The StatementConstraint rdfs:ConstraintProperty

The purpose is to represent in an RDF Schema, the properties a
specific satement can have.


Implementation
--------------

StatementConstraint
    subClassOf rdfs:ConstraintProperty

statementPredicate
    type StatementConstraint
    domain rdfs:Statement
    range rdf:predicate



Illustration
------------

Lets say I have a Person with an intrest in a Topic. (That would be a
resource of type Person and a resource of type Topic in a Statement
with the Predicate intrest.) Suppose I want to represent the strength
of an individual persons individual intrest. That would be a property
of the reified statement, not of the intrest property.


In the implementation, the statementPredicate will be used with the
intrest property.  A Person is said to have a intrest in a Topic.
For each of a persons intrest statements, there will be a property of
the statement indicating the strength of the intrest.

intrest
    type rdf:predicate
    domain Person
    range Topic

IntrestStatement
    subClassOf rdf:Statement
    statemetPredicate intrest

strength
    type rdf:predicate
    domain IntrestStatement
    range FloatZeroToOneInclusive


An application can now know what properties a specific statement can
have by matching the statement predicate with a specific subClassOf
rdf:Statement.


---------

<description about="http://lists.w3.org/Archives/Public/www-rdf-interest/2000Jun/0043.html">

Pierre-Antoine CHAMPIN wrote:
> 
> your example is not valid :
> IntrestStatement is not a statement : it is a class ;
> therefore, the domain of statementPredicate is not Statement, but Class,
> even if you intend to use it only on classes that are subclasses of the Statement class...

You are right.

How do I say that the predicate is only to be used on subclasses of
the Statement class, or the Statement class itself? ... If nothing
else, I could, as you say, change it to:

statementPredicate
    type StatementConstraint
    domain rdfs:Class
    range rdf:predicate


> looks like you are trying to provide RDF statements with Descrition Logic features :
> describing a class in terms of constraints upon its instances...

I think of it as defining propertis for the statement class in the
same way as we are defining propertis for other classes, except that
you don't have to explicitly state the type for each statement.

I find it natural to classify the statement on the basis of the
predicate used.

</description>

> have a look at OIL (http://www.cs.vu.nl/~dieter/oil/oil.nutshell.pdf),
> you may find it interesting...

Yes. I am reading it now. Thank you.

... Have also looked at CyC upper ontology, wordnet, Prolog, predicate
logic, etc. And now the ph ontology. I love it all and can't wait to
use it. So I'm implemnting it in Perl. :-)


-- 
/ Jonas  -  http://paranormal.se/myself/en/index.html