Re: [idn] Planning ahead for good IDNs

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

--On Friday, 12 January, 2001 15:11 -0500 Marc Blanchet
<Marc.Blanchet@viagenie.qc.ca> wrote:

> At/À 19:52 2001-01-12 +0000, Brian W. Spolarich you wrote/vous
> écriviez:
>> On 12 Jan 2001, D. J. Bernstein wrote:
>> >> | There seems to be widespread agreement, even among ACE
>> proponents, | that we will _eventually_ move to UTF-8 domain
>> names.
>> >>   There is?  I can't say I myself detect a rough consensus
>>   based on the recent conversation in the WG, or in private
>> conversation.  There still seem to be at least three credible
>> camps:  ACE only, ACE transition, and UTF-8 only.
> > and the more "exotic" ( a friendly joke targetted to John K.):
> directory, new class, etc... which, because of their property
> can use or not encodings at all, more work to be done on that
> side tho.

I would add to Marc's list an observation made on this list
several times before (but apparently as frequently forgotten).
Since it was first introduced, there has been feeling in some
parts of the community that UTF-8 is discriminatory toward, or
relatively disadvantageous to, languages in the "upper" part of
the UCS-2 (and beyond) code point sets.  Whether or not that is
true depends on the metrics used.  However, for those who are
concerned about that issue, and dislike UTF-8 as a consequence,
most or all of the ACE schemes (and other schemes, including
UCS-4 itself, that produce codes that are length-independent of
code points) are going to be more acceptable.

Consequently, claims about "widespread" belief that UTF-8 is the
correct long-term solution require at least some qualification.

I think we have drawn closer together, but, at one time, the
level of enthusiasm for UTF-8 in the non-ASCII community could
be predicted by taking the inverse of the numerical position of
the last character in the script used by that community in the
Unicode/10646 code tables.

     john