Re: [xsl] Time for an exslt for 2.0?

[Dimitre Novatchev <dnovatchev@xxxxxxxxx>]

> If yes, are there enough known deficiencies in the 2.0 spec to justify
> starting a new EXSLT 2.0-like project or is too early at this point to
> tell?  I'm not suggesting I know that answer and you most certainly
> would have a much better feel for this than me.  Just curious as to
> what exactly the focus would be. Higher-order functions I'm assuming
> would be one area of focus?

HOF support is already implemented with FXSL for XSLT 2.0. All F & O
functions, operators and constructors (with very few exceptions) and
even some XSLT features have their FXSL identically named HO
functions.

In my experience there are only two major features lacking in XSLT 2.0
 / XPath 2.0:

     - nested sequences (this is entirely XPath 2.0 DM problem)

    - no easy and convenient way to serialize xsl:variables
(especially ones containing
      sequences) in XML. There is no standard function that returns
the type of the value
      contained in an xsl:variable -- this type must be stored as part
of the serialization
      process, otherwise the value of the xsl:variable will not be
properly de-serialised (the
      type will be lost).

I don't think an EXSLT-like activity can solve the first problem
above. The second can be solved with a little effort in pure XSLT 2.0
(I have implemented a type-of() function but have to extend it to work
with list-types)

EXSLT can definitely contribute to the standardisation of such
features as function memoisation and accessing environments. However,
its long inactivity has prompted some people to ask if EXSLT is dead.



Cheers,
Dimitre Novatchev.

>
> Of course theres still the issue of dynamic XPath evaluation unless I
> have missed something and it has recently been added? (wishful
> thinking :)
>
> Anything else?
>
> On 12 May 2005 12:03:49 +0100, Colin Paul Adams
> <colin@xxxxxxxxxxxxxxxxxx> wrote:
> > EXSLT seems to be dead. Both Dimitre and I have fairly recently
> > submitted proposals for new features (memo-functions, and environment
> > variables for system-property() (as suggested by Michael Kay)
> > respectively.
> > Both were agreed, yet nothing has appeared on the web site (despite
> > Dmitre chasing up on his proposal).
> > Maybe it's time for an equivalent for XSLT 2.0?
> > --
> > Colin Adams
> > Preston Lancashire
> >
> >
>
> --
> <M:D/>
>
> :: M. David Peterson ::
> XML & XML Transformations, C#, .NET, and Functional Languages Specialist