Jim,
I accept the rationale you provided re why the authors of the
first ID elected to move the definition of the keys and signature
control body parts out of the first ID, but I maintain that my
characterization of why they should be moved is valid. I question the
utility of establishing control body part definitions that are so
generic as to embody almost no syntactic or semantic constraints.
However, if such a construct turns out not to be really generic, then
it clearly belongs in the document that defines the more detailed
syntax for a specific application. That was my interpretation of a
good reason for moving the body part definitions in question, though I
defer to your explanation for the author's rationale and I agree that
the minutes should be amended to reflect that author's rationale for
this proposed change.
As for the generality of originator and recipient IDs, I think
we disagree again. Here, the second I-D specifies a generic format
for both IDs and gives the impression that the generic character is
intended to unify expression of both forms of ID. My notes point out
reasons why this may be an inapporpriate generalization. While I
agree with your observation that one can fnd the necessary constructs
in the I-Ds to transfer all of the requisite information, there was no
argument, no supporting rationale, that a particular mode of use would
be more common than another, and thus the design should be optimized
on that basis. I don't necessarily agree with the suggestion that the
transmission of one or more certificates will be the rare case, as
that may require too much state or too much thought by originators,
especially for messages sent to multiple recipients. My point about
the different requirements for both forms of ID still stand. Since
the minutes characterize this point as a suggestion made by a WG
attendee, I think the minutes are accurate in that regard and should
not be amended.
Steve