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

Re: "Real OWL" on the Web

From: Mark Wallace <mwallace@3SigmaResearch.com>
Date: Wed, 15 Apr 2009 09:05:57 -0400
Message-ID: <49E5DBB5.5080709@3SigmaResearch.com>
To: Dan Brickley <danbri@danbri.org>
CC: Semantic Web at W3C <semantic-web@w3.org>
Hi Dan,

Good point about the OWL in FOAF.  Also, as I feared my language may be 
inflammatory in a way I did not intend. 

You mention that "how you compute with that data is a pragmatic choice", 
and I agree.  I guess the heart of my question is more about the 
processing than the language.  I guess I am looking for examples of 
systems out there that are using the OWL distributed around the web for 
significant DL-reasoning.  Do such systems crawl or use seeAlso?  
Examples using RDFS reasoning would also be similarly interesting.  
Again, they may well be there and I am just unaware of them.  Thus the 
question.  :-)

 -Mark

Dan Brickley wrote:
> On 15/4/09 14:41, Mark Wallace wrote:
>> Hi all,
>>
>> In my experience (it may be too limited) I see a lot of "real RDF" on
>> the web, especially in things like FOAF, and Dbpedia. But I am not aware
>> of much "real OWL" on the web, and by this I mean OWL-DL ontologies and
>> knowledge bases, embedded in web pages, and systems that reason over
>> such information. Now I know many folks are hosting ontologies at HTTP
>> URLs. This is not what I am talking about. I am thinking more of OWL-DL
>> statements out there like FOAF is now, and systems that crawl (?) and
>> reason over the combined knowledge.
>
> A quick comment -
> Whenever people do identity reasoning with FOAF data (eg. Sindice, 
> Garlik/Qdos) this is reasoning using properties like foaf:homepage, 
> foaf:mbox etc whose semantics are documented using (real!) OWL. 
> Whether or not a general purpose OWL inference engine is used at 
> runtime is pretty secondary. OWL is a data formats standard, just as 
> RDF is - how you compute with that data is a pragmatic choice. Any 
> vocabulary or dataset (eg. DOAP, SIOC) that is mixed up with FOAF in 
> terms of linked classes and properties, is also partially documented 
> using OWL. So I guess I disagree with any attempt to contrast "real 
> world" FOAF / RDF deployment with "non-real" OWL deployment. Or with 
> the idea that the only real OWL work is OWL-DL...
>
> cheers,
>
> Dan
>
> -- 
> http://danbri.org/
>
Received on Wednesday, 15 April 2009 13:06:33 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:42:11 UTC