Re: [xslt] XSLT 2.0

[Daniel Veillard <veillard redhat com>]

On Wed, Jun 18, 2008 at 09:51:08AM +0100, Tony Graham wrote:
> (Sorry to take a while to respond, but it's been busy here.)
> 
> On Tue, Jun 10 2008 01:31:05 +0100, Steve Ball explain com au wrote:
> > On 06/06/2008, at 7:18 AM, Tony Graham wrote:
> >> On Thu, Jun 05 2008 07:57:07 +0100, Steve Ball explain com au wrote:
> ...
> >> There probably are enough libxslt users that there would be quite a
> >> few
> >> of those users who could gainfully contribute, but that raises the
> >> question of why there aren't more than the dedicated few currently
> >> working on libxslt.
> >
> > That's the same with many (most?) open source projects...
> 
> Yes.  I have a semi-coherent theory about a log relationship between
> numbers of users and interested users and between them and people who
> can provide code.

  I would say that correlates my experience. There is also a net trend
factor, i.e. if what you're working on is related to current hype in 
media you're suddenly more likely to find people with 'free time'.

> >> I would be willing, but I have other commitments on my open source
> >> free
> >> time.
> >
> > Don't we all... ;-)

  of course.

> So how do we proceed?  It's not going to get very far if it's just you
> and me moaning about not having enough free time.  I'm sure a poll of
> people on this list (or of people you meet on the street) would show
> that most people feel they don't have enough free time.
> 
> It seems to me that it's still a bit premature to say "let's just make a
> wiki and thrash out the design", even if someone were to stump up for
> some non-free time.
> 
> Daniel, if I may ask, in an ideal world, how would you like to see an
> XSLT 2.0 version of libxslt be developed?

  Ouch, hard question. In an ideal world you get a person with passion
for the technology, who understand the technical bits and a sponsor
willing to pay him full time to implement it, until it's finished.
  I'm not sure we should plan for an ideal world. I'm personally not that
thrilled by XSLT-2.0/XPath-2.0 (probably because I don't use XSLT much),
and clearly my employer has no plan to push me to implement it (libvirt
takes most of my time nowadays).
  Maybe a progressive approach can be done, first a update of XPath in 
libxml2 to grok XPath2, and then maybe a separate libxslt2 library on
top of it which could reuse libexslt and bits of libxslt.
  It's a bit hard for me to draft technical directions because I didn't
tried to learn the version 2 of XPath/XSLT/...
  IMHO it's way to big, I have also a givem theory linking the size of a
specification to the manpower needed to implement it and test it fully,
and in C the ratio is especially large (though the implementation is IMHO
more useful). 

Daniel

-- 
Red Hat Virtualization group http://redhat.com/virtualization/
Daniel Veillard      | virtualization library  http://libvirt.org/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine  http://rpmfind.net/