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

rdfms-uri-substructure: some clarifying advice [was: A collection of issue resolutions]

From: Dan Connolly <connolly@w3.org>
Date: 12 Feb 2002 15:35:35 -0600
To: Brian McBride <bwm@hplb.hpl.hp.com>
Cc: RDF Core <w3c-rdfcore-wg@w3.org>
Message-Id: <1013549736.13109.121.camel@dirk>
On Tue, 2002-02-12 at 13:05, Brian McBride wrote:
> Propose the WG resolve that the proposed change would be a major change to 
> the RDF specification and is out of scope for this WG.
> 
> rdfms-uri-substructure: xmlns, uri+name pairs or just uris..? Clarification 
> needed (Sergey Melnik)
> 
> 
> Propose the WG resolves that changing how resources are named on the web is 
> a web architecture issue and beyond the scope of our charter.

I can live with that, but I have a proposal that I prefer:

It's just clarifying advice to users/implementors, not a change to the
language:


Whereas:

(a) the RDF 1.0 spec says that property and class names
are computed from element and attribute names
by concatenating their namespace names with their local names

(b) it's useuful to be able to process RDF with
XPath and XSLT, where even though
	concat(namespace-name(qname1), local-name(qname1))
is the same as
	concat(namespace-name(qname2), local-name(qname2))
the qnames themselves may not compare equal in XPath expressions.

(c) lots of implementors have looked for advice on
how to serialize RDF, and, in particular, how to
compute a namespace name and localname from
the name of a property or a class.

we advise RDF schema/namespace/vocabulary designers

(d) choose namespace names that end in non-xml-name-characters
such as / # ?

and we advise implementors of RDF serializers:

(e) in order to break a URI into a namespacename and a local
name, split it at the last XML name character. If the URI
ends in a non-name-character, throw a "this graph
cannot be serialized in RDF 1.0" exception.

Provided this is agreed, the syntax editor should salt to taste
and add it to the syntax spec, and the test cases editor
should add something close to what follows:

Test cases:

implementors are advised to throw a "cannot serialize in RDF 1.0"
exception if asked to serialize:

  <mid:something@example> <http://example/ill-advised-name/> "abc".

and implementors are advised to serialize

  <mid:something@example> <http://example/ok-name/> "abc".

as, for example:

<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">

    <rdf:Description rdf:about="mid:something@example">
        <ok-name xmlns="http://example/">abc</ok-name>
    </rdf:Description>
</rdf:RDF>

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Tuesday, 12 February 2002 16:36:17 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:45:09 EDT