Greg,
I think the main point of contention is your statement that
"MIME/PEM is near completion". I'm not sure I would agree with your
characterization of that document. We spent nearly all of the time
allotted for the MIME/PEM discussion arguing over the rathole of
whether or not to kill RFC1421 PEM, and not addressing several other
substantive issues still lurking about in the MIME/PEM ID. At the very
least, I can not imagine our reaching closure and sending up a MIME/PEM
specification as an RFC before the July IETF, and depending on how many
people show up in Amsterdam, and whether or not we actually resolve the
issues by then, we're looking at least a minimum 3 month delay, and
possibly as long as a 6 month delay if we need to go to the IETF meeting
after that.
Perhaps you will understand some of the exasperation of various
PEM working group members if you reflect on how (for at least the past
year), at nearly every single IETF meeting, we seem to be nearly there,
and at every single meeting, some little snag comes up which delays
things for "just another 3 months". At this point, what with impatient
users asking what is holding things up, and with rival technologies
available for the FTP'ing and compiling, there is definitely the feeling
in some quarters that the following saying definitely applies:
"There comes a time in any project when you must
shoot the engineers and start shipping the product."
Your main argument for killing RFC1421 PEM does not seem to be
that it is difficult or impossible to do. It seems to be that if people
use RFC1421 PEM UA's, this will delay the use of MIME/PEM. I don't see
how this is so. As you have rightly pointed out, RFC1421 PEM has many
limitations --- it does not support character sets other than US-ASCII;
it does not support the sending of 8-bit binary messages; and so on.
Therefore, people who really have these requirements will very quickly
switch over to MIME/PEM. As for people who only need to send textual
US-ASCII messages, they can use either RFC1421 PEM and MIME/PEM; all we
need is to have the compatibility filters that can translate back and
forth between these two formats. Merely having these compatibility
filters should not slow down MIME, if MIME is as wonderful as all the
MIME advocates believe!
So --- why don't we continue work on the MIME/PEM standard, and
stop arguing over what happens to RFC1421 PEM? We can decide this after
MIME/PEM is ready. If at that point, there is no significant RFC1421
user community, then things will be easy. If not, then there will need
to be compatibility code. The only cost to this path is that we need to
assure that there is a non-cryptographic method to translate back and
forth between a RFC1421 PEM message and a text-only MIME/PEM message.
This does not seem to be difficult.
- Ted