W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > June 2002


From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Date: Wed, 26 Jun 2002 12:19:18 +0100
To: "RDF Core" <w3c-rdfcore-wg@w3.org>
Message-ID: <JAEBJCLMIFLKLOJGMELDMECNCEAA.jjc@hplb.hpl.hp.com>

At the f2f one of the postponed issues that I wished to advocate reopening was:


In particular, given the amount of change we have made I think a small change to
rectify the bnode serialization problem should be considered.

However we were out of time, and so I have decided to raise this by e-mail.

This is primarily motivated for me by Jena development.
Within Jena we have repeated user requirement to enable round-tripping.

e.g. one of our summer students yesterday said words like: "we use N-triple as
our transport becuase bNodes are important."

Since we are not proposing N-triple as a standard, we do have some obligation to
make RDF/XML useable.

I propose that we should modify the RDF/XML syntax to permit bNode labels
"_:foo" wherever we currently permit URI refs.

This still leaves non-standard property names as non-serializable, but that is
much less of a problem in practice.


More detail:

5.23 Production URI-reference
An attribute ·string-value· interpreted as a URI reference defined in Uniform
Resource Identifiers (URI) [URIS] BNF production URI-reference.

uses of URI-reference should have the name changed to resource-reference and

Production resource-reference


Production URI-reference

as before

Production bNode-reference

An attribute ·string-value· matching "_:<NCName>". (where NCName is the
production from XML Namespaces). The string value gives a bNode identifier with
file scope as for NTriple.

by using two productions (the current URI reference
Received on Wednesday, 26 June 2002 07:19:21 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:24:13 UTC