Two document stages: Proposed and Full

Dave Crocker dcrocker at
Tue Jun 10 16:49:03 CEST 2003


(Harald's request the this solution discussion be moved to the solutions
list is duly noted.  So the reply on this list is non-solution related.)

CEP> I think that's a mistake.  I'd rather have it that Proposed Standard
CEP> requires interoperability.  Do you happen to know when the interoperability
CEP> requirement was relaxed?

It was never 'relaxed'.  When Proposed/Draft were first defined, Draft
called for interoperability and Proposed mostly did not.

The "mostly" is the interesting part.  The language I recall for
requiring interoperability at Proposed was something like:

The specification pertains to essential (ie, fragile) Internet
infrastructure, or the specification is sufficient complex (ie, its
physic are poorly understand) so that an interoperability demonstration
is needed to provide a degree of comfort about basic viability and/or
safety of the specification.

For most specification, we rely on a static review of the document, to
convince us that we understand how it behave and that it will not damage
the infrastucture. When we can't develop that code just from reading, we
require implementation.

A separate point that Marshall Rose has been good at pointing out is,
essentially, a marketing issue for a specification: If it's been
implemented, it is much more difficult for people to make abstract
claims about failings of the spec. Hence it is more difficult to reject
a request for Proposed status.

 Dave Crocker <mailto:dcrocker at>
 Brandenburg InternetWorking <>
 Sunnyvale, CA  USA <tel:+1.408.246.8253>, <fax:+1.866.358.5301>

More information about the Problem-statement mailing list