W3C home > Mailing lists > Public > www-rdf-logic@w3.org > June 2003

Partial Imports (was www-webont-wg@w3.org Re: Proposed response to Golbeck regarding imports issue)

From: Jim Hendler <hendler@cs.umd.edu>
Date: Tue, 17 Jun 2003 09:34:37 -0400
Message-Id: <p05200f16bb14c8b060ed@[10.0.1.2]>
To: Ian Horrocks <horrocks@cs.man.ac.uk>, www-rdf-logic@w3.org
Cc: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>, jjc@hplb.hpl.hp.com

[[ This issue is moving here from webont WG as it has become a 
generic discussion of importing (yet again) rather than about a 
specific response.

You are welcome to go to www-webont-wg@w3.org archives to see earlier 
messages in the thread, but I think you'll get the gist from the 
below ]]


At 2:11 PM +0100 6/17/03, Ian Horrocks wrote:
>On June 17, Jim Hendler writes:
>>
>>
>>  >
>>  >>  In certain contexts, I think OWL would be useful with some ontologies
>>  >>  preimported.
>>  >
>>  >Well, this would not be OWL.
>>
>>  With due respect Peter, this must either be the dumbest thing I ever
>>  heard you say or, more likely, we're somehow not understanding each
>>  other.  Most of our tools enable the user to start with ontologies
>                       =====
>                       =====
>
>What you describe here are OWL tools, not the OWL language.
>
>If I were Dan, what I would do is produce a <<tool>> that checked on
>the usage of namespaces in an OWL ontology and automatically added
>the import statements that I wanted. This would allow my applications
>to work the way I wanted using reasoning that would still be sound
>w.r.t. the OWL language spec.
>
>Regards, Ian

that's how you would do it.  My plan is to build tools that don't 
import whole documents when they see external pointers, they will 
only import some statements from those documents.  However, the 
graphs they create will still be legal OWL (Full) graphs and usually 
will be legal OWL DL.  What they won't necessarily be is documents 
that completely capture the intent of the ontology provider, and 
which may therefore not produce the same results as an imports would 
have stated.  They may also lose "ground truth" in some deeper sense, 
but they certainly won't be incorrect OWL.   (i.e. if I "invoke" a 
statement from your web page (ex:) that "Ian Horrocks" is an ex:male" 
but don't import the statement that "ex:male owl:DisjointClass 
ex:female", there is nothing wrong with my believing that you are 
ex:male - or even finding properties of that class and assuming you 
have them.  There's certainly nothing wrong in my OWL - and my 
application could still produce very useful (and sound and complete) 
results - it's only if I brought in other statements that involved 
ex:female that a problem would occur, and I see no reason why tools 
couldn't handle that correctly (by importing only the parts of 
documents that are needed for particular applications).  A real world 
example of this is that many people are using the portion of the CYC 
upper ontology for "time" without reading in the whole rest of CYC.
   This is why I care so much about this imports stuff.  You may think 
the above is stupid or illogical, I think that anything else will 
cause the semantic web to fail.  OWL is nice in that we both can have 
our mistaken beliefs and still build tools and ontologies that 
interoperate just fine.
>
>
>
>
>>  pre-imported -- for example we are building a cancer research project
>>  that starts from the NCI OWL Lite ontology.  It comes preloaded.  If,
>>  on the other hand, we started from having the user hit a button and
>>  import that ontology, it would not come preloaded.  I cannot see how
>>  this would make any difference to whether something is OWL or not.
>>
>>  My suspicion is there is some deeper issue which you are responding
>>  to. If we're just arguing about how the term "is OWL" is used, then
>>  it isn't worth much time, because the use of our vocabulary is out of
>>  our control once we publish it.
>>
>>    -JH
>>
>>  --
>>  Professor James Hendler				  hendler@cs.umd.edu
>>  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)
>>  http://www.cs.umd.edu/users/hendler      *** NOTE CHANGED CELL NUMBER ***

-- 
Professor James Hendler				  hendler@cs.umd.edu
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)
http://www.cs.umd.edu/users/hendler      *** NOTE CHANGED CELL NUMBER ***
Received on Tuesday, 17 June 2003 09:34:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:46 GMT