RE: [idn] Comparisons of the proposals

[Paul Hoffman / IMC <phoffman@imc.org>]

At 07:49 AM 3/21/00 -0800, Andrew Draper wrote:
>utf-5:  On-the-wire protocol redefines the meaning of DNS names which have
>         a well-known suffix.
>cidnuc: On-the-wire protocol redefines the meaning of DNS labels which
>         have a particular form.

Agree.

> > Changes to DNS protocol and resolvers
>
>Changes to DNS protocol and internationalisation aware resolvers
>
> >
> > utf-5:  None.
>
>utf-5: For some names DNS resolvers must append well known suffix and
>        encode name in utf-5 (or the reverse).  For other names resolver
>        sends/decodes as is.
>
> > cidnuc: None.
>
>cidnuc: For DNS labels containing non-ASCII characters DNS resolvers
>         must reencode label and prepend identifying prefix.  Reverse
>         process for received labels.

I would change both of these to simply "For names that are not currently 
legal, must encode before resolving, and must decode tagged names coming 
from a resolver." Does that match your thinking?

--Paul Hoffman, Director
--Internet Mail Consortium