Re: [idn] New protocol proposal: IDNRA

[Keith Moore <moore@cs.utk.edu>]

if the experience with RFC 2047 is any indication, it will be 
approximately 8-10 years before 75% of existing mail clients 
support display of IDNs, and much longer before mail clients
can be expected to display IDNs properly - i.e. before a
sender can reliably count on an IDN being displayed by a
recipient's mail reader in "native" format as opposed to ASCII
encoded form.

(even today, 9 years after the first MIME RFCs were released, 
mail clients cannot reliably be expected to decode MIME, and
RFC 2047 support lagged the rest of MIME)

on the other hand

- this timeframe would be many times worse if we adopted a  
  system which required *every* interface to change to support IDNs
  (i.e. if we had to change not only the input and display of 
  email addresses within user agents, but also the message format, 
  SMTP, DNS, etc.)

- the rate of adoption will be much quicker in newly installed
  mail clients (as opposed to existing mail clients being replaced)
  given that the non-English-speaking portion of the Internet 
  is growing very quickly, if new mail clients which support 
  IDNs can be made available in a year or two, there is an opportunity
  for those new users to receive IDN-capable user agents when they
  first connect to the Internet, rather than later.  this would
  allow IDN capable mail clients (and presumably also other applications)
  to be deployed more quickly than otherwise.

Keith