Lists Home |
Date Index |
- From: Dan Brickley <Daniel.Brickley@bristol.ac.uk>
- To: "'email@example.com'" <firstname.lastname@example.org>
- Date: Sun, 9 May 1999 18:50:17 +0100 (BST)
I'm hoping that someone with more XSL expertise might provide a little
advise here. I'm looking into feasibility of using XSL transformations
as an alternative mechanism for derriving an RDF directed labelled
graph from XML instance data. This could be either via a transforming
XML data into RDF 1.0 Syntax, or into some other format more trivially
converted into triples of (Resource subject, Property predicate, RDFnode
object) for pouring into an RDF-consuming application[eg. 2].
I've had variations of this working on my desktop using MS IE5.0, just
by hacking the XML-to-HTML demos to output RDF instead. What I'd like to
know now is how one might set about doing this in a principled fashion.
Suppose, for common scenario, that I have an application which can be
conceptually mapped into RDF nodes and arcs, but for one of several good
reasons I don't feel inclined to use the relatively verbose RDF Syntax
to make this explicit. A number of other options for RDF compatibility
come to mind...
1. write code that builds in knowledge about proper interpretation of my
markup to create a stream of RDF assertions(triples)
2. use an alternative graph-serialisation convention (eg. Andrew
3. interpret DTDs,DCDs, schemata etc to figure out which RDF triples are
implied by my data
4. Represent the mapping to RDF in an XSL style sheet
IMHO there's probably a role for all of these in various contexts. The
latter is the one I'm currently concerned with, as it raises prospect of
derriving a unified representation of a lot of deployed XML without
having to use the RDF syntax for all apps.
First question is: what convention, if any, might we use to make it
mechanically evident that an associated style sheet serves to transform
into machine-friendly RDF (as against, for example, human-friendly
thanks for any suggestions,
SiRPAC is the w3c parser for RDF syntax and has recently acquired a more
abstract design allowing for alternative sources of RDF assertions; RDF
Syntax 1.0 simply becomes one way for creating a stream of facts to write into
an RDF graph.
'Resource','Property','RDFnode' mentioned above show up as classes in
SiRPAC; the latter is a convenience superclass of the Literals and the
Resources, since an RDF property may have either as its value (we can
use Java's instanceof to check which we've got).
see also CVS inteface to SiRPAC files at
In passing -- does anyone have any thoughts on how this interface might
better be intergrated into the SAX filter way of doing things? ie. XML
plus interpretation rules in one end, RDF triples come out the other...
xml-dev: A list for W3C XML Developers. To post, mailto:email@example.com
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:firstname.lastname@example.org the following message;
To subscribe to the digests, mailto:email@example.com the following message;
List coordinator, Henry Rzepa (mailto:firstname.lastname@example.org)