W3C home > Mailing lists > Public > public-grddl-wg@w3.org > August 2007

Re: Comments on GRDDL (using 3rd-party XML schemas with GRDDL) [OK?]

From: Harry Halpin <hhalpin@ibiblio.org>
Date: Wed, 01 Aug 2007 10:38:56 -0400
Message-ID: <46B09B00.1000503@ibiblio.org>
Cc: GRDDL Working Group <public-grddl-wg@w3.org>

Chimezie Ogbuji wrote:
> On Fri, 2007-07-27 at 09:46 -0500, Dan Connolly wrote:
>   
>> It's inconsistent.
>>     
>
> Yes
>
>   
>> Any statement of the form "A GRDDL-aware agent may ..." specifies
>> conforming behavior. To then say that such behavior is not conforming
>> is inconsistent.
>>     
>
> Yes, see my previous email about the value of additional text regarding
> capabilities not required by a GRDDL-aware agent.  
>
>   
>> Rather than "allowing
>> transformations to be found in schemas not specified at 
>> the namespace document" it would make more sense to say
>> "allowing namespace documents to be looked up using
>> non-standard mechanism"; e.g. somebody could use
>> a local/custom URN resolver or a catalog or whatever to
>> overlay the public web with a local mapping of URIs
>> to schemas. 
>>     
>
> This makes *much* more sense IMHO.  The mechanism is not out-of-band,
> (it is very much within the boundaries of AWWW - albeit via a
> non-ubiquitous URI scheme), it doesn't speak specifically about schemas
> (of which GRDDL knows *nothing* and should remain this way).  I would
> strike out the last statement about schemas.  
>   
The XSL/XQuery WG really wants XML Schema to be mentioned as an example
of possible out-of-band information. In my personal opinion, as an
example of possible out-of-band information, I think it's pretty good,
and in general giving a concrete example of some non-normative behavior
that GRDDL that some people might think is reasonable but that GRDDL
explicitly does *not* support and is *not* using in the test-cases helps
make the point about under what precise conditions the test-cases should
be run.

Now, let's make sure we're discussing the *exact* edits the XSL/XQuery
WG suggested:

"The GRDDL specification states that any transformation identified by an
author of a GRDDL source document will provide a Faithful Rendition
<http://www.w3.org/TR/grddl/#sec_rend> of the information expressed in
the source document. The specification also grants a GRDDL-aware agent
the license <http://www.w3.org/TR/2007/WD-grddl-20070302/#sec_agt> to
make[REMOVE 's'] a determination of whether or not to apply a particular
transformation guided by user interaction, a local security policy, or
the agent's capabilities. [For example, a GRDDL-aware agent may have a
security policy that prevents it from accessing GRDDL transformations
located in untrusted domain names or it may be unable to apply
transformations given in a language it does not support, and so it may
be unable to produce the faithful rendition. Furthermore, in addition to
being GRDDL-aware, an agent may feature optional capabilities such as
allowing a schema and an associated transformation not at the namespace
URI to be looked up using the mechanisms defined in the W3C XML Schema
specification [XMLSCHEMA], and the results of applying such a
transformation may not be a faithful rendition.]  In defining these
tests it was assumed that the GRDDL-aware agent being tested is using a
security policy which does *not* prevent it from applying
transformations identified in each test [, supports XSLT 1.0, and does
not rely on any capabilities outside those defined in the GRDDL
Specification]. Such an agent should produce the GRDDL result associated
with each normative test, except as specified immediately below."



> I think we can satisfy the XQuery WG's concerns with statements which
> demonstrate that HTTP dereference of a namespace URI is not the *only*
> way to resolve additional content to consider in calculating a derived
> RDF graph (even though this is the only normative mechanism which
> results in a 'faithful rendition') without saying anything about XML
> schemas.  
>   
Again, I think they want XML Schemas to be explicitly mentioned as a
possible example and will be unhappy with anything else.
> This is very similar to something I touched on in a recent presentation
> [1] where I investigated how GRDDL can be extended to support "faithful
> renditions" via transformations identified by mechanism in a CMS
> independent of direct markup in the source.  
>   
> The suggested solution was to use "GRDDL-like" mechanisms where the
> application of a transformation is the same as in GRDDL, but the means
> by which the transformation is identified is outside of GRDDL.  This
> does not require a modification to GRDDL but an understanding of how it
> can be used as the base framework for additional (in this case, more
> specific) behavior.
>   
Note the exact terminology used in the suggested change is "Furthermore,
in addition to being GRDDL-aware, an agent may feature optional
capabilities" which I think covers the fact there are many non-normative
types of possible behavior *and* that these are explicitly not part of
being GRDDL-aware.
> I think the comment thread is conflating identification of schema
> documents with calculation of faithful renditions (the two have
> *nothing* to do with each other).
>
>   
>> Then the result is a faithful rendition
>> inasmuch as the author of the source document agrees
>> that the non-standard lookup mechanism gives a
>> reasonable representation of the namespace document.
>>     
>
> I'm not certain it would be a 'faithful rendition' as it the resulting
> RDF would have been computed through transformations identified in a
> manner outside of GRDDL.  However, this doesn't mean that GRDDL-like
> computation of RDF graphs are not useful or cannot be done.
>   
Again, that's why the suggested text features the sentence "such as
allowing a schema and an associated transformation not at the namespace
URI to be looked up using the mechanisms defined in the W3C XML Schema
specification [XMLSCHEMA], and the results of applying such a
transformation may not be a faithful rendition."

The "not" is important there, as it explains that non-standard
mechanisms may not lead to faithful renditions.
> [1] http://copia.ogbuji.net/files/stc07/GRDDL-XML-CMS.odp
>
>   


-- 
		-harry

Harry Halpin,  University of Edinburgh 
http://www.ibiblio.org/hhalpin 6B522426
Received on Wednesday, 1 August 2007 14:39:07 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:11:50 GMT