[
Lists Home |
Date Index |
Thread Index
]
As far as I can see, Charmod can be stepped around by including a note with
your software
"This application does not completely satisfy the requirements of the Character
Model for the WWW. Operations involving accented characters may
be unreliable."
or something similar. Charmod does not require that you support all
Unicode characters (that is a SHOULD not a MUST).
There may be very good reasons not to support accented characters,
such as the unavailability of small libraries or the wrong market share.
But don't blame the messager: Charmod shows that there is a problem
and gives an architecture for incrementally reducing the problem until
it goes away. The problem is not caused by Charmod, it is just
intrinsic to Unicode.
Of course, it would help if transcoding library vendors made it clear
which form of normalization they had used, too.
From: "Michael Kay" <michael.h.kay@ntlworld.com>
> On the contrary, charmod, as I read it, expressly says that users are
> required to ensure that all strings are normalized, and expressly
> prohibits anyone from providing text processing software to help users
> perform this task.
It is the basic law of interpreting specifications that when faced with a silly
interpretation and a sensible one, you pick the sensible one: "Don't panic,
Private Godfrey!" The idea that Charmod bans anyone from providing
text-processing software to help fix problems is so silly it is seems obstinate.
Cheers
Rick Jelliffe
|
