RE: This WG in context of RFC 2277

[Andrew Draper <ADRAPER@altera.com>]

Patrik Faltstrom wrote:
> Patrik Faltstrom wrote:
> > This is a major work item for the IETF which impacts a lot of things.
One
> > can not, I repeat not, "just" fix things by storing things in DNS and
hope
> > that applications using domainnames can handle some encoding without
> > reviewing the RFCs specifying the applications, and probably updating
> > them.
> 
> I have to emphasize that I said that we should REVIEW the RFCs. I don't
see
> that we SHOULD update all other protocol that exists in the world, and of
> course is it excellent if we don't have to.

I agree that we should review other protocols as necessary and allow the
appropriate experts to change their protocols.

But this brings out another important point.  Is "should try hard not to
change other protocols" a requirement of the internationalised DNS protocol
or not?  I would argue that it isn't.  

Anyway, to start the list of protocols for review, here are a few
suggestions (any my evaluations of them):

SMTP: Domain names appear in commands and responses throughout.  Not assumed
to be 8-bit clean.

HTTP: Domain names appear in various headers.  I think it's guaranteed to be
8-bit clean.  Many clients and servers are not expecting non-ASCII
characters.

H.323: Domain names appear in addresses, but there is a UTF-16 variant
available. Name storage is 8-bit clean). Most systems deal with domain names
transparently, but there are always UI and logging issues.

Regards,

    Andy