Andrew Sullivan ajs at
Tue Dec 1 15:00:35 CET 2009

On Tue, Dec 01, 2009 at 01:45:14AM +0000, Shawn Steele wrote:
> Assuming the rules are applied consistently, then they would be
> clearly understood by all.

Hold on.  I think that what you say there is _exactly_ what the
IDNA2008 work was chartered to address, because it's false.

A basic fact of human-computer interfaces is that the humans bring
with them a bunch of assumptions, and those assumptions vary somewhat
depending on thier cultural background.  In the case of IDNA, the
relevant cultural pieces involve their languages.

Nobody is trying to make domain name labels into language (well, maybe
Jefsey's disciples are, but leave that aside), but we have to be
realistic and understand that the utility and ease of a given set of
labels is going to be heavily conditioned by the socio-linguistic
background of the person.  But what that means, of course, is that the
rules _are not_ applied consistently, because different
socio-linguistic contexts have subtly different rules for what turn
out to be "the same character". 

If you now want to insist that rules have to be applied consistently
across all these contexts, then what you're really doing is rejecting
the basis for the IDNA2008 work.  That's ok, but that's quite a bit
different from quibbling over how particular characters are to be


Andrew Sullivan
ajs at
Shinkuro, Inc.

More information about the Idna-update mailing list