Re: GRDDL in non-HTML XML - how, exactly?

On Thursday, March 24, 2005, 6:01:41 AM, Dan wrote:

DC> On Wed, 2005-03-23 at 20:33 +0100, Chris Lilley wrote:
DC> [...]
>> On Wednesday, March 23, 2005, 8:11:33 PM, Dan wrote:
>> DC>    xmlns:data-view="http://www.w3.org/2003/g/data-view#"
>> DC>   
>> DC>
>> data-view:transformation="http://www.w3.org/2003/g/embeddedRDF.xsl"
>> 
>> Ok so transformation, not interpreter.

DC> yes... (we changed names in 1.53 of 2004/12/07 23:19:58).

I request that, since the last actual public draft (ie, on the TR page)
is dated 13 Apr 2004 that you publish the current draft on the TR page
at your earliest convenience and give consideration to a regular three
month updating of the TR page with this work in future.


DC> I added several diagrams in 1.62.
DC>  http://www.w3.org/2004/01/rdxh/spec

DC> And in the process I thought of some better ways to explain
DC> things, but I haven't updated the text yet.

DC> Meanwhile, I wonder if the diagrams help?

They do. Another request, please use the HTML object element to give a
choice between SVG and PNG directly.


>> DC>  but you can disregard those diagnostics;
>> DC> the resulting ,svgm.rdf has the right answer. (I hope we'll
>> DC> upgrade the SVG namespace document to RDDL or the like soon.)
>> 
>> Yes, good idea.

DC> I have an example of using RDDL as a profile document...
DC> see the "transformation applied to profile" figure.

DC> It would work just as well as a namespace document.

Thanks, will look at that.

DC> I have a grokRDDL.xsl transformation...
DC>  http://www.w3.org/2003/12/rdf-in-xhtml-xslts/grokRDDL.xsl

DC> but that's for one of the newer versions of RDDL, which
DC> don't seem to have a critical mass of support, so I intend
DC> to revert.

I agree that the original RDDL 1.0 seems to be the one with the critical
mass of users.

>> DC> As Dom explained, an XSLT stylesheet is a representation of
>> DC> an algorithm.
>> 
>> An, yes - its not a list, or at least, only a very short one.

DC> I think I see how you were mislead... the attribute takes
DC> a list of URI refs; each one refers to an algorithm, which
DC> should be represented in XSLT. Gotta work on that part
DC> of the spec.

Aha. Ok. But you are glad that IRI didn't make space a legal unescaped
character, then :)

What is the expected processing if I provide, say, three URIs each of
which points to an XSLT? Do you execute the first one that resolves
correctly? The first one in a supported language? All of them?

What is the motivation for a list, here?

>>> Other representations may be used by prior agreement of all
>>> concerned parties.

Who are 'all concerned parties' in a Web context? Everyone who has
access rights to dereference the resource?

DC> [...]
>> DC> In the 2nd example, I see this bit of RDF:
>> 
>> DC>         <rdf:Description>
>> DC>           <!-- In case of a well-known Coordinate Reference System
>> DC>                an 'Identifier' is enough to describe the CRS -->
>> DC>           <crs:CoordinateReferenceSystem
>> DC> svg:transform="rotate(-90) scale(100, 100)">
>> DC>             <crs:Identifier>
>> 
>> DC> That just says "there is a coordinate reference system." It doesn't
>> DC> say what the relationship between that coordinate system and this
>> DC> SVG file is.

DC> Actually, I misread it... it's conventional to start properties
DC> with lower-case letters, but CoordinateReferenceSystem is in
DC> the syntactic position of a property there... so what it's saying
DC> is "there is some thing (not named here) that is related
DC> by the CoordinateReferenceSystem property to something whose
DC> crs:Identifier is...".

>> Suggestions on how to fix that?

DC> It could be as simple as changing
DC>     <rdf:Description>
DC> to
DC>     <rdf:Description rdf:about="">

DC> so that it would be saying "this document is related
DC> by the CoordinateReferenceSystem property to something whose
DC> crs:Identifier is...".

OK, that's easy to fix. So, a specified but blank value is different to
an unspecified value?

If I say rdf:about="foo" does it make it about just the element with ID
foo?

>>  Its supposed to say 'here is a
>> coordinate system' and 'here is a projection' and 'this svg file
>> represents a map in that coordinate system with that projection'. It
>> seems we are missing the RDFese, or the GRDDLese, to make the third
>> statement.

DC> Right. RDFese.

OK, good. Sounds easy for me to fix so far. Then these cases can be
added to your test case collection.

-- 
 Chris Lilley                    mailto:chris@w3.org
 Chair, W3C SVG Working Group
 W3C Graphics Activity Lead

Received on Friday, 25 March 2005 15:13:34 UTC