Re: Unicode Public Review Issues update

[Patrik Fältström <paf@cisco.com>]

On tisdag, feb 4, 2003, at 17:35 Europe/Stockholm, RJ Atkinson wrote:

> Example is that the lunar new year is called "Te^'t",
> with both ^ and ' being above the "e".  And the question is just 
> whether
> the 3 different ways of composing that character are properly handled
> by normalisation.
It should (in theory), because the normalization consists of:

 - Decomposition of characters into as many codepoints as possible
 - Sorting of the diacritics (canonical ordering)
 - Recomposition to as few codepoints as possible

See 3.10 of the Unicode standard 3.0, pages 50-52, about canonical 
ordering.

Now, the rules for canonical ordering might be wrong for Vietnamese, 
and then that should be pointed out.

   paf