Reply to Reply to Proposed change to HARPOON

["Kevin E. Jordan" <[e-mail]>]

> So, I suggest to keep to bilateral with Binary, and IA5 with Quoted-Printable
> or Base64 as the legal encodings, bilateral preferred, IA5 admitted only
> when we know the destination is not bilateral-capable.
> Comments?

It looks like you are suggesting that we allow either Bilateral or IA5, but
make the default Bilateral.  Is this correct?  If more than one format is
available, then there must be a mechanism defined for making the choice,
and possibly a default in case the mechanism fails to produce a definite
result.  So, it appears that you are advocating two possible formats, with
Bilateral being the default format.

Erik has reminded me that an important goal of HARPOON is to define how
MIME encoded in X.400 '88 is downgraded to MIME encoded in X.400 '84.  Ned
is arguing very strongly for preserving existing X.400 UA's and gateways
such that they will continue to work somehow without having to be changed
to accommodate HARPOON.  The only way I see that both goals can be achieved
is to specify that MIME must be encoded in '84 X.400 in exactly the way that
it is encoded today by RFC987-compliant gateways.  This implies that a
leading IA5 body part is created to contain the initial MIME headers, and that
subsequent MIME message parts are appended as text-encoded IA5 body parts.
It also implies that HARPOON must be rewritten to specify that MIME encoded
in '88 X.400 must be translated to the RFC987-compatible representation when
downgrading to '84 X.400.

Allowing both Bilateral and IA5 is impractical.  Administrators and
implementors will tend to do the easiest thing possible; support/implement
the default choice only.  Besides, if we allow both, who will define the
mechanism for making the choice?  Would it be necessary to add a new MTA
attribute to the COSINE WEP tables?  To the set of MHS-DS X.500 routing
attributes?

HARPOON should define exactly one of the two representations.  My personal
perference is the representation which uses BilaterallyDefined, but I
understand and sympathize with arguments for the IA5 approach.  Based upon
Erik's reminder, HARPOON is needed, but the technical content of HARPOON will
differ based upon whether we choose to encode MIME in BilaterallyDefined or
IA5Text.