Yes, I'll concede that a 7bit, 8bit, or binary encoded text message
with iso-2022-jp charset is unlikely to survive being passed
through an EBCDIC environment. q-p and base64 should work okay,
though (depending on which q-p characters are encoded).
I'm afraid you might misunderstannd what happened. There was some
conversion performed between ISO-2022-JP and IBM local encoding to
try to make the content survive, which failed completely.
I know how badly ISO-2022-JP-encoded plain text messages corrupted
through EBCDIC environment, which drove BITNET users to
TCP/IP in Japan.
See, EBCDIC *is* good for something after all!
And MIME drives people from RFC 822 to X.400?
Masataka Ohta