Re: [xslt] Implementation of unparsed-text() (XSLT 2.0)?

[Frans Englich <frans englich telia com>]

On Wednesday 16 August 2006 12:28, Daniel Veillard wrote:
> On Wed, Aug 16, 2006 at 01:03:20PM +0200, Daniel Leidert wrote:
> > Hello,
> >
> > I wanted to ask, if there are any plans to implement the XSLT 2.0
> > function unparsed-text() [1]. It seems the only possibility to read in
> > the content of a text-file and assign it to a variable in a XSL
> > stylesheet. So what is the current policy regarding XSLT 2.0 functions?
> >
> > [1] http://www.w3.org/TR/xslt20/#function-unparsed-text
>
>   Since XSLT-2.0 canibalize the same namespace, this doesn't allow for
> a graceful transition.

It probably doesn't hurt to mentation it is Daniel's viewpoint that 
XSL-T "canibalize the same namespace". Versioning is a debated topic, and 
that a separate namespace should be used for extensions is one of many 
possible views. In other cases versioning is handled in other ways -- which 
XSL-T is an example of with its version attribute and fallback mechanisms.

Personally, I like XSL-T 2.0 in this regard. I would find it quite unpractical 
to use different namespaces depending on if the function is from XPath 1.0 or 
2.0, or even if it's from XSL-T 1.0 or 2.0. I do not immediately see the 
drawbacks of not using a second namespace(s).

> If XSLT-2.0 functions were in a separate namespace 
> adding them like any other extension would have been relatively clean,.
> but in the current situation I'm not too fond of adding partial support.

Yes, I think so too.

> Apparently 'http://www.w3.org/2005/xpath-functions' is available for XPath
> 2.0 function which is way cleaner, but for XSLT-2.0 functions there is
> no separation as far as I can tell, maybe I missed it though...

XSL-T 2.0's functions are in the same namespace as XPath 2.0. As I see it, 
this is by the same principle as 1.0 with the only difference that it is an 
explicit namespace instead of an empty namespace URI, or "none".


Cheers,

		Frans