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

Re: (Round 2) Proposed Extensions to OWL

From: Roger L. Costello <costello@mitre.org>
Date: Fri, 04 Jul 2003 11:32:22 -0400
Message-ID: <3F059E06.341538F@mitre.org>
To: www-rdf-interest@w3.org
CC: jon@spin.ie, "Costello,Roger L." <costello@mitre.org>

Hi Folks,

The discussions have been excellent.  I believe that we are making good
progress. 

Originally we had just one issue.  Interestingly, now we have two
issues:

(1) What is the proper method for creating descriptions that involve
units of measure?  This is an instance document issue.

(2) What is the proper method for defining the relationships between two
descriptions that use different units of measure?  This is an ontology
issue.

Hopefully, a good solution to (1) will facilitate a good solution to
(2).

I believe that we are converging on a solution to (1).  In this message
I will focus on (1).

We have seen many suggestions for creating descriptions that involve
units of measure.  I think that we all agree that a description (model)
is the result of a transformation of a physical entity into a
representation that makes it easier for humans and machines to deal
with.

Example: 

   length-transform(Yangtze River) = length-description

There are many transformations possible: one to generate a description
that uses a kilometers unit of measure, another to generate a
description that uses a miles unit of measure, etc.

Example: 

   area-transform(Wilson Farm) = area-description

Again, there are many transformations possible: one to generate a
description that uses a square miles unit of measure, another to
generate a description that uses a miles-inches unit of measure, etc.

The latest discussion is that there must be a flexible way of describing
the transformation that was used to produce the description.  A flexible
version of <transform> is being proposed.

Below are two examples which uses the latest proposals, including some
new element names that Benja proposed, as well as a twist of my own. 
Note especially the use of the transform elements.

This document describes (models) the length of the Yangtze River:

<River rdf:ID="Yangtze">
    <length>
        <PhysicalQuantity>
            <measurement>
                <Measurement>
                    <length-transform rdf:resource="LengthInMiles"/>
                    <number>3914</number>
                </Measurement>
            </measurement>
        </PhysicalQuantity>
    </length>
</River>

This document describes (models) the area of the Wilson Farm:

<Farm rdf:ID="Wilson">
   <area>
      <PhysicalQuantity>
         <measurement>
            <Measurement>
               <area-transform>
                  <TransformProduct>
                     <length-transform rdf:resource="LengthInMiles"/>
                     <length-transform rdf:resource="LengthInInches"/>
                  </TransformProduct>
               </area-transform>
               <number>625</number>
            </Measurement>
         </measurement>
      </PhysicalQuantity>
   </area>
</Farm>

where length-transform and area-transform are subproperties of
transform:

                        transform
                         |     |
               -----------     ----------
               |                        |
           length-transform         area-transform

Thus, the transform property has considerable flexibility in how it
expresses the transformation of a physical entity into a description
(model).

What do people think about this approach?  

A real good test of this approach would be to describe the "location" of
something using a coordinate system.  Since I have long forgotten about
coordinate systems I dare not take a stab at it.  Is there anyone who
deals with locations and coordinate systems on a regular basis that
could do this?  /Roger
Received on Friday, 4 July 2003 11:34:25 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:00 GMT