition,
there is exactly one Working Group for each problem that the Technical
Board wants to deal with. Although multiple Design Teams can go off on
their own and solve a problem in different ways, all their solutions are
reviewed by the one Working Group for that problem. Presumably the
Working Group will not want to take the time to consider more than one
solution for its problem. So all but one of the Design Teams will be
screwed.

I find this appalling. Interoperability is the ONLY justification for
an open standards group to prohibit competition between two solutions.
As long as we're careful to keep non-backwards-compatible protocols off
of assigned ports, to assign numbers properly, etc., we have NO reason
to prohibit multiple working groups taking different approaches to a
single problem.

I suggest that Malamud and Crocker replace their problem-oriented
approach by a solution-oriented approach. There should be one Working
Group per solution. If the people solving a problem one way and the
people solving a problem another way want to join forces, fine; but
if they don't, the Technical Board should not have the authority to
choose one group over the other. As long as both groups satisfy the
objective requirements of the standards process, both should be able
to proceed independently.

---Dan