Re: [idn] Openendedness of Unicode/10464 (was: RE: [APTLD iname 24] Re: [id n] Proposed suggestions from AsiaPacific Top LevelDomain meeting)

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

At 07:11 PM 3/4/00 +0100, Karlsson Kent - keka wrote:
> > -----Original Message-----
> > From: Paul Hoffman / IMC [mailto:phoffman@imc.org]
>...
> > We probably only have one shot at doing this, at least for
> > the next five
> > years. There will be massive complaints if we do IDNv1 now
> > and then try to
> > do IDNv2 in two years, particularly if v2 allows some characters not
> > allowed in v1.
>
>But, but, ...  This is what we MUST allow.  New characters are added
>to Unicode/10646, and will continue to do so for many years to come.

Sorry if I was unclear here. I didn't mean that IDNv1 should say "no new 
characters are allowed after today"; I fully agree that it cannot. What I 
meant was that IDNv1 should say "all characters from present or future 
10646, other than those in the table of unallowed characters, are allowed."

The disagreement is how we handle things that get added in future 10646 
that match our criteria for disallowed *types* of characters. An easy 
example: if we choose to disallow "non-name symbols" (such as U+263B, BLACK 
SMILING FACE) as a type, and list all the disallowed non-name symbols in 
IDNv1, what should be the case for non-name symbols that is added in a 
future version of 10646?

Are they allowed until we come out with IDNv2? That wouldn't work, because 
they would be allowed for some period of time but then not later. Are they 
allowed forever? That gives a strong incentive for companies to try to add 
symbols to 10646 later, unfortunately.

Maybe the only resolution for this problem is to come up with an IANA set 
registration and have the reviewer for that registration closely follow ISO 
additions to 10646, looking to add everything other than the types that are 
disallowed in IDNv1.

FWIW, this is not an academic exercise. Dozens of symbols and punctuation 
were added to ISO 10646 in the past five years.

>Yes, I know, this WG is not about what registrators are allowed to do,
>but this issue falls somewhere inbetween 'protocol' and 'policy'.

There's another reason why this group isn't considering "registration": 
there is no registration for many names. I own the domain name imc.org, and 
I don't need to register with anyone if I want to resolve the name 
"&&&.imc.org". The IDN should say what is and is not allowed on the wire, 
period.

--Paul Hoffman, Director
--Internet Mail Consortium