W3C home > Mailing lists > Public > www-rdf-interest@w3.org > April 2003

RE: Taking advantage of OWL's standardization on a relationship vocabulary - need an OWL API?

From: Ralph Hodgson <Ralph.Hodgson@attglobal.net>
Date: Thu, 24 Apr 2003 10:28:23 -0400
To: "'Roger L. Costello'" <costello@mitre.org>, <www-rdf-interest@w3.org>
Message-ID: <001401c30a6d$c7f12ad0$0201a8c0@TQBEDFORD>

Roger,

You raise important points. In the object paradigm, typically an
interface layer of object brokers mediates the exchange of information
from object instances to relational databases. ODBC and other
Object-Relational APIs have become important as middleware layers.

However you are not talking about instantiating objects, you are
interested in inferencing pver relationships, maybe even transitive
relationships. Your note reminds me of logic-based approaches that I
have seen in Prolog-based systems. 

In the OWL world, I am seeing the importance of an API that uses
Description Logic and/or F-Logic. Semantic middleware vendors such as
Network Inference and Ontoprise have and are creating capabilities to
address the needs you mention. There is a paper on Semantic Integration
on our WEB site at
http://www.topquadrant.com/Docs/TQ0303_Semantic%20Integration.pdf. This
provides an overview of vendors, positioning their capabilities. It
might also interest you how we are busy helping people understand what
business problems can be solved using semantic technologies. Our
galleries of Capability Cases might also be of interest - at
http://www.topquadrant.com/TopDrawer_Gallery/TQgallery_home.htm. Through
scenarios and conceptual architectures, we are using Capability Cases to
explain to business and IT what these technologies can do.

By the way, I am finding the camera example very interesting and I think
there is the potential of a good solution story for us to create here
that explores and promotes the value of OWL for eCommerce applications.

Ralph Hodgson
Executive Partner
TopQuadrant, Inc. 
Office: (724) 846-9300, Fax: (425) 955-5469, Cell: (781) 789-1664
http://www.topquadrant.com



-----Original Message-----
From: www-rdf-interest-request@w3.org
[mailto:www-rdf-interest-request@w3.org] On Behalf Of Roger L. Costello
Sent: Thursday, April 24, 2003 9:24 AM
To: www-rdf-interest@w3.org
Cc: Costello,Roger L.
Subject: Taking advantage of OWL's standardization on a relationship
vocabulary - need an OWL API?



Hi Folks,

My idea of what OWL is and what are its benefits continues to evolve. 
Here's my current thinking:

OWL provides an XML vocabulary for defining terms and their
relationships.  That is, it provides a *standard* set of elements and
attributes, with *defined semantics*, for defining terms and their
relationships.[1]

No question, this is a huge step forward.  XML was the first step.  This
is the second step.

What I see as lacking is: "how does an application utilize a document
which has been written using the OWL XML vocabulary?"

Consider the Camera OWL Ontology.  It is a document that uses the OWL
XML vocabulary (subClassOf, equivalentProperty, etc) to define terms
(SLR, Camera, etc) and their relationships (SLR is a subClassOf Camera,
etc).  How does an application utilize this information?

Let's take an example.  Suppose my application is processing an XML
instance document that contains <SLR>...</SLR>.  Further, suppose that
my application is searching for Camera information.  Thus, my
application would like to know the relationship between SLR and Camera. 
That relationship information is in the Camera Ontology, but how does my
application find it?

One solution is for my application to parse through the Camera Ontology,
looking for SLR, and any relationship it has to Camera.  If every
application has to write this parsing code then there will be a lot of
rundandant effort.

Another solution is to provide a standard OWL API.  The API has all the
parsing smarts in it. Thus, this parsing code isn't written over and
over.

An example method in an OWL API might be:

    isRelatedBy(term1URI, term2URI)
      possible return values: 
          ---> "subClassOf"
          ---> "equivalentProperty"
          ---> "sameIndividualAs"
          etc

The bottom line is this:

1. I know how to define terms and their relationships using the standard
OWL XML vocabulary. 2. I know how to create XML instance documents that
employ the terms defined in 1. 3. It is not clear to me how to utilize 1
in processing 2. 4. Is a standard OWL API useful? 5. What other
approaches are there to programmatically harvesting information from an
OWL document? 6. What do you think?  

/Roger

[1] Note: RDF Schema is, of course, part of this discussion.
Received on Thursday, 24 April 2003 10:40:12 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:58 GMT