W3C home > Mailing lists > Public > www-rdf-interest@w3.org > October 2001

Re: a new way of thinking about RDF and RDF Schema

From: Brian McBride <bwm@hplb.hpl.hp.com>
Date: Sun, 21 Oct 2001 23:24:38 +0100
Message-ID: <3BD34B26.50102@hplb.hpl.hp.com>
To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
CC: www-rdf-interest@w3.org, simeon@research.bell-labs.com

Peter F. Patel-Schneider wrote:


>>Most RDF parsers are built on a standard XML parser.  This is true of SiRPAC, 
>>RDFFilter, ARP and Redland's parser.
> How can this be?  RDF is not XML. 

RDF/XML (to name the xml syntax for RDF, not the RDF datamodel) is a subset of 
XML.  It conforms to XML syntax.  Parsers deal with syntax.

> How do they handle parseType?  An RDF
> system that handles parseType has to get its hands on the raw bits, before
> an XML parser sees them.

That statement is not correct.  RDF parsers such as RDFFilter, ARP and SiRPAC 
are built on top of an XML parser.

Typically they operate in streaming mode, where the XML parser outputs a stream 
of SAX events which are interpreted by the RDF parser.

parseType="Literal" is something the RDFCore WG is, err ..., working on, right 
now.  Its not really clear exactly how to handle this.  Most parsers currently 
turn it into a Literal, but there is no agreement on exactly what Literal. 
Other suggestions include generating an RDF representation of the infoset 
representation of the embedded xml.

But whatever the outcome of that discussion, I'd expect RDF parsers to continue 
to be built on XML parsers.  Am I missing something?

> Nope.  They can map to the same ``resource''.  There is no particular harm,
> as long as you can't tell the difference.  :-)  (In fact this is a feature
> of a true model theory.  If you can't tell the difference (as in via
> entailment) then it doesn't matter.)

Back to school for me on model theory.

>>><rdf:Description rdf:type="a:b"/> is not valid RDF because a:b is not a
>>>literal, not (just) because QNames are not allowed as attrib values. 
>>This is a nit, but the rdf:type attribute is a special case in RDF.  The value 
>>of the attribute is the URI of a resource, not a literal.  You may wish to add 
>>that to your list of mistakes in RDF.
> Precisely.  Therefore RDF can't use the abbreviation you wanted.

You might like to have a look at the following para in M&S


Thanks for the lesson on model theory.  I'll get back to school tomorrow.

Received on Sunday, 21 October 2001 18:29:21 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:44:32 UTC