W3C home > Mailing lists > Public > www-archive@w3.org > April 2004

Re: Named Graph Homepage - second draft

From: Patrick Stickler <patrick.stickler@nokia.com>
Date: Tue, 13 Apr 2004 09:06:29 +0300
Message-Id: <B43ED041-8D10-11D8-8853-000A95EAFCEA@nokia.com>
Cc: www-archive@w3.org, ext Chris Bizer <chris@bizer.de>, Pat Hayes <phayes@ihmc.us>
To: "ext Jeremy Carroll" <jjc@hplb.hpl.hp.com>

On Apr 08, 2004, at 12:45, ext Jeremy Carroll wrote:

> I agree - this is the correct practical algorithm, I'll see if I can 
> get some text in there.
> i.e. the URI is the base URI in scope on the outermost RDF element of 
> the document (typically the rdf:RDF of an RDF/XML or OWL document). 
> This may be the xml:base or the retrieval URI of the document.

Right. That's a better way to put it. And if there is no such base URI 
in scope for the root rdf:RDF element, then the graph is not named.


> Jeremy
> Patrick Stickler wrote:
>> On Apr 07, 2004, at 17:10, ext Chris Bizer wrote:
>>> Hi,
>>> I updated the draft for the Named Graph homepage and included the 
>>> comments
>>> from Patrick and Jeremy. I also added RDF/XML as a third possible 
>>> syntax for
>>> Named Graphs to section 3:
>>> 3.3 RDF/XML
>>> A collection of RDF documents can be seen as a set of Named Graphs. 
>>> This
>>> gives Named Graphs upward compatibility with RDF/XML, but has the
>>> disadvantage that retrieval URL, document name and graph name are 
>>> mixed up.
>> I've been chewing on this a bit more recently, and I still think that 
>> it
>> makes alot more sense to derive the name of an RDF/XML encoded graph 
>> based
>> on the xml:base value *of the root <rdf:RDF> element*.
>> Yes, it is true that any element in the RDF/XML can have its own 
>> xml:base
>> attribute defined, but there can be at most one such attribute defined
>> for the <rdf:RDF> element, so there's no ambiguity there.
>> And since the xml:base value need have no correlation to the URI via 
>> which
>> the RDF/XML instance was obtained, we avoid the URI denotation 
>> ambiguity
>> otherwise introduced by taking the access URI as denoting the graph.
>> Otherwise, I'd prefer to simply state that there is no obviously 
>> correct
>> and reliable means to associate a graph name URI with an RDF/XML 
>> instance
>> in the instance itself, and avoid (being misunderstood) proposing that
>> the access URI be used (which I think is a mistake/hack/etc.).
>> Thus,
>> <rdf:RDF xml:base="http://example.org/foo" ...>
>>    ...
>>    <rdf:Description xml:base="http://example.org/bar" ...>
>>    ...
>> </rdf:RDF>
>> equates to
>> <http://example.org/foo> {
>>    ...
>> }
>> Note that the second xml:base on the description element has
>> no affect on the name of the graph.
>> Also, this works even when mulitiple RDF/XML fragments are
>> embedded in the same e.g. XHTML document, since each root
>> <rdf:RDF> element can have its own xml:base value and hence
>> a distinct name.
>> We'd restrict the interpretation of xml:base to explicit
>> attributes occurring on the root <rdf:RDF> element, not
>> inherited from a higher XML scope.
>> Eh?
>> Patrick
>>> Chris
>>> <NamedGraphsPage.zip>
>> -- 
>> Patrick Stickler
>> Nokia, Finland
>> patrick.stickler@nokia.com


Patrick Stickler
Nokia, Finland
Received on Tuesday, 13 April 2004 02:08:40 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:32:27 UTC