Re: GRDDL and XML Schemas/vocabularies [was: how about a mix ...?]

On Fri, 2006-08-18 at 11:02 -0400, Chimezie Ogbuji wrote:
> 
> 
> 
> On Fri, 18 Aug 2006, Harry Halpin wrote:
> 
> > Our charter specifically mentions RDF/A. embedded RDF, XML Schema, and
> > microformats. And we need a fairly comprehensive set of examples and
> > transforms, but we can't guarantee or "standardize" our transforms as
> > that is out of scope. However, our examples and test suite should work,
> > and to the extent they work people using GRDDL  may adopt  our transforms.
> 
> This is one area of the charter that I've wanted to ask about for some 
> time.  With regards to XML Schema, the charter says:
> 
> "tutorial materials and use cases sufficient to bootstrap adoption of 
> GRDDL .. with XML Schemas"
> 
> It's not clear if that means (1) using GRDDL to glean RDF from an XML 
> schema (using a pre-written profile), (2) using some hueristic to 
> 'generate' a GRDDL profile for instance documents that conform to the XML schema, or 
> (3) Using GRDDL with 'stand-alone' XML vocabularies documented by 
> existing XML schemas (as opposed to microformats, which are embedded in XHTML specifically)?

It's (3).

I meant (3) when I wrote it. But even if we discard the author's intent
(which isn't really supposed to matter that much), what remains,
literally, is that the WG succeeds if it delivers tutorial materials
sufficient to bootstrap of any of those three mechanisms;
they're all reasonable interpretations. As you say, (2) is an
open research problem, and (1) is pretty uninteresting, so I encourage
the WG to pursue (3).

In particular I think the clinical records story in your message of
Thu, 17 Aug 2006 15:12:53 -0400 is a good example. At least: I hope
the records have a root namespace URI that has an associated
schema.

> I have some observations / concerns with each interpretation.  If it's 
> (1), I wonder how much semantic content you can glean from a dialect specific to expressing 
> structural (and datatype) constraints, what would be the target RDF 
> vocabulary, and what would you gain that you don't already have from 
> processors that understand the XML schema dialect?

Indeed, that's an academic excercise of little interest.

> If it's (2), I'm *very* interested and it is an unprecedented area we've 
> (the research department I work for) invested quite a bit of research & 
> development into, but the problem statement (as least as it currently is in the charter) is quite vague.

Yes, open research question. Let's not put that in our critical path.

> If it's (3), I'm assuming the idea is tutorial material and best practices 
> for writing GRDDL profiles (XSLT profiles, most likely) for 
> well-established vocabularies with well-established schemas (but not 
> neccessarily in an automated fashion as with the previous interpretation) 
> such as DocBook, MathML, or any of the other major XML dialects mentioned 
> in Tim Bray's very well written article [1] "Dont Invent XML Languages" or 
> a follow-up thread [2] with some rather relevant commentary by alot of 
> people that know much more about XML vocabulary modelling than I do.

Getting GRDDL integrated with well established schemas would be great...
For example, getting a GRDDL pointer in the Atom namespace document
is technically straightfoward but politically sensitive. I am
hopeful, but I don't want to be too pushy about it.

Danny mentions the possibility of GEORSS; I haven't looked into
that one; I thought it was RDF/XML to start with.

But the charter doesn't say "well-established schemas"; as long
as our tutorial materials are sufficient that a few people with
moderately interesting schemas get on board, I think we're succeeding.


> [1] http://www.tbray.org/ongoing/When/200x/2006/01/08/No-New-XML-Languages
> [2] http://copia.ogbuji.net/blog/2006-01-12/Learn_how_
> 
> I'm sorta trying to get my head around what the intent was for that 
> particular reference to XML schemas in the GRDDL charter, especially if it 
> represents a 'requirement' we must meet.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Friday, 18 August 2006 15:25:09 UTC