RE: [regrep-semantic] Mantech/DoD EII Workshop

["Jeff Pollock" <Jeff.Pollock@networkinference.com>]

Farrukh-
 
Well, I think that is sort of a contentious question possibly.
 
Some might say "no" - because you could theoretically pre-classify the inferences into the RIM and query that using the native query approach.  
 
I would say "yes" because there are inferences that you can do at query time that go beyond a pre-classification approach (depending on what you query for) and also because using the inference engine for query would effectively enable one to treat the semantic content registry as a changeable virtual schema - whereby content changes could cause a reclassification of other registry data in realtime. 
 
Without the inference, you may as well be querying RDF because all you can get out is what you've put in, and even worse, like UML, schema-level changes would become compile-time type procedures.
 
But both ways would certainly work, and as some in this group have pointed out, the use of OWL's built in semantics alone is valuable - even without inference.
 
Cheers!
 
-Jeff-
 
 

	-----Original Message----- 
	From: Farrukh Najmi [mailto:Farrukh.Najmi@Sun.COM] 
	Sent: Tue 6/29/2004 10:00 PM 
	To: Jeff Pollock 
	Cc: regrep-semantic@lists.oasis-open.org 
	Subject: Re: [regrep-semantic] Mantech/DoD EII Workshop
	
	

	Jeff Pollock wrote:
	
	Jeff,
	
	Thanks for sharing this very informative presentation. The content was
	easy to follow.
	
	I like the distinctions you make between the roles played by Inference
	Engine, Mediation Engine and Semantic Content Registry.
	In my mind I have been assuming that the Semantic Content Registry would
	rely upon an inference engine in order to support Semantic Query use
	cases. Does this kind of dependency fit your mental model?
	
	--
	Regards,
	Farrukh
	
	--------------------------------------------------------
	Going to Java One 2004 June 28 - July 1?
	
	http://java.sun.com/javaone/
	
	Come see the newly released freebXML Registry 3.0
	at pod 1220 in the Java One Pavilion:
	
	http://ebxmlrr.sourceforge.net
	http://ebxmlrr.sourceforge.net/presentations/freebXMLRegistryBrochure.pdf
	http://ebxmlrr.sourceforge.net/presentations/xmlEurope2004/04-02-02.pdf
	       
	--------------------------------------------------------
	
	
	
	
	> Friends and Colleagues,
	>
	>     
	>     I just concluded a two-day workshop held by Mantech International
	>     for the Department of Defense on the subject of enterprise
	>     information integration. There were many interesting presentations
	>     from thought leaders out of Unicorn, MetaMatrix, Contivo, IBM,
	>     Microsoft and more - each on the subject of information level
	>     interoperability.
	>     
	>     The fascinating thing about this workshop was the hands-on
	>     interaction between vendors, and also the CIO architect thought
	>     leaders from Army, Navy, Air Force, DLA, OSD and others. This was
	>     clearly a battle of competing ideas.
	>     
	>     We had the inventors/founders of UDEF, Topic Maps, BizTalk, and
	>     Contivo in the room, and of course I was doing my best to
	>     represent the W3C OWL/RDF standards-based view of information
	>     interoperability - and Network Inference's strengths in those
	>     areas of course.
	>     
	>     I had the honor of doing the keynote for the two day event, I
	>     wanted to share that with all of you so that you can see the kinds
	>     of things I am speaking towards these days - I hope you enjoy the
	>     attached briefing (and workshop outbrief).
	>     
	>     Also, over the coming few days, you will be able to view and
	>     download all the presentations from the speakers at the following
	>     URI: http:\\www.meiim.com
	>     
	>     One final thought; in speaking with bright minds from so many
	>     different backgrounds I have come to distill one simple
	>     proposition that OWL gives users that no other technical approach
	>     can claim: using OWL-driven inference, your data schemas become
	>     virtual - they can adapt in response to outside business change
	>     (and propagate those changes to instance data). Other systems, so
	>     XSD/XML-centric, are locked into a fixed, unresponsive set of
	>     schematic metadata/data that remain ever-so brittle in these
	>     changing times. Build for change!
	>     
	>     All the best,
	>     
	>     -Jeff-
	>     
	>     ------------------------------------------------------------------------
	>
	>      
	>
	>     Jeff T. Pollock
	>     Vice President, Technology
	>
	>Main:   760-476-0650
	>Direct: 415-971-2223
	>Fax:    760-476-0648
	>Web:    www.networkinference.com
	>Email:  jeff.pollock@networkinference.com <mailto:jeff.pollock@networkinference.com>
	>         
	>
	>         5900 La Place Ct., Suite 250
	>         Carlsbad, CA 92008
	>
	>  The schema is virtual.
	>
	>     
	>