W3C home > Mailing lists > Public > www-rdf-logic@w3.org > February 2002

Re: [semanticweb] The level of abstraction (modeling problem)

From: Seth Russell <seth@robustai.net>
Date: Tue, 19 Feb 2002 12:44:24 -0800
Message-ID: <004d01c1b986$374353c0$657ba8c0@c1457248a.sttls1.wa.home.com>
To: "ejchen77" <jchen@itri.org.tw>, <rector@man.ac.uk>
Cc: "RDF-LOGIC" <www-rdf-logic@w3.org>, <semanticweb@yahoogroups.com>
From: "Alan Rector" <rector@cs.man.ac.uk>

> This view does require some clarification
> of what we mean by 'things in the
> world' and 'having kinds' and special
> problems about ideas, numbers, etc. which are
> beyond a short email, but the simplest test is 'can
> it have a kind?'  If it can, it is a class; if it
> can't it is an instance.
> For most things, that is a useful test.

Indeed it is and thanks for the clear explanation of classes and instances.
But I still have a problem thinking about this where it comes to things that
can be exactly copied, like documents and statings.  Obviously these copies
carry on their unique existences in the world with different histories after
they are copied.   So I arrive at the conclusion that if we want to model
things and their copies, then we must refer to the copies as a class.

Example: The statings of the triple: {:Bush :wonThe :Election2000} seems to
be a class of things ... there are instances of this class in your email
inbox box, and there are instances of it in my email box.  Please see my
previous post relative to this on RDF-LOGIC [1]


Is there any precedence for this ?

Seth Russell
Received on Tuesday, 19 February 2002 15:47:50 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 2 March 2016 11:10:37 UTC