Re: Associations in RDF

Mike,

Suppose it will be interesting for you to look at some more deep
foundation for Properties and Associations. I mention a Theory in the
paper "To keep abreast of the 21st Century". I have put an English
version on my site http://ototsky.mgn.ru/it under the "For CIO" item.


Best regards,
 Leonid
mailto:leo@mmk.ru and copy to leo@mgn.ru
=====================================================
Leonid Ototsky,
http://ototsky.mgn.ru
Chief Specialist of the Computer Center,
Magnitogorsk Iron&Steel Works (MMK)- www.mmk.ru
Russia
=====================================================

Wednesday, July 17, 2002, 9:35:43 PM, you wrote:


Mac> Hi Everyone,

Mac> I have always felt uncomfortable with the notion of predicates
Mac> being represented as properties.  To me, predicates seem to have
Mac> two types: relation (or association) and attribute.  Attributes
Mac> are suitably represented as properties (could be considered synonyms).
Mac> By forcing all predicates to be properties ... are we not shortchanging
Mac> predicates as associations?

Mac> For example, what about independent or temporary associations that should
Mac> not be tightly bound to a subject?  Additionally, a relation that has an
Mac> aspect
Mac> of degree is poorly modeled via a property.  Take an example of
Mac> friendship which could have the following degrees (and many more):
Mac> acquaintance, recurring_acquaintance, new_friend, childhood_friend,
Mac> lifelong_friend, intimate_friend, ... etc.  If we modeled all of these as
Mac> properties, the property list of an object would grow to be unmanageable.
Mac> Instead we need to put characteristics into the association itself.  Thus,
Mac> we model an Association as a class and we can have a simple property
Mac> that points to a particular association subclass hierarchy.

Mac> Has anyone else experienced this?  Is this a deficiency or is it covered
Mac> in ways that I am unaware of?

Mac> If it is not covered, I would propose that there be two subclasses of
Mac> Predicate: Property (which exists) and Relation (or association).
Mac> Thus as a separate first-class object, associations can be rapidly
Mac> queried against and not be confused with simple properties.

Mac> I think associations are important enough to be explicitly
Mac> called out.  In seeing the RDF mapping to UML, these are made
Mac> explicit.  I just think this is so basic that it should be in the base
Mac> RDFS spec.

Mac> What do you think?

Mac>  - Mike
Mac> ----------------------------------------------------
Mac> Michael C. Daconta
Mac> Director, Web & Technology Services
Mac> www.mcbrad.com

Received on Thursday, 18 July 2002 02:01:02 UTC