RE: UTF-8 in Internet Drafts (was Re: [idn] I-DACTION:draft-jsen g-idn-admin-01.txt)

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

--On Thursday, 31 October, 2002 08:09 -0500 "Hollenbeck,
Scott" <shollenbeck@verisign.com> wrote:

>> I have to fully agree with Mark. Getting the policy changed
>> would be great. For James' document, the us-ascii version
>> is just a bunch of numbers. No help to those who
>> don't know Chinese characters, and no help to those who do.
>> Just a waste of bits and paper, even if the topic may be
>> important and interesting.
> 
> I-Ds can be submitted in PostScript format if that helps,
> though a text version (that can be different due to issues
> with figures and non-ASCII characters) is still required to
> have both published.

Assuming, obviously incorrectly, that the same "simultaneous
posting in text and PS or PDF" principle applied to I-Ds as
well as RFCs, this one was sent to the I-D administrator in
text and pdf formats.  The latter was discarded.   If we would
like a PDF alternative, I think that is a small matter of the
IESG expanding the rules.   

On the other hand, "post in UTF-8 as an _alternative_ to text
strikes me as a bad idea: realistically, few of us have UTF-8
capable display systems that can display any Unicode
character. Citing Paul's examples, I can personally see
Patrik's or Martin's names displayed correctly, but the
Chinese characters in "draft-jseng-idn-admin..." turn into
either black blotches (losing information) or back into the
Unicode code points (which are present already).   James and
the rest of the JET team have been discussing (at my request)
ways to present the characters in ASCII text that would make
the document more comprehensible than the code points alone
for those reading the document in ASCII text.  A URL for the
PDF has been posted, and the next version will presumably be
submitted in Postscript as well (and will probably contain a
pointer to the PDF version).   But, speaking for myself, I
would far prefer clarity and unambiguousness in an I-D to a
religious principle. At this stage, "permit I-Ds to be
submitted in UTF-8" is, for too large a percentage of the
community, too close to a religious principle at the expense
of clarity.

And, while this document was copied to the IDN list for
information, neither the document nor its format are task
items for this WG.

     john