W3C home > Mailing lists > Public > semantic-web@w3.org > May 2009

Re: [Welcoming feedback] Semantic Web: Information wants to be useful

From: Pat Hayes <phayes@ihmc.us>
Date: Thu, 7 May 2009 12:40:23 -0500
Cc: semantic-web@w3.org
Message-Id: <1ABE880F-AD15-4F38-8B93-7BD8F5AEC28A@ihmc.us>
To: Nicolas Raoul <nicolas.raoul.lists@gmail.com>

On May 7, 2009, at 1:44 AM, Nicolas Raoul wrote:

> Dear all,
>
> On Saturday I will give a presentation at Tokyo's Linux Users Group,
> the goal being to entice attendees to use the Semantic Web data,
> create mashups, and publish their data. The goal is not to teach them
> exactly how to do, nor to explain theories, but to simply show what
> the Semantic Web enables, and make them want to know more. Attendees
> are Linux/Java/Open Source/IT specialists, most of them with no
> Semantic Web experience.
>
> I wrote these slides: http://nrw.free.fr/tlugsw.pdf
> I am sure the slides contain errors and misrepresentations, and I
> would be really glad to get some feedback so that I can fix them (and
> learn).
>

Nice presentation, but one or two quibbles.

Slide 10 says "any information in the world" [can be represented as  
RDF triples.] Well, no, not really (which is why we need OWL and SWRL  
to write ontologies.) This could be rephrased to say something  
slightly less all-inclusive. Most of the data in the world... maybe,  
or some such phrase.

Slide 11 the 'mosquito is in class insect' example reads strangely.  
The natural way to say this would be [Mosquito] is a kind of [insect],  
or better still:  Mosquitoes are a kind of Insect.

Slide 25 says "Writing an ontology is a bit similar to writing a  
hierarchy of classes and
attributes in Object Oriented Programming."

Aaaaargggh. I guess you have to say something like this to a LINUX  
audience, but this is WILDLY misleading. I think there are more errors  
caused by people coming to RDFS and OWL with an OO mindset about what  
'class' means than from any other single cause. Classes in ontology  
languages are NOT AT ALL like classes in OOP. Maybe you could say it  
was "superficially" similar to an OOP hierarchy, or something like  
that? Or put in a warning?

Slide 26. Be careful. Using owl:SameAs too freely is very dangerous.  
This is a current hot-button topic in several SWeb application areas.  
People should only use owl:sameAs when the entities are not only  
*exactly* identical, but also what each of their 'defining' (home)  
ontologies say about them is compatible. This is rarely the case, in  
practice. In your example, I know both Cyc and dbPedia say their  
concepts are sameAs one another, but they are both wrong. Cyc defines  
a 'piece' of carbon; dbpedia defines the chemical element carbon.  
These concepts are NOT owl:sameAs one another, no matter what the  
websites say.

I would omit this whole owl:sameAs discussion from an primer-intro  
slideshow for newbies, myself. Its just too fraught with difficulties  
and unresolved issues.

Pat Hayes

> Thank you very much for your time!
> Nicolas Raoul.
>
>
>

------------------------------------------------------------
IHMC                                     (850)434 8903 or (650)494 3973
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
Received on Thursday, 7 May 2009 17:41:05 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 21:45:29 GMT