W3C home > Mailing lists > Public > public-swbp-wg@w3.org > January to March 2004

Re: top-3's received so far

From: Jim Hendler <hendler@cs.umd.edu>
Date: Tue, 9 Mar 2004 19:44:12 -0500
Message-Id: <p0602040fbc7413f42fc6@[]>
To: Alan Rector <rector@cs.man.ac.uk>, public-swbp-wg@w3.org

At 15:14 +0000 3/9/04, Alan Rector wrote:
>1) I'd support Natasha's suggestion.  I agree we can't come down 
>from on high, but having just taught a course in using OWL, there 
>are many common errors - in the sense that people don't actually 
>succeed in representing what they intended to and can't understand 
>the result. There are also some patterns that work well and produce 
>modular ontologies, some that seem to work for a while and then 
>break down (e.g. using universal restrictions rather than 
>existential restrictions), and others that are just plain surprising.
>Judging frmo the semantic web list there are also questions about 
>how to represent things that keep coming up again and again.  We 
>need to get a way of presenting issues and options for addressing 
>We'll try to get some of the stuff out for comment.

I think the latter of these two (i.e. the questions that keep coming 
up) are the really important ones to work out..  I agree w/Pat that 
we need to be very careful or we will teach how to build non-Semantic 
Web ontologies fooling ourselves into thinking that the linking isn't 
important (or is somehow bad) and etc.  My instinct is if we start by 
keeping things pretty simple and teaching the representational stuff 
a little at a time, we'll end up in pretty good shape.

>2) The second thing I want is lists of tools with comments and ideas 
>for new tools.  This is quite deliberately self serving.  We are 
>working on new tools in conjunction with Stanford - see the OWL 
>plugin on protege.stanford.edu-->plugins-->backends and the 
>www.co-ode.org websites for the first results.

well, as long as thing are open this could be okay -- but remember 
that if you try to keep out my tools, I'll try to keep out yours (and 
then multiply this by a hundred) -- we need to make sure that we're 
not trying to procribe rather than teach

>We'll try to get some of the stuff out for comment.

Professor James Hendler			  http://www.cs.umd.edu/users/hendler
Director, Semantic Web and Agent Technologies	  301-405-2696
Maryland Information and Network Dynamics Lab.	  301-405-6707 (Fax)
Univ of Maryland, College Park, MD 20742	  240-277-3388 (Cell)
Received on Tuesday, 9 March 2004 19:44:11 EST

This archive was generated by hypermail pre-2.1.9 : Tuesday, 9 March 2004 19:44:12 EST