Discussions about whether good I-Ds always become RFCs or
not misses the point. You wont stop marketing droids from
making overblown statements about the relative worth of
'publishing' an I-D.... you could make I-Ds last only
the week of an IETF meeting, and marketing droids will
gleefully announce "our proposal was contributed to
the XYZ IETF!".
Anyway the original question wasn't about the archival
worthiness (or not) of I-Ds, but rather the wisdom of
citing I-Ds given their presupposed unworthiness. As
has been noted, RFCs dont (or at least aren't supposed
to, modulo mistakes) make *normative* cites to I-Ds
(although historical cites to 'work in progress' are
sometimes valid). As has also been noted, works
outside the jurisdiction of the IETF (e.g. books)
can do what they please.
So, since we already have in place rules on when
I-Ds can be cited in 'IETF approved' documents, and
the IETF can't stop anyone else from pointing to
IETF works in progress, is there anything left to
discuss here?
cheers,
gja