Re: Responses to "Draft of charter for NextWebOnt (Proposed) Working Group"

>At 5:02 PM -0500 1/12/07, Kendall Clark wrote:
>On Jan 12, 2007, at 1:15 PM, Jim Hendler wrote:
>
>  I would like to see one "OWL Ultralite" that is as close to RDFS as possible
>
>
>I'd be happy to look at the model theory or axiomatization of such a 
>beast, if it's available. Not having seen anything yet, it's hard to 
>say whether it's interesting, either practically or theoretically.
>
>
>
>Kendall - you've seen the model theory!  This has been said all 
>along to be a subset of OWL, so the OWL documents provide the model 
>theory (and reference, and examples, and test cases, etc.)
>

Obviously (?) I would also be very interested in getting this right, 
also. But I think that rather than basing the model theory on OWL, 
the right way to do this is to follow the ideas developed by Herman 
ter Horst, who has written a lot on extending the RDFS style 
semantics. In particular, it ought to follow the RDF/S semantics in 
being non-extensional rather than the OWL insistence on 
extensionality: this gives a 'lighter' logic which is simultaneously 
more useful AND more tractable than a rigidly extensional logic. The 
pragmatic benefits of removing the extensionality condition have now 
been well validated (in a wider context) by the experience of using 
CL in the IKRIS project. Insisting that OWL be extensional was IMO a 
mistake, a good example of allowing theoretical 'elegance' to 
override useability considerations.

This would still be a subset of OWL in the sense that it would be a 
weaker logic than OWL, ie if an inference is OWL-valid then it is 
OWLUL-valid. (There might be some edge exceptions to this involving 
sameAs, but we can find ways to hack those with a little work.)

Kendall, I can send you a sketch of the model theory (which is OWL 
Full-ish but built on an RDFS foundation) and a brief summary of how 
it weakens the logic but in a GOOD way, some time early in February: 
the rest of my January is all booked up.

Pat

>While I disagree that the TFs are motivated by "theoretical aspects" 
>only, they have the virtue of having been written down and can be 
>discussed publicly.
>
>Well of course you haven't seen it yet - we're discussing whether 
>this WG should be the place to create one, whether people will 
>submit through the notes process, or whether this will occur in some 
>other WG.  I certainly have opinions as to what I'd like to see in a 
>language, but that's not the discussion we've been having in this 
>thread.
>  To be clear, my comment is that, as I have stated publicly, and 
>continue to do so, my WG made a mistake in not considering users 
>enough in the design of the "lite"  fragment of OWL.  I do not want 
>to see another WG make the same mistake, this is called experience. 
>So I ask the WG to either take on the issue by adding usability to 
>the design criteria (as many WGs do) or to remove this fragment from 
>the scope of THIS working group so they can be worked on by a group 
>that will be responsive to the commercial comments we've been 
>hearing.
>  -JH
>p.s. note that this is only one of many things in the scope of this 
>WG, so this is only an argument at the moment on this small piece of 
>the charter, not on chartering the WG more generally
>
>
>--
>Prof James Hendler				hendler@cs.rpi.edu
>Tetherless World Constellation Chair		http://www.cs.umd.edu/~hendler
>Computer Science Dept			301-405-2696 (work)
>Rensselaer Polytechnic Inst			301-405-6707 (Fax)
>Troy, NY 12180


-- 
---------------------------------------------------------------------
IHMC		(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32502			(850)291 0667    cell
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes

Received on Sunday, 14 January 2007 18:45:20 UTC