Re: Inputting mixed SC/TC (Re: [idn] A question...)

[John C Klensin <klensin@jck.com>]

--On Saturday, 09 February, 2002 10:10 +0800 xiang deng
<deng@cnnic.net.cn> wrote:

> Chair's responsibily is not to create conflict but to collect
> all different argument and coordinate to find a possible
> solution.

Under normal IETF procedures, which apply to this Working Group
as well, the Chair's responsibility, at this stage in the
process, is to move work along.  That includes, quite
explicitly, 

	* Discouraging the further discussion of topics on which
	the WG has already reached a conclusion.
	
	* Discouraging discussion of problems that are either
	out of the WG's scope, or for which realistic solutions
	(e.g., solutions that deal with all appropriate cases)
	have not been proposed.

James (and Marc) have, if anything, been more tolerant of
extended discussions of topics the WG can't fix, or has already
reached consensus that it won't try to fix, than is normal in
the IETF.

Since I have started on this, I am very concerned about the very
real and serious problems with TC<->SC.  I am also very
concerned about problems with look-alike characters in
alphabetic character sets.  I am also very concerned about the
implications of non-language characters in Unicode that, from my
point of view, don't belong in the DNS (however coded).  

I have raised those issues with the WG, discovered that the WG
isn't interested in the problems, and gone on to worry about
contexts in which those problems _can_ be solved.   That does
not make me worry less, nor does it mean that I think the
problems are unimportant.  But, until and unless I think I have
new arguments or evidence, I don't bring them up again, and I
would _expect_ James or Marc to tell me that I'm out of order if
I did.  I also don't try to convince the WG that those topics
are important by getting many people --who have not been
participating in the WG's work-- to send identical notes
complaining, nor do I circulate notes to other groups in the
hope that they will somehow overrule the WG.  I avoid doing
those things for two reasons: they are considered in very bad
taste in the IETF and they never convince anyone; instead, they
tend to irritate people enough that they stop listening.

What is strange, to me, about this discussion is that there is a
way to delay CDN deployment without impacting Korean or
Japanese.  It is fairly simple, and does not require that the
IDN WG agree. And I had assumed it would be obvious to everyone.
Specifically, the four NICs simply refuse to register such
names.  You try to convince Singapore and other countries with
large Chinese-speaking populations to join them in that
prohibition.  You expand on Erin's document, some of Prof
Tseng's notes, and other comments and examples which have been
introduced here to provide other registries cautions about the
problems they could cause by registering such names.  And you
try to convince ICANN to prohibit those names in gTLDs.   From
the perspective of this WG, that is a policy move, made in
policy forums.  The fact that the WG provides a _mechanism_ for
encoding CDNs does not imply that any registry need _permit_
CDNs.  And it seems to me that, at this late date and given the
overlapping Japanese and Korean issues, that may be your most
efficient way forward on this particular problem.

regards,
     john