Marc Mutz wrote:
...
OpenPGP UIDs are essentially free-form,
Yes, this is one of the advantages of OpenPGP
over other PKI contenders. It says nothing
about how the UserIds are handled. Which means
we are free to design the contents to suit our
requirements.
It's certainly a key benefit for my company's
use. You can't pay enough money to get that
sort of design feature!
but with the
Name (Comment) <mail(_at_)address>
convention for interop.
The whole string is in utf-8. This opens up two possible ways to encode
a IDN (or an internationalized email address later) in the UID:
1. In ACE form (ASCII compatible encoding)
2. In UTF-8
...
In any case, I think that rfc 2440bis as an IETF protocol that uses
UTF-8 needs to include at least a stringprep profile to use. I'm no
IETF expert, though, and it may suffice to reuse whatever comes out of
IMAA, but that's still at least half a year away.
If nothing is said and it's accepted as such, then the slot is
automatically an IDNA-unaware one, to be filled with the ACE form.
I think the key word here is _convention_. If
there was a desire to document that and other
conventions more clearly, than an adjunct
informational track RFC may be a better bet
than putting it into the base RFC.
That way, applications could be aware of the
convention, but can happily know that any other
form is also quite acceptable.
iang