Re: [idn] WG last call summary

["D. J. Bernstein" <djb@cr.yp.to>]

Once again: IDNA has received strong written objections from at least
fifteen regular WG participants and _hundreds_ of other people.

Here are some typical quotes from IDNA proponents explicitly refusing to
take these objections into account:

   * ``Just protesting doesn't count, if an alternative or fix isn't
     included'';

   * ``the Chair's responsibility ... is to move work along ...
     discouraging discussion of problems ... for which realistic
     solutions ... have not been proposed'';

   * ``unless you have a TC/SC solution which you willing to contributed
     to the group, I consider this discussion closed'';

   * ``You have only repeated problems that we already knows. You have
     not demonstrated any solution which is technical possible now.''

These responses are all missing the point. When a user objects to IDNA,
saying--for example---that IDNA will produce ``conflicts and chaos for
Internet users of Han characters,'' you can't dismiss his objection by
saying that you believe that the other proposals are even worse.

As I commented before, the IETF procedures don't say ``It's okay to make
an incredibly destructive modification to the Internet protocol suite if
you have to _do something_.'' Until the IDN WG settles on a safe course
of action, we will have to stick to the status quo.

I also summarized why people are objecting to IDNA: ``IDNA will cause a
tremendous amount of damage, including bounced email, web link failures,
widespread user confusion, and massive costs---much higher than
necessary---for software development and deployment.'' Crocker asserts
that ``such false claims have been dealt with repeatedly.'' Let's go
back to the videotape:

   * IDNA co-author Adam Costello claimed in an IDN message on Sun, 27
     May 2001 21:30:52 +0000 that, under IDNA, ``nothing will actually
     break (mail will get through, web pages will load, etc).''

   * After the IDN WG identified several serious interoperability
     problems in the IDNA architecture---the result being that mail
     would bounce, web links would fail, etc.---I challenged Costello's
     ``nothing will actually break'' claim.

   * In an IDN message on Thu, 19 Jul 2001 04:31:48 +0000, Costello
     admitted that IDNA _would_ break things, and that his previous
     claim was wrong: ``I overstated it. I was wrong. Sorry.''
     (Naturally, he continued by saying that even more things would be
     broken by another proposal.)

I recently asked a simple question about how IDNA is supposed to work,
from a programmer's perspective: under UNIX, if LANG is en_US.UTF-8,
should the MH/NMH ``show'' mail-displaying program convert names from
the IDNA character set to UTF-8? Costello, aware that a ``yes'' answer
would cause interoperability problems and that a ``no'' answer would
mean that users see gobbledygook instead of non-ASCII glyphs, ignored
the question.

---D. J. Bernstein, Associate Professor, Department of Mathematics,
Statistics, and Computer Science, University of Illinois at Chicago