On Wed, 15 Jan 2003, Csongor Fagyal wrote:

> I agree - partially. I would say we needed a FINAL definition of IDNs,
> their encoding, nameserver setup and so on. Once these are ready, I am
> sure that all open source guys around will insert IDNs into their
> software - say, Mozilla, BIND, sendmail,postfix, and so on and so on...
> That would solve most of the problems, but it will not happen until the
> specs are final.
> However, one cannot expect IDNs to work on all browsers and e-mail
> agents: IDNs are not backward compatible, so to say. I would be happy
> with a simple "it is settled, green light to IDNs" from ICANN (or
> Verisign or whoever).

i stand to be corrected here - my understanding of events in the idn world
is fuzzy.

a standard already exist which is supported by a number of software
developers including IE.  thats utf-8 - those interested should find the
rfc at the ietf archives.

but verisign for some odd reason went with their own standard which
required plug in support.

now i remember reading a notice from verisign on nanog which detailed some
weird proxy majik with the dns and port 80.  all queries to the tld
servers carrying their idn's will be answered by a proxy server which will
then look up the respective domain name and proxy the session.  i always
say if it looks like a duck and quacks like one - then we most likely have
a duck and the egg laying is next.

this whole proxy service to make idn work without a plug in defeats the
general purposes of the dns.  and introduces another point of failure -
i.e. the verisign proxy.

so i assume there's some hope.  i wonder if anyone has tried to register a
utf-8 name in the dns maybe a cctld to test and see if it would work with
a browser without any plug in.  would be fun to know the results.

cheers
joe baptista
yes .gods on our side www.dot-god.com

Reply via email to