Re: OWL Lite semantics

>In the current semantics draft, OWL Lite just gets the same semantics as OWL
>DL on the syntactic subset.
>It would be possible in our one-dimensional layering to give OWL Lite a
>reduced semantics.
>
>This has the following advantages:
>- clearly differentiation between the two
>- much easier to implement OWL Lite so it really becomes an entry level

Not obvious to me it makes it easier to implement, and it has the 
severe disadvantage of multiplying semantic theories. Seems to me 
that as far as semantics is concerned, the fewer distinctions we have 
the better all round. Having clear differentiation at this level is 
exactly what we want to avoid, unless there are *very* good reasons 
for it. The whole Lbase integration effort is aimed at eliminating 
semantic differences like this.  (If I had my druthers I would give 
up on the extensional model theory for OWL-DL; its only there to make 
some theoreticians feel better, it plays no actual useful role, and 
at some time its going to start making problems for future layering 
efforts.  But that's just a comment, Im not trying to rock the boat.)

>I attach a modified version of section 5 of the semantics doc that changes a
>few iffs to if-then's, and drops comprehension.
>
>As far as I can tell, teh vast majority of the entailments discussed in the
>feature synopsis under OWL Lite are preserved, at much easier
>implementability.

I reserve judgement on that, but would be interested in seeing an 
argument for why it would be. If it supports the same entailments 
then I don't see how it can possibly be simpler to implement; and if 
it doesn't, then its defining a different language.

Pat
-- 
---------------------------------------------------------------------
IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola              			(850)202 4440   fax
FL 32501           				(850)291 0667    cell
phayes@ai.uwf.edu	          http://www.coginst.uwf.edu/~phayes
s.pam@ai.uwf.edu   for spam

Received on Monday, 9 December 2002 00:37:48 UTC