> Suppose foaf:Person has as rdfs:comment something like
> "A class of http retrievable web documents, which
> contain information about persons" -
>
> If thats the case i can't see any problem to assert:
>
> There is a resource typed foaf:Person, which has format
> application/xhtml+xml
The fact is though that foaf:Person has an rdfs:comment of "a person".
> If one accepts
>
> foaf:Person<--rdf:type-- ? --dc:format--> application/xhtml+xml
>
> as true and the only knowledge you have...Why the hack one
> could come to the idea, that ? would be, would have been or
> will be a person.
Exactly the problem. Since ? *is* a person the above triples are obviously
wrong.
However knowing that ? is a person, and knowing that the dc:format of a
representation (or *the* representation) is application/xhtml+xml are both
useful things to know.
It is only by considering the representation as a resource, and defining the
relationship between ? and that representation that we have a hope of giving
both pieces of information in a flexible (i.e. without depending on
assumptions certain applications might make) way.
|