SYNTAX: Refactoring comments, schemas and status

This is a summary of some outcomes after publishing the refactoring
WD including comments, schema developments and other related documents.


We have received several comments on the refactoring doc via the
www-rdf-comments and myself some encouraging ones privately.  I also
posted a note to the xmlschema-dev list to prod them for suggestions.
The results of this are that we have two of the schema champions
(great phrase!) who have made new draft schemas for RDF/XML.

The latest versions are:

  Non-XML RelaxNG schema by James Clark, 18 Sep 2001
  http://lists.w3.org/Archives/Public/www-rdf-comments/2001JulSep/0248.html

    RelaxNG: http://www.thaiopensource.com/relaxng/nonxml/

  Schematron schema by Rick Jelliffe, 17 Sep 2001
  http://lists.w3.org/Archives/Public/www-rdf-comments/2001JulSep/0237.html

    Schematron: http://www.ascc.net/xml/resource/schematron/schematron.html

We need to get these tested properly to check they do match our
decisions.


Henry Thompson replied to my xmlschema-dev posting indicating a
suggestion on how to use XML Schema and XSLT to do the mapping.
  http://lists.w3.org/Archives/Public/xmlschema-dev/2001Sep/0051.html

This was quite similar to the suggestion from James Clark in his
message above:

  .. identify a minimal subset of the syntax that is suffificient to
  express any triple and can be mapped straightforwardly onto
  triples, specify that syntax using RELAX NG, and then specify how
  to transform the full syntax into this minimal subset.
  -- James Clark, 
     http://lists.w3.org/Archives/Public/www-rdf-comments/2001JulSep/0248.html

Jeremy has done a discussion doc using a more formal declarative
approach than the current procedural form, including a method of
mapping to triples.

  A Lexical Functional Grammar for RDF (1st version)
  Jeremy Carroll, 14 Sep 2001
  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Sep/att-0169/01-lfg-rdf.htm

  2nd version (PDF only): 17 Sep 2001:
  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Sep/att-0194/02-rdf_and_lfg.pdf
  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Sep/att-0194/01-rdf_and_lfg_A3.pdf

although Jeremy admits some things like rdf:_<n> and rdf:aboutEach
cannot be done this way.  I'm still trying to understand this.

I talked to Jeremy about mapping and suggested that it would be very
good idea to use deployed XML technologies if possible; especially
those which have tools that are very familiar to developers.

Last week he wrote a second discussion doc:

  Transforming the Abbreviated Syntax into Very Basic RDF
  Jeremy Carroll, 18 Sep 2001
  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Sep/att-0224/01-Transforming_RDF.html

The idea is related to what we discussed although I feel this precise
method might not be sufficent, since you can't express all legal
models in RDF/XML, or any true subset.  With N-Triples you can do
this, so if we do need a canonical form, in that case we are heading
towards an XML N-Triples (which I'm going to leap in and name XNT,
pronounced "xant").


Separately I've been looking at at the existing XML Schema for
RDF/XML that has been around for a while (by Rick Jelliffe, updated
by Ralph Swick), and trying without a lot of success to update it to
the W3C REC form.  The latest version is available at

  XML Schema for RDF/XML
  http://ilrt.org/discovery/2001/09/rdf-xml-schema/

Dave

Received on Thursday, 20 September 2001 12:33:28 UTC