W3C home > Mailing lists > Public > uri@w3.org > August 2001

Re: draft-masinter-dated-uri-00.txt

From: Sean B. Palmer <sean@mysterylights.com>
Date: Sun, 26 Aug 2001 17:07:37 +0100
Message-ID: <00fd01c12e49$5c78de80$91ed93c3@Palmer>
To: "Larry Masinter - LMM@acm.org" <lmnet@attglobal.net>, "Aaron Swartz" <aswartz@upclink.com>, "Al Gilman" <asgilman@iamdigex.net>
Cc: <uri@w3.org>
> > Note that earl:TestCase is generally an Object/Class,
> > not a Predicate/Property value. Properties can be
> > Objects too, of course.
>
> AG:: No.

Yes. My terminology was just confusing. In RDF, statements are comprised of
subjects, predicates, and objects. RDF introduces three properties for this
that can be used to model each of these. The domains of these three
properties are, of course, rdf:Statement. The range for "subject" is
rdfs:Resource, the range for "predicate" is rdf:Property, and the range for
"object" is a conjunction of rdfs:Resource, and rdfs:Literal. Confusingly,
because literals are resources. Anyway...

*Properties* can, in the RDF model, be used as the value of an object
predicate on any generic rdf:Statement. Because Classes and Properties are
both sub classes of rdfs:Resource. However, they are both disjoint from
each other, so Classes cannot be used as the value of an predicate
predicate on any given rdf:Statement.

> [...] The object is not the property, it is the 'value' of the property.

Yes. What is sometimes annoying is that people confuse:-

   :name rdfs:range :Name .

with simply:-

   :Name a rdf:Class .

RSS, for example, does this. It uses some URI as a property and a class.
Yuck.

[...]
> Properties are memes, patterns, and not objects.

They are not classes, but they can be objects. Once again, it was the
terminology that was confusing. If properties could not be used as subjects
or objects, then it would be impossible to define them! We are, of course
talking past each other. I understand that properties are not things which
belong to any other class than rdf:Property. In particular, they are
certainly not instacnces of rdfs:Class.

[...]
> > The semantic assumptions for URNs are the same as
> > those for any URI.
>
> AG:: Yes, that's an assumption, but not a fact.

I believe it to be a theoretical fact, but not a practical fact.

> It fails to be true of the large category of de_facto URIs known
> as search URLs.

Once again, the persistence of search URLs is given by their context. In
particular, if you believe that the URI identifies the representation of
its resource, which is simply "a serach result for X", then of course it's
not going to be persistent. In the same way that no changed page is
"persistent". Persistence is defined by the context in which an identifier
is being used.

--
Kindest Regards,
Sean B. Palmer
@prefix : <http://webns.net/roughterms/> .
:Sean :hasHomepage <http://purl.org/net/sbp/> .
Received on Sunday, 26 August 2001 12:07:55 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 13 January 2011 12:15:29 GMT