[
Lists Home |
Date Index |
Thread Index
]
Simon St.Laurent wrote:
>jonathan@openhealth.org (Jonathan Borden) writes:
>
>
>>I didn't mean to imply that you could *calculate* intersections. The
>>intersection is the shared part, whatever is outside the intersection
>>is your own private Idaho.
>>
>>
>
>And I need RDF/OWL for that how?
>
>
You certainly *don't*. I never meant to imply that you did. All I meant
to say was that neither the RDF nor OWL model theory based semantics
suggest that when you send a (for the sake of argument here: XML based,
or better RDF/XML based) message to someone else, that your and the
other party's semantics for the message need be identical, only that you
and the other party share the semantics as identified by the model
theory (assuming you both agree that you are conforming to the
appropriate model theory).
For example, we need to agree that (in N-triples) "1"^^xsd:integer
represents the number _1_, not the number 2. It could be one of
*anything*. Gosh, if we aren't able to agree that one is one, then we
have a tough time communicating anything useful to either of us. Note
that the XML
<foo>1</foo> or
<foo bar="1" />
need not have *anything* to do with numbers, so syntax alone won't get
you where most people want to go.
Do you need RDF/OWL for *that*. Certainly not, there are a whole variety
of ways that you can specify shared semantics e.g.
1) "Hey Simon, I am going to email you a quote for that purchase you
want to make. Just paypal me the amount if it's ok"
2) A schema, our programs agree to use a particular schema.
3) RDF/OWL
etc.
The point is that RDF/OWL is *a way* to specify shared semantics, not
the only way. It is as useful as is the subset of all software that is
written to process it. The more software that is written to conform to
RDF/OWL, the more useful it gets, just like the browser itself, or PDF
Word, etc.
Jonathan
|
