Eszett

[Andrew Sullivan]

On Wed, Aug 13, 2014 at 11:41:29AM -0700, Markus Scherer wrote:
> One big reason why sharp s and sigma are problematic in IDNA2008 is that
> users have gotten used to domain names being case-insensitive

Sort of, and those are certainly not the only cases.  For instance,
ECOLE.CA and ecole.ca will match, but école.ca does not match ECOLE.CA
(which is unexpected for French-from-France readers) and ÉCOLE.CA
doesn't work at all.

There's no getting around this, however, because the case-folding
rules outside the ASCII range are far more complicated than "add
0x20", and therefore simple case folding doesn't help.

Worse, domain names are not case-insensitive, but case-insensitive
_for matching_.  DNS is still case-preserving.

This very pickle is in fact _why_ IDNA2008 said, "No upper case, and
do local mapping where it makes sense."  The idea was that the user
agent could know its locale and could do mapping that was sensible in
that environment.  Nobody else would be in the position to do it.  My
recollection is that this was well-hashed-out at the time.

> In addition, in German the use of ß is confusing even to natives, and has
> changed in the 1996 spelling reform, and ß is never used in de-CH, etc.

At the same time, there was in fact considerable noise about the
mapping away of ß; I don't think we can dismiss that either.

Best regards,

A

-- 
Andrew Sullivan
ajs at anvilwalrusden.com