Re: [xsl] dynamic template invocation

[ac <ac@xxxxxxxxxxxxx>]

Hi,

Metaprogramming is fine but apart from your proposition being non 
standard, it seems a bit like using a bulldozer to pick-up a dime, in 
this case.

HOR are great and I am looking forward to their standardization and wide 
availability.  MAPing functions is especially interesting.

I like Haskell and even F#, but they are not XSLT.  Templates are also a 
bit different from functions.

Setting the mode seems like a very simple and natural extension of the 
existing paradigm.

Thank you for your interesting suggestions.

Cheers,
ac


> XSLT itself is good base for metaprogramming, XSLT even now can
> perform self analyzing & self modification according to input xml,
> params or lookup at document xsd - see
> http://www.saxonica.com/documentation/extensions/functions/transform.html
>
> On Tue, Sep 8, 2009 at 7:30 AM, Dimitre Novatchev<dnovatchev@xxxxxxxxx> wrote:
>   
> > Dear ac,
> >
> > Higher Order Functions (HOF) are powerful enough for the solution of
> > any problem.
> >
> > One needs just a little bit of practice in functional programming
> > (using a suitable language as Haskell) in order to gain the necessary
> > skills.
> >
> > HOF's are in the process of being specified by the corresponding W3 WG
> >  as a standard feature in XQuery, XPath and XSLT.
> >
> > Cheers,
> > Dimitre Novatchev
> > ---------------------------------------
> > Truly great madness cannot be achieved without significant intelligence.
> > ---------------------------------------
> > To invent, you need a good imagination and a pile of junk
> > -------------------------------------
> > Never fight an inanimate object
> > -------------------------------------
> > You've achieved success in your field when you don't know whether what
> > you're doing is work or play
> > -------------------------------------
> > I enjoy the massacre of ads. This sentence will slaughter ads without
> > a messy bloodbath.
> >
> >
> >
> > On Mon, Sep 7, 2009 at 7:15 PM, ac<ac@xxxxxxxxxxxxx> wrote:
> >     
> > > Hi Dimitre,
> > >
> > > It is a nice solution with a clever and elegant <xsl:variable
> > > name="templates" select="document('')/xsl:template" />
> > >
> > > One can always pre-include the stylesheets into an integrated one, in a
> > > "pre-pass" to get access to the included/imported templates, which may not
> > > be so trivial especially for imported stylesheets.
> > >
> > > The context also needs to be parametrized, but mostly it does not seem to
> > > provide dynamic template mode selection which brings back the requirement to
> > > use call-template on templates that typically have both name and match
> > > attributes, when running various process modes on the same source nodes (eg.
> > > payroll trial, log, post, archive, purge), either with as xsl:choose or a
> > > saxon:call-template.
> > > There are different ways to get around the problem but none of them as
> > > elegant as <xsl:apply-templates mode="$mode"/> or variations thereof.
> > >
> > > I have encountered quite a few other "similar" use cases, as when processing
> > > music, for example, with analysis, scoring, loging, debugging, performing
> > > modes, and for image and video processing, as well as for animation, case
> > > management, process modeling, ...
> > >
> > > I was suggesting and wandering if others feel that it may be a natural and
> > > useful improvement for the standard.
> > >
> > > What do you think?
> > >
> > > Thank you.
> > >
> > > Cheers,
> > > ac
> > >
> > >
> > >       
> > > > Search about FXSL. This feature is there (in  a way that is allowed by
> > > > the XSLT 1.0 language) since 2002.
> > >
> > >
> > >       
> >
> > --