Re: Crazy idea

James Clark wrote:
> 
...
> 
> The trouble with that is that functions would only be able to return
> result tree fragments. I had been thinking of something like:
> 
> <xsl:variable name="reverse"
>  select="function($s)
>           (concat($s ? $reverse(substring($s,2)) : '',
> substring($s,1,1)))"/>
> 
> <xsl:value-of select="$reverse('able was I ere I saw Elba')"/>
> 
> Something like you suggest would be useful too (I would think of it as a
> way to call named templates from functions).

and you'd be able to call them from your real functions, which would add
value to both features. Nice. 

> 
> There are plenty of interesting possibilities for XSLT/XPath v2.
> 

Those two would make a really nice pair of enhancements, IMHO. 

I've adjusted to the idea of immutable variables, but still find the
lack of compound data structures frustrating. I was trying to think of a
sensible way of using immutable compound data structures when I realised
that in XSLT terms it  would simply be a matter of allowing result tree
fragments to be used as expression inputs. (I say "simply" without
having thought it through - would this be tricky to design / implement?)

Francis.


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