Whoah. Did I miss something somewhere?
The sub-types name should be indicative of the specific compression
algorithm. If a second gets standardized, then we should specify
an additional set of subtypes (one each for binary, 8-bit, etc.) with
the name indicating this specific, new algorithm.
This implies
Content-Transfer-Encoding: base64/compressed
rather than
Content-Transfer-Encoding: compressed-base64
which is what I *thought* we were debating.
David