Re: ETA on XSLT for POWDER to POWDER-S

Hmmm... this is interesting.

The XSLT engine I'll use in my (Perl) POWDER Processor is XML::LibXSLT 
[1] - which, a bit of poking around, suggests only implements XSLT 1 
(I'm not 100% sure but it looks that way). It would be pretty easy to 
use Perl techniques to apply the Reg Ex transforms to a POWDER doc in 
Perl without using XSLT but, as Toby says, it's the transformation to 
POWDER-S where we need the heavy lifting.

Also, note that the XSLT is not normative, it's a (very) handy tool. 
Furthermore, there is no _requirement_ to use the template regular 
expressions - in situations where milliseconds count there will probably 
be better ways.

All of which leads me to conclude that an XSLT that goes from 
POWDER-BASE to POWDER-S just using XSLT 1 might actually be preferable 
to one that handles POWDER to POWDER-BASE to POWDER-S if that entails 
using XSLT 2. It's certainly workable and I for one will probably find 
it easier to implement.

Phil.


[1] http://search.cpan.org/~pajas/XML-LibXSLT-1.66/LibXSLT.pm

Toby A Inkster wrote:
> 
> On 3 Oct 2008, at 10:12, Smith, Kevin, ((R&D)) VF-Group wrote:
> 
>> Good question -  I had started in XSLT 1.0 but changed to XSLT 2.0 to
>> utilise the replace() function: the part I was having trouble with was
>> to ensure that a host, e.g.
> 
> As I understand it (and I may be wrong), there are three flavours of 
> POWDER:
> 
> * POWDER, the easy-to-author XML format.
> * POWDER-BASE, identical to above, but with all the matching stuff 
> compiled into regexps.
> * POWDER-S, the RDF version.
> 
> Going from POWDER to POWDER-BASE or to POWDER-S may be difficult without 
> the string processing ability of XSLT 2.0. But going from POWDER-BASE to 
> POWDER-S should be more achievable with XSLT 1.0, as the regular 
> expressions shouldn't need any changes.
> 
> Personably I'd be happy enough with a POWDER-BASE to POWDER-S 
> transformation in XSLT 1.0. I could hanndle POWDER to POWDER-BASE as a 
> pre-transformation step.
>

Received on Friday, 3 October 2008 11:49:33 UTC