AWWW Mechanism for Interpreting use of RDF Symbols was - Re: ISSUE-58: Scalability of URI Access to Resources

Thanks for the example, comments below (I've changed the title since
this is no longer relevant for ISSUE-58 - at least this particular
part of the thread)

On 9/7/07, Xiaoshu Wang <wangxiao@musc.edu> wrote:
> I think, what is at debate or unclear is this.  Given a sentence, such as
>
> _:me a foaf:Person.
>
> What is the meaning of it?  Is the meaning expressed by just that
> sentence alone or is it by that statement plus the FOAF ontology (due to
> follow the link)?

Yes.  Depending on whether you rely (alone) on the transport layer
('follow-your-nose') to determine meaning or if you try to interpret
via model-theoretic constraints (OWL) you can come to two different
conclusions.

In the first case, you would simply dereference _:me, conclude it is
an information resource (or 'something else') depending on what the
transport layer says, and proceed recursively (decode the media-typed
representations, dereferenced, etc.).

In the second case, the author of that triple might have (and I
*emphatically* believe it is a good practice to do so) also included a
link to a model-theoretic 'definition' via:

foaf:Person rdfs:isDefinedBy <http://xmlns.com/foaf/spec/>

In which case, the agent, resolves the FOAF ontology and comes to
understand that the only interpretation which holds true (against the
ontology) is one where _:me is a person (to the extent that the
ontology has enough constraints to describe what a person is).

> I think OWL and RDF uses two different models.  OWL
> uses explicit import, but RDF use follow-your-nose approach because
> there isn't any "import/include".

I wouldn't go as far as saying RDF 'uses' follow-your-nose, but it
appears that this is the most prominent mechanism being suggested for
RDF.

> The "RDF URI" that Chimezie coined may mean the the latter usage while
> the "symbol" is the URI used in OWL.

This is precisely what I mean, and why I prefer the terms
'symbol'/'sign' because it calls out to mechanisms for interpretation
that are well-founded, have a rich precedent, and are (to some extent
anyways) deterministic.

> The confusion is: given or writing
> an RDF document, how do I know if I should follow the nose or not
> because the interpretation and complexity can be quite different for two
> respective treatment.

The context is important IMHO.  If the RDF document has assertions
(rdfs:isDefined By for example) that are unambiguously understood to
point to a model-theoretic pipeline (if you will), then it would be
prudent for agents to go in that direction.  Otherwise, you simply
don't know, and follow-your nose would be the next best thing.

-- Chimezie

Received on Friday, 7 September 2007 15:21:15 UTC