I think we would do well to require a reference implementation as a
condition for Proposed Standards from new working groups or individual
submitters...but there are other conditions that we should impose that are
far more important. Such as, a requirement for formal cross-area review of
the design goals document and of preliminary specifications as a
prerequisite before producing a reference implementation.
How utterly sclerotic. And what is the IETF, the code police? If ever
there were a need to make the IETF utterly and completely beside the
point, this suggestion would be the perfect way.
Heaven forbid the internet Engineering task force should actually do
engineering. Remember engineering? That's the discipline of producing
efficient designs that meet predetermined goals and requirements.
IETF is already plunging toward irrelevance at terminal velocity. The only way
arrest the descent is for it to start producing better quality and more
specifications. A good start would be for it to actually pay some attention to
the problem definition and rough specification phases and to conduct them in an
environment where they can get meaningful review outside of a narrow community.
Ietf mailing list