Re: SEM: Light review of semantics document

pat hayes wrote:
> 
> Its OK to say that RDF docs using OWL-defined terms are *often* used
> to encode data (particularly as it happens to be true :-) and to
> refer to such docs as OWL data documents, and say that the OWL doing
> the defining can be seen as playing something like the role of a data
> schema. But make sure that you also say that this is a *common
> practice*, not a matter of definition; that people should not assume
> that all RDF docs are like this; that one can use both RDF and OWL in
> other ways; and that as Jonathan emphasized, there is in fact no
> built-in sharp boundary between instance data and defining ontologies
> in this family of languages. It is particularly important to
> emphasize this latter point as many of our readers will find the idea
> that there is no such sharp boundary new, and may be surprised by it,
> and we want to make sure they don't accidentally assume it is there
> and incorporate it into software, say. (That is, if they do so
> incorporate it, that they do so with their eyes open.)
> 

Pat,

I think this was all Jim and I were suggesting. We realize that you
could easily include OWL terminology in any RDF document, not just those
that have be annointed "ontologies" by use of the owl:Ontology tag. It
seems there are essentially three classes of documents:

OWL ontologies - those that have an explicit owl:Ontology tag. These may
include instance data

OWL data documents - those that do not "define" any classes or
properties

mixed documents - those documents that do not have an Ontology tag but
define one or more classes and/or properties. This third class
emphasizes your point of there being no sharp boundary.

Of course, all of these documents are RDF documents.

Jeff

Received on Wednesday, 20 November 2002 10:00:34 UTC