Two comments:
1) I did not like, do not like and will not like any proposal that
has the RFC specify 14 headers to be treated in a special manner,
with no possibility for extension.
This has been discussed before when the question arose as to whether
to cover headers used by the network news standard. The decision,
supported by a network news guru, was that we should only cover the
headers in the e-mail rfcs. If other groups want to handle extended
character sets in the headers they define then they should specify
their own rules. For example the news rfc could say "If a Header-charset
header is present then it also applies to the Organization header".
Or do you want to cover any of the other existing e-mail headers?
If so which and why?
2) I lost a lot of enthusiasm for "just use mnemonic" when I discovered
the following "mnemonic encoding": :"< (escape-char doublequote
left pointed bracket).
[Isn't the escape-char for Menemonic an "&"? I assume so below.]
Even if the encoding is 7bit this becomes "...&\"<...". Is there
any reason to believe that angle brackets cause problems in quoted
strings? Or in comments? Let's make some allowance for buggy software but
let's not be frightened by ghosts. I think we should have genuine
broken software pointed out and not just worry about the possibility.
However if you read the whole document you will see that it allows
for the possibility of using quoted-printable to avoid explicitly
using any characters that might cause problems: it mentions double-quote
which would otherwise need to be converted to a quoted-pair. It doesn't
mention angle bracket because I didn't (and don't) think it a problem.
If I'm wrong we can easily extend the use of quoted-printable quoting.
Bob Smart