W3C home > Mailing lists > Public > www-rdf-interest@w3.org > March 2001

RE: Again: Anonymous Resources

From: Lee Jonas <lee.jonas@cakehouse.co.uk>
Date: Thu, 8 Mar 2001 14:43:13 -0000
Message-ID: <51ED29F31E20D411AAFD00105A4CD7A770BA@zingiber.cakehouse.co.uk>
To: "'Stefan Kokkelink'" <skokkeli@mathematik.uni-osnabrueck.de>
Cc: Aaron Swartz <aswartz@swartzfam.com>, RDF interest group <www-rdf-interest@w3.org>

Stefan Kokkelink [mailto:skokkeli@mathematik.uni-osnabrueck.de] wrote:
>Lee Jonas,
>I completely agree with your interpretation of
>anonymous resources! But this has some consequences
>for (valid) manipulations of RDF models, see [1]
>for an example. And there are other points that
>need clarification. Let's do some odd things ;-) :
>  <rdfs:subPropertyOf>
>    <rdfs:Property>
>       <rdfs:subPropertyOf>
>         <rdfs:Property/>
>       </rdfs:subPropertyOf>
>    </rdfs:Property>
>  </rdfs:subPropertyOf>
>According to RDF Schema this leads to the conclusion
>that the first Property is also subPropertyOf the
>third one. But: you can't express this fact in the
>XML serialization (although you are the owner of
>the data).

Yes, you can infer from the transitive nature of rdfs:subPropertyOf that 
if A is a B and B is a C then A is also a C.

However, this is implied.  If you tried to represent the statement of {A ->
rdfs:subPropertyOf >- C}, directly in RDF it would actually lead to a
multiple inheritance triangle:

      | \
      |  B
      | /

I assume this is not what you intended.  Nevertheless, the RDF spec leads me
to believe that if you want to refer to some arbitrary resource from
multiple places, you shouldn't make it anonymous - give it a name.  In the
example above, associate a rdf:ID with C and refer to that from A and B.

Can you think of any more examples that would stress the interpretation we
agree on?

>I agree with Brian (great paper!): 'It is unfortunate 
>that the XML serialization defined for RDF does not 
>permit the representation of all possible
>graphs containing anonymous resources.' (Section 2.2) 
>I think this is a bug in the RDF serialization. Perhaps
>the serialization should provide an 'rdf:anonymous_ID'
>attribute in addition to rdf:ID. rdf:ID specifies
>an URI, rdf:anonymous_ID a local identifier that is 
>allowed to be changed (more precisely: mapped one
>to one) by an RDF application (e.g. 
>parser) processing the data. And this identifier should
>not be interpreted as a URI, it's private data!

What would the rdf:anonymous_ID actually accomplish over rdf:ID?  If you
give a name to something it is no longer nameless.


Received on Thursday, 8 March 2001 09:43:38 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:44:28 UTC