As someone else pointed out, POP and IMAP is where the users feel an
impact. I would argue that few people run SMTP servers on slow links, and
while most users at work are on fast links (typically ethernet), most
users at home or on the road are on slow links.
Thus, perhaps an IMAP (and POP) extension to allow on-the-fly compression
would give the biggest bang for the buck. It has an additional advantage
of not causing the interoperability problems with a new Content-Encoding.
Again, you can use TLS with both POP and IMAP (and this is supported by
quite a few servers). And TLS provides compression as an option, which
is good since TLS will severely limit the ability of modems and such to
perform link-level compression.
Such an extension would be useful even after RESCAP and some kind of
Content-Encoding compression were being deployed and would provide a
benefit to users even if the sender didn't support RESCAP or the new
Content-Encoding compression.
Actually, the extension that's needed in IMAP is one to remove CTEs completely.
Do this and use TLS compression and you should have a reasonably effective
solution for large messages. And its one that even works for messages already
in message stores.
Ned