Re: RDF vocabulary scope guidelines, promoting properties to classes - property identifiers

I should add something that I realize I did not cover:

> However, we cannot do that in RDF as standardized. As a result,  
> hundreds of domain ontologies are doomed to do exactly what the tag  
> ontology does: reify a domain object and place annotations on it.

Individual ontologies can reduce some duplication by 'lifting' some  
aspects: for example, the Tagging and Knowing classes in this  
discussion could both be subclasses of Event (the former likely being  
instantaneous, the latter having a duration). Event would be the  
domain of a "began" property, thus taking care of both f:beganKnowing  
and tag:taggedOn.

This saves some duplication, but doesn't avoid the fact that almost  
every relationship would benefit in some way from the ability to be  
annotated: RDF still imposes the need to reify simple assertions like  
knowing and tagging.

2 more...

-R

Received on Sunday, 8 February 2009 00:47:39 UTC