RE: Proposed changes to address issue-dbooth-3 (ambiguity)

> From: Jeremy Carroll [mailto:jjc@hpl.hp.com] 
> 
> Advocacy:
> (still noting that we do not have an HP position as yet ...)
> 
> It seems strange to propose a change that supposedly is motivated to 
> better control variability, 

No, it is motivated by the desire to give the transformation author the
*ability* to better control variability *if* the transformation author
so chooses.  This change does not attempt to reduce variability in cases
where the transformation author is not concerned about such variability.

> that, actually, for the transforms that we 
> have experience with, increases variability.
> 
> The heart of the contradiction lies with the #which-langs issue.
> 
> David claims reduced variability for transforms written in 
> perl or using XProc. I believe:
> a) we have no examples of such transforms

Correct.  AFAIK, all of the examples in the test suite to date use XSLT.

> b) we have decided that such transforms, while interesting, and ones 
> that we think may develop in the future, are not in-scope for this 
> version of GRDDL

No, the working group specifically decided that transformations written
in other languages *are* permitted:
http://www.w3.org/2006/08/30-grddl-wg-minutes#item06
[[
RESOLUTION: to address [#issue-whichlangs] as per the current draft
(1.83 2006/08/25 20:23:09). SHOULD support XSLT 1; MAY support others.
]]

And the GRDDL spec in sec. 6 
http://www.w3.org/2004/01/rdxh/spec#txforms
specifically says that XProc "merits consideration for expressing more
complex or sophisticated transformations which require control over the
flow of processing through a variety of XML processing tools".  The goal
of this change is to *enable* the transformation to exercise such
control if it so chooses.  Without this change a transformation written
in XProc would be *unable* to prevent unwanted pre-processing.

> 
> I have demonstrated increased variability for transforms 
> written in XSLT 1.0.

Yes, and as I pointed out, that document was designed with GRDDL in mind
and *chose* to put a GRDDL transformation inside a default attribute.
This proposed change does *not* attempt to prevent people from either
shooting themselves in the foot or from producing variable results if
that is what they wish to do.  


David Booth, Ph.D.
HP Software
+1 617 629 8881 office  |  dbooth@hp.com
http://www.hp.com/go/software

Opinions expressed herein are those of the author and do not represent
the official views of HP unless explicitly stated otherwise.

Received on Tuesday, 19 June 2007 14:02:34 UTC