Re: concrete extension

From: Jeff Heflin ([email protected])
Date: 01/23/01


"Peter F. Patel-Schneider" wrote:
> 
> Hi:
> 
> There are differences between the abstract and concrete relationships that
> show up in the syntactic definition.  For example, toClass has a range of
> Class, which is different from the range of toDataType.  This indicates
> that we would lose something if we used toClass for both abstract and
> concrete properties.
> 
> peter

Looking at daml+oil+concrete.daml, toDataType doesn't have a range
specified, which I think in RDF means anything (including a Class) could
be a valid object for the property, but that's not really the issue...

The issue is: are these range constraints artificial (i.e., created just
so we can differentiate the properties) or are the properties
fundamentally different? I argue that they aren't fundamentally
different because they both have identical semantics per
semantics-concrete.html:

x in IC(?R) iff IR(?P)({x}) <= IC(?C)

If the only difference in meaning is that one applies to the concrete
domain and the other to the abstract, why confuse the user with
different syntactical constructs?

Jeff


This archive was generated by hypermail 2.1.4 : 04/02/02 EST