Re: Associating a GRDDL transform with an XML schema... and it's not XSLT

Quick note - I'll be in the south of France and can go over this in
detail then, but as pointed out to e-mails to the OWL WG, a GRDDL
transform is supposed to be machine-readable transformation such as
something in XSLT, not a human-readable document. But in essence, I
think Kevin Smith has it right.

Thus, what you want to do is put a pointer to the XSLT in the Schema,
no? The *important* key is that the XSLT is "findable" from whatever the
POWDER namespace document is. Since the namespace document hosts the XML
Schema, that's where the GRDDL transform should be, and the instructions
(and example) given in the Recommendation should be correct.

I am quite overwhelmed right now, but will be more than happy to sit
down and talk this over, and even do whatever changes are needed, in
          -harry

Phil Archer wrote:
>
> Thanks Kevin,
>
> I did wonder whether the simplest/correct thing to do would be to put
> a reference to the POWDER-BASE to POWDER-S XSLT in wdrb.xsd. But...
> here's where the difference between GRDDL and XSLT comes into focus
> (or my understanding goes out of focus, one or the other ;-) ). The
> transformation from POWDER to POWDER-BASE is a defined transform -
> which is defined in the Formal doc. Actually, I guess it's an open
> question whether this is a GRDDL transform or not since the output is
> still XML and not RDF. The POWDER-BASE to POWDER-S transformation is
> defined in the same document -  it just so happens that you are (very
> helpfully) providing an XSLT to effect that transformation which does
> create RDF/OWL from the XML input.
>
> Hence, my query about whether it is actually beneficial to include a
> reference to a human-only processable GRDDL transform.
>
> As you say, however, unless we hear from the GRDDL experts otherwise,
> your plan seems sensible. With a bit of luck, there may be a GRDDLer
> lurking around the south of France next week, you never know...
>
> Phil.
>
> Smith, Kevin, (R&D) VF-Group wrote:
>> Hi Phil,
>>
>> The normative part is the XML Schema section under
>> http://www.w3.org/TR/grddl/#ns-bind . The section of the primer you
>> refer to (ref [3] in your mail) applies to XML instance documents (in
>> the example given, an XML instance of an MS Excel spreadsheet), and not
>> schemas.
>>
>> Since the GRDDL transform is from POWER-Base to POWDER, then any
>> GRDDL-processing instruction would reside in wdrb.xsd [1], the schema
>> for POWDER-Base.
>> This would be via inclusion of
>>             xmlns:data-view="http://www.w3.org/2003/g/data-view#"
>>  
>> data-view:transformation="http://www.w3.org/2003/g/embeddedRDF.xsl"
>>
>> ...attributes within the xs:schema element.
>>
>> The RDF within xs:appinfo is non-normative, but I'm happy to include it
>> that way as it's (a) easily done and (b) may assist GRDDL processors by
>> being in the same format as suggested in the spec.
>>
>> So unless there are any objections I can add the attributes above, and
>> the annotation RDF to the POWDER-Base schema.
>>
>> Cheers
>> Kevin
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> [1] http://www.w3.org/2007/05/powder/wdrb.xsd
>> -----Original Message-----
>> From: Phil Archer [mailto:parcher@fosi.org] Sent: 15 October 2008 12:17
>> To: public-grddl-wg@w3.org
>> Cc: Smith, Kevin, (R&D) VF-Group
>> Subject: Associating a GRDDL transform with an XML schema... and it's
>> not XSLT
>>
>> Hi,
>>
>> As some of you on this list area aware, POWDER associates a GRDDL
>> transform with its namespace, the result of which produces POWDER-S.
>> (POWDER is an XML dialect, POWDER-S is RDF/OWL with a twist).
>>
>> The transform is defined in our Formal Semantics document [1], currently
>> undergoing minor edits to take account of LC comments received. What I'm
>> working on right now is getting the right references in the right places
>> to make the GRDDL association explicit... and, sorry, I'm confused and
>> would greatly appreciate some help.
>>
>> We have a (rather complicated) schema at [2]. We have a GRDDL transform
>> defined as human-readable text at [1].
>>
>> Do we just add a couple of lines like this to the root element of the
>> schema?
>>
>> xmlns:data-view="http://www.w3.org/2003/g/data-view#"
>> data-view:transformation="http://www.w3.org/TR/powder-formal/"
>>
>> I ask because the Primer suggests this is enough [3] but the main
>> documentation [4] suggests we need to include a chunk of RDF as an
>> xsd:annotation.
>>
>> And given that the normative transform is defined in a human readable
>> document, not an XSLT, is this going to break something? Actually, we do
>> have an XSLT but a) it's not normative and b) it handles some, but not
>> all of the transform (see the thread starting at [5] if you care why).
>>
>> Any and all comments gratefully received.
>>
>> Thanks
>>
>> Phil.
>>
>> [1] http://www.w3.org/TR/2008/WD-powder-formal-20080815/
>> [2] http://www.w3.org/2007/05/powder/wdr.xsd
>> [3] http://www.w3.org/TR/grddl-primer/#spreadsheets-section
>> [4] http://www.w3.org/TR/grddl/#ns-bind
>> [5]
>> http://lists.w3.org/Archives/Public/public-powderwg/2008Sep/0034.html
>>
>> -- 
>> Phil Archer
>> w. http://www.fosi.org/people/philarcher/
>>
>
>

Received on Wednesday, 15 October 2008 23:18:30 UTC