Re: [idn] internationalized email addresses

[John C Klensin <klensin@jck.com>]

--On Monday, January 08, 2001 11:52 PM -0700
ned.freed@innosoft.com wrote:

> Exactly. There is little point in considering local parts now
> unless they additionally constrain the problem. (AFAIK they do
> not.) And even if they did, the requirements for host names are
> already so onerous that adding additional ones imposed by other
> identifier spaces seems like a really bad idea. We need
> closure, not further muddying of the waters.

While I agree, we do need to be a bit careful about this, e.g.,

(i) As we move functionality from DNS resolvers to the nameprep
function we need to remember that local-parts are case-sensitive
in ASCII while FQDNs are not.  Preserving that distinction may
require slightly different rules in the two situations.

(ii) While I don't see any obvious issues with local-parts that
would argue for either UTF-8 or ?ACE against the other, we again
will need to be careful:  the variety of conventions used in
local-parts around the world is such that a string that looks
like an ACE prefix is perhaps more likely to occur by accident in
a mail local-part than in a DNS name.   That said, we probably
have more experience with adverse impact of unexpected appearance
of 8bit strings in email clients than with other classes of
clients.  And much of that experience has been fairly horrifying.

Or is that what you meant by "other facets"?

    john