My sense is that the demand for IDNs arose primarily in the web space. It
is entirely reasonable that businesses with well known,
non-ASCII-expressable, names would want to employ them directly in URLs.
Clearly once IDNs are out there, they will start leaching into mailbox
addresses. Displayed directly in their ASCII form, they may be
unintelligible, but so too will be their decoded UNICODE equivalent to
many foreigners. Those of us with a long history in email remember the
days of totally unintelligible local parts (i.e. UX1234), and even today,
who the hell is JimB? In these cases, we depend on RFC 2822 display-names
to represent a meaningful description of the local-part, which post RFC
2047 can be expressed in a range of alphabets. My expectation is that with
the introduction of IDNs, the display-name will have to begin describing
the domain-name as well. This leaves the question of non-ASCII local-parts
unaddressed. I'm not sure they need to be.
Nick
Nick Shelness
Independent Technology Consultant
Contact Details
Office Tel: +44 (0) 1828 640 632
Office Fax: +44 (0) 1828 640 647
email: nick(_at_)old-mill(_dot_)net