Collations used for sorting vary depending on what XSLT product you are
using, and what "locale" the system thinks you are in (Denmark or Sweden
perhaps?).

American software vendors have some strange ideas about the collating
sequences used in various European countries. As far as I can tell they are
often based on 19th century practice recommended by fussy national
academies, not on the actual practice of modern publishers.

Saxon 8.x allows you to select any collation supported by the Java VM, or it
you are really keen, to define your own.

Michael Kay
http://www.saxonica.com/


> -----Original Message-----
> From: Jon Breten [mailto:valemon@xxxxxxxxx]
> Sent: 01 February 2005 09:40
> To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> Subject: [xsl] Issue with xsl:sort - does not differ between
> 'V' and ''W'?
>
> Hi all,
> I am trying to sort a node-set by an element that contains lastnames -
> but it seems the sort algorithm does not differ between 'V' and 'W'.
>
> The code is simple enough:
> <xsl:sort select="LAST_NAME" data-type="text" lang="no" />
> I have tried both with and without the lang="no" property.
>
> The result is a list like the following:
> "Wabbah"
> "Vebah"
> "Weccer"
> "Vubbih"
>
> So, the point is it seems the algorithm regards 'V' and 'W' as the
> same thing, but correctly sorts on the following characters. Any ideas
> or input on this? I have checked around, and I cant find anything
> implying that this is by design - so I'm left wondering what I'm doing
> wrong here? The output is apparently sorted correct with regard to
> nodes not starting with V/W.
>
> Regards,
> Jon