Re: [idn] Re: Another solution?

[Dan Oscarsson <Dan.Oscarsson@trab.se>]

>I think many people have been thinking about this.  However, people
>seems to fear that designing a new protocol would increase the speed
>of namespace fragmentation.  foo.com would be available in DNS and not
>in INS and vice versa.  I believe the "namespace fragmentation"
>argument is a poor one -- namespace fragmentation isn't a protocol
>issue, it is a political issue.  Put pressure on organizations
>handling the namespace to not fragment it and they won't fragment it.
>If the pressure isn't there, they will continue to do whatever they
>can get away with.

Instead of designing a new protocol we could do as John Klensin has
proposed: use a separate class or some other method so that the
non-ASCII names exist in the same DNS world but invisible for the
old world. Backward compatible need not mean that old software must
be able to handle or see new things. The important is that all old
things
still works as before.

Why not require that all protocols over IPv6 MUST use UTF-8?
All software need to be rewritten to handle IPv6, we not
require them to use UTF-8 in protocols?

   Dan