W3C home > Mailing lists > Public > www-tag@w3.org > November 2002

Re: Another RDDL/RDF proposal

From: Brian McBride <bwm@hplb.hpl.hp.com>
Date: Wed, 13 Nov 2002 00:24:47 +0000
Message-Id: <5.1.0.14.0.20021112235639.026fcd08@0-mail-1.hpl.hp.com>
To: Dave Beckett <dave.beckett@bristol.ac.uk>, Paul Prescod <paul@prescod.net>
Cc: Tim Bray <tbray@textuality.com>, WWW-Tag <www-tag@w3.org>

Hi Folks,

Please forgive me butting in with a few short comments.

o firstly of the various options of RDDL/RDF I've seen, the one suggested 
by Paul in

   http://lists.w3.org/Archives/Public/www-tag/2002Nov/0046.html

  looks pretty clean from an RDF perspective.  Jonathan seemed to like it too.

o its a really good idea not to rely the URI you used to fetch a document 
to affect what it means.  I've been burnt by taking copies of things on my 
hard drive to work on them on a plane and finding when I should be 
processing uri's like http://www.w3.org/.../schema#, I'm processing instead 
file:c:\\temp\schema.rdf.  Very embarrassing.

o if you want to check out your RDF/XML syntax, the RDF validator is a 
useful tool.  It is not completely uptodate with the latest rdfcore 
decisions, but its not bad:

   http://www.w3.org/RDF/Validator/

There are other ways to check out your rdf/xml, but the validator has the 
advantage it does not breathe fire on folks who get it wrong.

o if you ever want to discuss approaches to representing stuff in rdf, in 
general we'd be glad to help.  Send an email to me pesonally or to 
www-rdf-comments@w3.org, or to the rdf interest list.

o the rdf schema link in the Tim's rddl spec 404'd on me when I tried it today.

o the language in tim's rddl doc isn't sometimes not quite right, e.g. it 
refers to rddl:nature being an attribute with a value which is a URI 
reference.  That's syntactic language.  In rdf terms, rddl:nature is 
property not an attribute.  Values of the property are resources, in this 
case.  Not being able to read the schema I couldn't tell whether you had 
defined a class of for the values that are permitted, i.e. whether there is 
a range constraint.  Another example is that rddl:purpose is defined to 
take a URI as a value.  Does that mean its an rdf datatype anyURI, i.e. the 
value really is the string of characters, or is it the resource named by 
the string of characters.  Best to be clear about such things.

o there seems to be some duplication of terms.  What is the difference 
between rddl:nature and rdf:type, which I saw used I think in some of 
Jonathan's examples.  Is rddl:nature a subProperty of rdf:type with a range 
constraint?  Also there is rddl:description - could you use the dublin core 
property for that?  Similarly rddl:title.

o there does seem to be a lot of confusion about regarding what rdf/xml can 
and cannot do.  There seemed to be a number of statements in this thread 
which didn't seem quite right.  RDFCore is close to completing a new set of 
specs.  Any chance you guys could give some of them a look over.  It would 
certainly be useful to us to know if we have managed to improve the 
presentation/explanation of RDF.  Feedback on how to improve them would be 
very timely.

Do let me know if there is any way I can help.

Brian




   
Received on Tuesday, 12 November 2002 19:24:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:12 GMT