Re: RDF document?

On Thu, 2005-06-09 at 09:30 +0800, Jeremy Wong wrote:
> Thanks for the correction of the term "document". It is an important concept 
> of RDF.
> 
> Consider the property foaf:workplaceHomepage [1], the domain of this 
> property is the class foaf:Document. Let's see the example..
> 
> <foaf:Person>
>  <foaf:name>Dan Brickley</foaf:name>
>  <foaf:workplaceHomepage rdf:resource="http://www.w3.org/"/>
> </foaf:Person>
> 
> I always think that this example is misleading. I would prefer to put the 
> above example in the following way..
> 
> <foaf:Person>
> <foaf:name>Dan Brickley</foaf:name>
> <foaf:workplaceHomepage 
> rdf:datatype="http://www.w3.org/2001/XMLSchema#anyURI">http://www.w3.org/</foaf:workplaceHomepage>
> </foaf:Person>

Thanks for the feedback. I don't see it as misleading at all:
foaf:workplaceHomepage is a relationship between a foaf:Person and 
a foaf:Document that is a homepage of a workplace of that person. 
This is explained in prose and in RDFS in the FOAF 
specification. I agree that you could, instead, have defined a similar 
relationship that holds between a person and a datatype-represented 
URI that identifies such a homepage. These are different modeling 
styles, each with their own benefits. For FOAF, we focus on the
document,
and not its identifier, since it is the document that is of interest,
and we 
can mention the document even if we don't know it's URI.

There is a larger issue here, which is that of the relationship between 
RDF's built-in notion that resource nodes can have URIs, and its
property-based
mechanism for datatyping some literals as being URIs. I think there is
a 
reasonable case that some future (hypothetical!) version of RDF could
do 
things differently, eg. by representing URIs as properties. I expect
that 
would create a lot of work at the model theory layer, and might not be
worthwhile.
But right now, RDF doesn't have a very natural idiom for describing more
than
one URI for the same resource. You either have to use OWL, and describe
two resources as being 'sameAs', or you have to use properties such as 
dc:identifier, with URI values (as strings? as resource nodes?). The
former is a little awkward, especially in RDF/XML notation. The latter
lacks well-established 
conventions, eg. regarding use of datatyping, and could result in
information
being hidden from RDF-based data merging tools.

To be clear, there really is nothing wrong with that FOAF example. It is
compatible
with the FOAF spec, the RDF semantics, RDFS and RDF/XML syntax.
Using...
<foaf:workplaceHomepage 
> rdf:datatype="http://www.w3.org/2001/XMLSchema#anyURI">http://www.w3.org/</foaf:workplaceHomepage>
...in the present day would hide information from RDF tools, since it
would
fail to merge with other information that attached properties to a
resource node with URI http://www.w3.org/ 

Try it and see: 

<link rel="alternate" type="application/rss+xml" title="RSS"
href="http://www.w3.org/2000/08/w3c-synd/home.rss" />

http://www.w3.org//Overview-about.rdf

Actually if you look at those 2 files, you'll see in the RSS feed that 
the W3C homepage URL is hidden as a string literal in an rss:link
property.

http://www.w3.org/2002/01/tr-automation/ has some more RDF about W3C,
though I forget if it mentions the homepage URL anywhere.

Dan





Dan


> 
> Regards,
> Jeremy Wong 黃泓量
> 
> [1] http://xmlns.com/foaf/0.1/#term_workplaceHomepage
> 
> ----- Original Message ----- 
> From: "Frank Manola" <fmanola@acm.org>
> To: "Jeremy Wong 黃泓量" <jeremy@miko.hk>
> Cc: "Chris Purcell" <cjp39@cam.ac.uk>; <semantic-web@w3.org>
> Sent: Wednesday, June 08, 2005 1:41 AM
> Subject: Re: URI ends with #
> 
> 
> > Section 0.1 of RDF Semantics says:  "There are several aspects of
> > meaning in RDF which are ignored by this semantics; in particular, it
> > treats URI references as simple names, ignoring aspects of meaning
> > encoded in particular URI forms [RFC 2396] and does not provide any
> > analysis of time-varying data or of changes to URI references. It does
> > not provide any analysis of indexical uses of URI references, for
> > example to mean 'this document'."
> >
> > Section 1.2 goes on to say: "The semantics does not assume any
> > particular relationship between the denotation of a URI reference and a
> > document or Web resource which can be retrieved by using that URI
> > reference in an HTTP transfer protocol, or any entity which is
> > considered to be the source of such documents. Such a requirement could
> > be added as a semantic extension, but the formal semantics described
> > here makes no assumptions about any connection between the denotations
> > of URI references and the uses of those URI references in other 
> > protocols."
> >
> > IOW, RDF doesn't associate any special meanings with various URI forms,
> > including those with empty fragids.  They are simply names, to be
> > compared as strings (as noted already). RDF, strictly speaking, doesn't
> > really include the concept of a "document" per se;  the closest thing is
> > a "graph".  I think phrases like "RDF document" in various specs should
> > be understood as meaning "some document that contains RDF statements"
> > (such as an RDF/XML document;  since RDF/XML is XML, it's legitimate to
> > talk about "documents" in this case, since XML formally defines what a
> > "document" is).  The RDF Concepts material referred to here doesn't say
> > that there *are* "RDF documents" corresponding to the URIrefs it's
> > talking about;  it says you can reconcile the RDF and RFC 2396 uses of
> > fragids by imagining that there such documents.
> >
> > --Frank
> >
> > Jeremy Wong 黃泓量 wrote:
> >>
> >> "Two RDF URI references are equal if and only if they compare as equal,
> >> character by character, as Unicode strings" [1]
> >>
> >> It is actually a Simple String Comparison [2].
> >>
> >> Another point is that, the meaning of a fragment is scheme dependent.
> >>
> >> I don't know whether an empty fragid reference the entire document 
> >> either.
> >>
> >>
> >>
> >> Regards,
> >> Jeremy Wong 黃泓量
> >>
> >>
> >> -- 
> >> [1] 
> >> http://www.w3.org/TR/2004/REC-rdf-concepts-20040210/#dfn-URI-reference
> >> [2] RFC3986, Section 6.2.1.  Simple String Comparison
> >>
> >>
> >> ----- Original Message ----- From: "Chris Purcell" <cjp39@cam.ac.uk>
> >> To: "Jeremy Wong 黃泓量" <jeremy@miko.hk>
> >> Cc: <semantic-web@w3.org>
> >> Sent: Tuesday, June 07, 2005 8:05 PM
> >> Subject: Re: URI ends with #
> >>
> >>
> >>> Does an empty fragid reference the entire document? That would mean
> >>> <http://www.example.org/#> is the same URI as <http://www.example.org/>.
> >>>
> >>> Chris
> >>>
> >>>> Thanks Arjohn,
> >>>>
> >>>> I am actually writing a serializer that put me into the question. I
> >>>> am sure that I should throw an exception for this case and those
> >>>> cases mentioned in the Serialising section.
> >>>>
> >>>>
> >>>> Regards,
> >>>> Jeremy Wong 黃泓量
> >>>>
> >>>>>> Consider the rdf:Property http://example.org/#
> >>>>>>  in Notation 3, I can write a triple as
> >>>>>> <http://example.org/A> <http://example.org/#> <http://example.org/B> 
> >>>>>> .
> >>>>>>  in RDF/XML, how can I write the triple? Let me try...
> >>>>>>  <Description xmlns:ex="http://example.org/#"
> >>>>>> rdf:about="http://example.org/A">
> >>>>>>   <ex: rdf:resource="http://example.org/B" />
> >>>>>> </Description>
> >>>>>>  However, "ex:" does not match the definition of Qualified Names
> >>>>>> [1] and the character "#" is not an NCNameChar [2]. I don't have
> >>>>>> any idea to express the predicate which its URI ends with #. Any
> >>>>>> suggestion?
> >>>>>
> >>>>>
> >>>>> This triple cannot be serialized in RDF/XML. See [1] for a short
> >>>>> discussion on the serialization of RDF in XML. The only way to
> >>>>> serialize
> >>>>> such a triple is to use one of the other formats (N3, Turtle, ...).
> >>>>>
> >>>>> Arjohn
> >>>>>
> >>>>> [1] http://www.w3.org/TR/rdf-syntax-grammar/#section-Serialising
> >>>
> >>>
> >>>
> >>
> >>
> >>
> >>
> > 
> 
> 
-- 
Dan Brickley <danbri@w3.org>
W3C

Received on Thursday, 9 June 2005 12:54:17 UTC