[topicmapmail] Meaning of URIs - ongoing debate on new W3C forum

[Lars Marius Garshol]

* Thomas B. Passin
| 
| That is part of my point - there are other ways to use uris in RDF -
| I meant mainly when they are the values of properties, and a literal
| is one kind of property value.  It is only in their use in rdf:about
| or rdf:ID that a uri is used as an abstract identifier.

Again, I wish you wouldn't use the syntax as your reference. It makes
it really difficult to discuss this, because the syntax is such a
mess, and the model is so much simpler. 

I guess what you mean to say is that it is only when a URI appears as
the subject of a statement that it is used as an abstract identifier.
That is wrong: a URI is used the same way regardless of where in a
statement it appears (subject, property, or object), so long as it is
actually a URI and not a literal.
 
| <ex:organization rdf:about='uri:example:oneMoreOrganization'
|     ex:resourceRef='http://www.w3.org/index.html'/>
| 
| This is another example of a uri used as a property value.  With the
| proper definition, ex:resourceRef could mean essentially the same
| thing as a topic map resourceRef.  

You could use it that way, but if someone used the same URI as the
subject of a statement like this

  ('http://www.w3.org/index.html', rdf:type, ex:organization)

in the graph the two URIs would compare equal, and so they would be
considered to identify the same resource, but in this case one
statement is treating the resource as an organization, while the other
is treating it as a document.

If you read Steve Pepper's identity crisis paper it's all about this
issue, and I think it would explain where the problem lies.

| Note that rdf now allows one to type a property, so the above could
| have been typed with the type xsd:anyURI. The rdf processor would
| have to know that it is a uri then.

You can't do that, since that kind of typing is for literals, and you
are not using a literal there. You could say

  ('http://www.w3.org/index.html', rdf:type, tpassin:document)

but that wouldn't affect the interpretation of the URI.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
GSM: +47 98 21 55 50                  <URL: http://www.garshol.priv.no >