Proposal to Close issue 5.3, Semantic Layering

I realize discussion is still going on with respect to issue 5.3, but 
in an effort to help focus the discussion, I will make a proposal to 
close it. I put my relatively long rationale for this proposal after 
the proposal itself


to again CLOSE issue 5.3, changing the closing text of issue 5.3 to read:

RESOLVED: to close the layering issue (5.3) as described in Consensus 
on semantic layering,  with the change that KB large-OWL-entails C if 
KB-fast-OWL-entails C (but not only if -- Large OWL is now known as 
Owl Full, Fast Owl is now known as OWL DL).

and actioning the editors of S&AS and Test to make the changes needed 
to be consistent with this.  We also action the editor of Reference 
to add a note mentioning that while it is legal in both OWL DL and 
OWL Full to change the class definition of owl:thing, it should be 
done with great care as it can have complex consequences.

  Some background - our current closing is based on a consensus described in

which states among other things that:

  KB large-OWL-entails C  iff  KB Fast-OWL-entails C

(where large owl is now called OWL Full and Fast OWL is now called OWL DL)

which is what we are discussing now.  However, it also states:

MODUS OPERANDI -------------- - Given a knowledge-base, tool1 must 
support the tight semantics   while tool2 may support the liberal 
semantics - On any expression in Fast OWL, these two would agree, 
For any expression outside Fast OWL,   Tool1's behaviour would be 
undefined,   while tool2 would return answer according to liberal 

which I believe is the current state of things.

Furthermore, our closing statement for issue 5.3 states;

Closed as described in Consensus on semantic layering, provided 2 
technical pieces of work can be completed (see minutes).

However, from the minutes and the consensus email (the one cited 
above) one cannot actually determine what the two pieces of technical 
work are!!  However, it is clear from the discussion that one of them 
is likely the iff clause -- so, in fact, out WG decide we would only 
close 5.3 if we could do two things, one of which is the IFF.  We 
didn't achieve that fully, so my proposal above is essentially having 
us go with the original proposal, noting that we didn;t fully achieve 
it the "two technical points" but can live with the current result.

I base this to a large degree on email from Jeremy, representing HP, 
(July, 0007) which says:

At 7:23 PM +0100 7/1/03, Jeremy Carroll wrote:
3) It violates the concensus statement:

      "KB large-OWL-entails C  iff  KB Fast-OWL-entails C"

but its not so clear that it fails to satisfy the original issue.
The original issue required layering on RDF/RDFS which OWL Full
achieves. OWL/DL is, in any case, not part of that original requirement
through not being an RDF/RDFS supserset.
Having entailments over OWL/DL be only a proper subset of the OW

and from Dan (July 0008) which says

I discussed the situation with my colleages in Semantic Web Advanced 
Development at W3C, and we don't see a big problem either.

Other fixes that have been proposed seem to require significantly 
more work in what seems to be a relatively sound design.

I do think we need a caveat which is to let people know that changing 
the owl:thing (and I suspect owl:nothing) classes can cause 
consequents.  We don't have to prohibit it, just to warn people to be 
careful if they do it.

Professor James Hendler
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)      *** NOTE CHANGED CELL NUMBER ***

Received on Thursday, 3 July 2003 09:10:34 UTC