Heh. Other viewers present only the first few headers. Or only headers
that are of a certain length. The software I support, PMDF, allows
user-selected headers, but some people use this to strip off all headers.
I've even seen header stripping facilities used to wipe out headers on
messages being transferred via SMTP.
A significant factor in the display of headers is length. Beyond a certain
point additional information simply confuses things and makes people
unhappy.
Both the mnemonic proposal as well as the encoded-variable proposal
have the potential to produce reasonable header displays in extended
viewers. I'm not sure your proposal does, since it is never clear what
information duplicates what. Unextended viewers have to live with an
increase in header size no matter what we do (unless we do nothing) so I
don't worry too much about them.
The notion of alternative versions of header information is a nice one, but
I don't think it will work out to be as useful as you think in practice, for
the reasons I have given here. If it does I'd like to see it done in a
different way -- I have no problem with having multiple alternative
encodings of a single entity specified, and you pick the one your viewer
can best handle. I still prefer mnemonic, however.
Ned