I'm just a guest here, but I agree with you. The 'code fragment wars'
are all very amusing, but miss the point of reliability of implementations
entirely. The current state of HTTP stuff with 'CRLF' just reinforces
this view.
If you use an 8-bit clean channel and then coerce your attribute
names (identifiers) and lengths to 7-bit 'printable' ASCII, all you
gain is that the long-term hope of good quality localization support
for other languages and other code sets is lost irretrievably. It
simply won't get realized, because it's too much hassle for the
implementors.
Justifying the protocol attribute encoding because the first
transport is HTTP is nearsighted and an argument that the IETF
area directors certainly won't sympathize with.
Cheers,
- Ira McDonald (outside consultant at Xerox)
High North Inc
PO Box 221
Grand Marais, MI 49839
906-494-2434