Re: xsl:sort with msxml english language, danish characters, weird res

Michael Kay wrote:

> The UCA is written as if there is a single correct
> answer, but there isn't.
The UCA doesn't define a particular collation sequence for any 
languages, rather it defines the requirements for how collation 
mechanisms should allow you to define the collation rules for a given 
language and script. The Unicode standard is very clear that collation 
is highly variable and that there is no single answer for any language 
or script. [Even for a single language you might have different 
collation rules for glossaries and indexes, for example.]
Java's built-in RuleBasedCollator class implements a collation mechanism 
that, as far as I know, conforms to the Unicode UCA in that it provides 
the functionality needed (althought it may not fully address issues of 
how to handle composed and uncomposed characters--I'm not sure about the 
details there). The IBM ICU package provides a more complete 
implementation of the UCA and the ICU4J package provides an alternative 
set of built-in language-specific collators that are more complete and 
accurate than those shipped with Java.
Cheers,

E.
--
W. Eliot Kimber
Professional Services
Innodata Isogen
9390 Research Blvd, #410
Austin, TX 78759
(512) 372-8122

eliot(_at_)innodata-isogen(_dot_)com
www.innodata-isogen.com