Re: FXSL and 'saxon:function()'

On 11/25/05, drkm <darkman_spam@xxxxxxxx> wrote:
>  Hi
>
>  I first heard about FXSL a fee weeks ago on XSL-List.  I
> seen there was a Saxon-specific implementation, that is now
> removed from the release manager.  Is this a choice?

You probably confuse here FXSL 1.x (developed for XSLT 1.0) with FXSL
2.x (developed for XSLT 2.0).

In the early days of FXSL 1.x development it was necessary to have
different XSLT-Processor-specific versions only because the
xx:node-set() extension function has different name/namespace in any
gicen XSLT processor. A later version of FXSL 1.x was developed to be
XSLT-Processor independent by using the EXSLT function
common:node-set(), which most of the major XSLT processors implement.

In XSLT 2.0 the RTF "datatype" was eliminated completely, which
allowed FXSL to be written without needing any extension functions.

Therefore, there is absolutely no need and no meaning to create
separate versions of FXSL, specific to given XSLT 2.0 processors.

>
>  I just see Saxon has native HOF.  Is would be intresting
> to have a Saxon-specific FXSL that uses these native HOF?
> So the intresting pqrt of FXSL will not be qnymore to
> provide HOF, but in the functions library it provides.
>
>  Does this make sense?
>

No -- see the explanation above.

And thanks to Michael Kay for explaining the main purpose and use of
saxon:function().

--
Cheers,
Dimitre Novatchev
---------------------------------------
To avoid situations in which you might make mistakes may be the
biggest mistake of all.