Re: xsl:lang (was Re: <sort lang="sv"/> in Saxon)

Jeni Tennison <jeni@xxxxxxxxxxxxxxxx> wrote:
> Well, from my reading of the F&O WD, it appears that everything will
> be Unicode-normalized. So if you concatenate two strings together,
> then you get a normalized concatenation, for example. There's also an
> explicit function for Unicode-normalizing a string in various ways.

I used to think that Unicode is the end of all character encoding related
headaches. :-/ :-\
Well, automatic normalization will face stiff opposition, except as
part of a collation process. Apart from the usual fears of unnecessary
normalizations eating too much performance, some people will have reasons
to work on a normalization other than the standard normalization. After
all, there is also a reason why space normalization has to be explicitely
invoked with xsl:strip-space or normalize-space().
Explicit normalization is welcomed and an advance.

> I'm not sure which type of normalization will be applied by default
> (W3C, presumably), nor whether normalization occurs before strings are
> used by functions as well as after (you'll have to excuse my
> innaccurate terminology, but I'm thinking about if you have an
> unnormalized string with a pair of characters that would be normalized
> into a single character, then are the indexes used the ones in the
> unnormalized string or the ones in the normalized string?)

Currently, nobody seems to care about the fine points. More
precisely: i'm not aware of anyone mentioning on the XSL list of
being trapped by such effects. Perhaps decisions may be postponed?

Regards
J.Pietschmann

 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list