J-F C. Morfin
jfc at morfin.org
Wed Apr 27 23:43:02 CEST 2011
At 20:58 27/04/2011, John C Klensin wrote:
>That is at least partially because, if someone wants to register a
>label that would naturally include a DISALLOWED character, people
>will make compromises and register whatever they consider as
>similar as possible.
this is what, I believe, makes Stephane differentiating characters
and code points.
Characters are what people consider. Code points and their properties
what Unicode considers. Characters exist before Unicode code points.
Unicode changes at this stage are to better match characters. The way
people have used code points to support a character may be the same
or different from the way Unicode will do it in its new version: this
is something a registry can easily address through a policy statement
and in validating domain names using the new Unicode approach to
replace the people compromise.
Also, remember that if IDNA2008 is Unicode-version independent, it
means that it is necessarily UCS system independent. I consider that
Unicode is inappropriate to network semiotics support but that
IDNA2008 does a good job to support Unicode and pave the way for new
kind of supports.
This means that U-labels and A-labels may go for ever, but N-labels
(from an UCS network oriented system?) could be introduced and
supported [that internally may very well use ISO 10646 tables or
not]. At this stage, IDNA2008, on the machine side uses Unicode which
is a computer oriented (hence code point) system, while people use
signs (i.e. when scripting, characters). The whole issue is to make
them better and better correspond. However, IMHO we are reaching the
limits of the Unicode kind of typographic oriented system (ex. lack
of French majuscules support, but also logos, gestures, sounds,
thoughts, etc. new signs for new kind of communication system. etc.)
More information about the Idna-update