schema vs ontology

At 06:20 AM 4/26/02 -0400, Dan Brickley wrote:
>Another worry re terminology is that we have two Semantic Web working
>groups. One group calls RDF schemas 'schemas', the other calls them
>'ontologies'. To date we've not really used the latter term in RDF Core
>specs. I personally don't like it, but that's not important. What is
>importnat is terminological consistency at least within the RDF-based
>corner of W3C's work. I'm stumped as to what we do about that. People will
>ask us (of some RDF/XML schema that uses WebOnt machinery) whether it is
>an 'ontology' or a 'schema'. Maybe we don't need an answer, but it does
>seem a bit odd to not know collectively (in WebOnt + RDF Core) what we
>think we're talking about. Editorial suggestions / contribs on this
>welcomed...

Hmmm... I think we ought to be wary about conflating schemas and 
ontologies.  They are, I think, related but different ideas.  Let's see if 
I can say why...

According to John Sowa ["Knowledge Representation", appendix B], an 
ontology is:
"""
a catalogue of the types of things that are assumed to exist in a domain of 
interest [[D]] from the perspective of a person who uses a language [[L]] 
for talking about [[D]].
"""

By contrast, I'd say that a schema is a catalogue of terms in some 
language, together with descriptions of how they can or should be used.


The first difference that appears to me is that
- an ontology is focused on the domain of discourse (semantics), even if 
with reference to a particular language, where
- a schema is focused on the language (syntax).

Further, it seems to me that a schema describes a vocabulary, only some 
terms of which may relate to ontological categories.

In summary, I think we should restrict our discussions to the term 
"schema", as used to describe the vocabulary of RDF.  Other groups may have 
a proper desire or need to use the term "ontology" to talk or reason about 
categories in their domain of discourse, using RDF language and schema as a 
vehicle for expressing some of those ideas.

#g
--

[More...]

John Sowa's Knowledge Representation Appendix B goes on to discuss 
languages like CGs and KIF as ontologically neutral, and then to discuss 
informal and formal ontologies, and in particular axiomatized ontologies.

Against that background, I think that RDF schemas are language tools that 
is/will be used and augmented by DAML/OWL to describe axiomatized ontologies.

I think that the when a schema is described as an ontology, it would be 
more correct to call it a description of an ontology.  This seems to be one 
of those conveniences of expression that is all very well as long as the 
underlying distinctions are understood.



-------------------
Graham Klyne
<GK@NineByNine.org>

Received on Monday, 29 April 2002 07:00:28 UTC