WOWG: OWL Lite vs OWL DL-Lite

On yesterday's Semantic Web Coordination Group telecon [1] and in a 
less formal session following it (same log) there was, yet again, 
discussion of the issues of how RDF, RDFS and OWL fit together -- 
there is concern that we haven't explained this well in some of our 
documents, and it may cause problems come LC time.  The problem is 
that by going with what we once referred to as our "1-dimensional" 
approach to Lite subset of DL subset of Full, we convey the idea of 
an upgrade path in which RDFS documents can "easily" be upgraded to 
Owl Lite.  Problem, of course, is that this is not really true -- 
RDFS is easily upgraded to OWL Full (using the Lite vocabulary 
subset) and several of our implementors - HP, Protege, various DAML 
sites, have expressed an interest in supporting what is essentially 
our un-named sublangauge -- OWL Lite language restrictions with Full 
Semantics.
  It is unclear whether what is needed would be actual design work, or 
more likely just better explanation in the Overview and Reference 
documents (or maybe a separate document or Note discussing this). 
Looking at our documents - the Overview as it is currently written 
primarily defines Owl Lite as an easier-to-implement vocabulary 
restriction, so it really does make sense to have OWL DL and OWL Full 
and then to have Lite subsets of each.
  We should consider how we as a group feel about addressing this.  It 
has become very clear that if we go to LC with just our current set 
of documents, we are likely to get LC comments on this and have to 
address them formally.  It also effects the resolution to some of the 
issuettes Jeremy raised -- if we are less concerned with Owl Lite 
being the direct migration path for RDFS, then we don't have to worry 
as much about whether some of the other RDF language features (bags, 
seqs, etc.) are in Owl Lite (as they would be in Owl Full Lite, but 
not OWL DL Lite)
  We originally decided not to do this (no resolution, just general 
discussion) because we were afraid it would confuse people too much. 
The feedback we are getting now is that maybe we are doing the 
opposite and confusing people the way we have it (consider: the 2-D 
diagram is actually easier to draw then the 1-D if you add RDF(S) 
into the diagram)
   -Jim H.
P.S. Please note that this is NOT feedback against having Owl Lite - 
it's just the opposite, people seem to think Lite is a good idea, 
just having it only for OWL DL is what they question.
p.p.s. (Humour) We even have great names for the two subsets Brian 
McBride suggested we could rename our current sublanguage "owl:dlite" 
and I thought we could name the other "owl:flite" (pronounced 
'flight') -- so we'd have pronounceable subset names :->


-- 
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	  240-731-3822 (Cell)
http://www.cs.umd.edu/users/hendler

Received on Tuesday, 11 February 2003 09:46:42 UTC