"This case isn't the important one" (was Re: Visually confusable characters (8))

[Andrew Sullivan]

Hi,

I've trimmed the cc:s

On Sun, Aug 10, 2014 at 01:35:18PM -0700, Asmus Freytag wrote:

> I can't help the impression that this case is an ad-hoc elevation of
> a single arbitrary instance of what are otherwise several doze
> edge cases, some of which might quite well be worse by an
> order of magnitude.

More than one person has made that claim, and I thought I'd answered
it, but let me answer it again.

The problem here _is not_ this particular case.  The new character
seems rare enough that it would obviously be sufficient to produce the
advice, "Don't use this character or this other combining sequence
unless you're absolutely sure you need it," and then trust zone
administrators to do the right thing.

No, the problem here is that we have discovered rather late in the
game that the thing NFC is for _is not_ what we thought it was.  I'm
not interested in discussing exactly how it came to this, and I've
even less interested in exploring the nature of the error that IDNABIS
made when making that decision.  The question now is whether there is
anything that _would_ be useful for what IDNA is trying to do.  Asmus,
you just suggested that the answer might be "no".  That seems bad,
given not only IDNA, but also that the IETF is just about to use the
same theoretical framework for other protocols -- sometimes, protocols
that might permit transient and _ad hoc_ registration, where automatic
rules were even more important than for IDNA.  I therefore would like
us to focus on that larger issue, and not on the narrow one of this
particular character.

Best regards,

A

-- 
Andrew Sullivan
ajs at anvilwalrusden.com