Re: [idn] Domain names and ASCII compatibility

[James Seng <jseng@pobox.org.sg>]

"David R. Conrad" wrote:
> I suspect this largely depends on the definition of "fully-functional",
> however I do not disagree with your sentiment that the existing requirements
> draft, as written, cannot be met using a UTF-8 solution.

The existing requirements can be meet with UTF-8 solution. DNS protocol itself
is 8bit clean (or dirty) so I do not forsee any problem using UTF-8 on DNS. At
least, I am sure a UTF-8 solution can meet all the "MUST" requirements.

However, I forsee there would be a lot of problem using UTF-8 in hostname.
Hostname are presumed to be non-8bit clean, in fact more restricted, in many
other protocols. This is where UTF-8 or other "just send 8bit" proposal to
have a lot of problems, especially those "SHOULD" requirements.

Then again, I have yet to see a proposal which meet all "MUST" and "SHOULD"
requirement yet :) Lets keep trying...

> Rightly or wrongly, I am doubtful that the IETF will buy into a solution that
> does not take into consideration backwards compatibility.  As such, I suspect
> it would be important to ammend the WG's "deliverables" to include some sort
> of analysis of what might break should a non-backwards compatable solution be
> chosen.

It is part of the requirement. In fact, it is inside the charter.

"The document(s) should also provide a technical evaluation of the proposals
by the Working Group."

"All idn protocol documents must fully detail the expected effects of leaking
of the specified encoding to protocols other than the DNS resolution protocol.
They must also contain a summary of the technical opinions of the IDN Working
Group."

-James Seng