W3C home > Mailing lists > Public > www-rdf-interest@w3.org > March 2000

Re: RDF data model: too flat?

From: Dan Brickley <Daniel.Brickley@bristol.ac.uk>
Date: Thu, 9 Mar 2000 10:41:12 +0000 (GMT)
To: R.van.Dort@Everest.nl
cc: www-rdf-interest@w3.org
Message-ID: <Pine.GHP.4.21.0003091032460.20932-100000@mail.ilrt.bris.ac.uk>


It sounds like you're arguing for some distinction such as that between
'intrinsic' and 'extrinsic' properties of an object. If you're not, it
might be that properties that have literal (non-URI /
non-Resource) values would be much the same sub-set of properties as
your 'attributes'. Either way, it's a hard distinction to make
in a water-tight fashion (eg. Philosophers have bickered for centuries
about such things as whether 'colour' is intrinsic or relational),
particularly when the Web architecture encourages us to use URI-namable
objects whenever possible.

Rather than extend the core RDF model to deal with
intrinsic-vs-relational properties, such distinctions might be layered
on. For example, a sub-class of rdf:Property called
rdfutil:IntrinsicProperty. This sort of distinction I'm sure would prove
useful for many applications (eg. user interface generation), but I
can't personally see a need for it to be in the core data model itself.

Dan


On Thu, 9 Mar 2000 R.van.Dort@Everest.nl wrote:

> Working with the basic RDF model I get the feeling that the RDF data model
> is too flat.
> 
> One of the aspects that has a contradiction with the world around us is
> that in RDF there is (speaking in terms of frame based theory) only one
> type of slot: Property.
> 
> In real world we see objects or things (RDF: Resources) with attributes and
> relations between objects.
> Attributes are dependent on their objects in which they are contained, in a
> (binary) relation two objects are independent in existence but connected.
> 
> For example the weight, color of the eyes and day of birth are typically
> attributes of a Person instance; the father, mother, spouse, car and bank
> account are independent objects related to a certain Person.
> 
> I would make a plea for a second type of slot in the RDF model: Relation or
> BinaryRelation.
> 
> I think that my suggestion is close to the CKML fundamentals, maybe the
> CKML model would be the outcome when we think things all over.
> 
> 
Received on Thursday, 9 March 2000 05:43:41 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:42 GMT