proposal for last session of July face-to-face

As I was reading Guus's agenda for the face-to-face, it occured to me that
there is very little information there about planning for the future.
There is a session partly reserved for this, but no background or plan is
given for this session.  I feel that without some background and a plan,
this session will not be productive.

My proposal is that this session should be mostly devoted to planning with
respect to finishing the definition of OWL.  The documents currently being
developed in the working group do not form a complete definition of OWL and
the missing sections need to be filled in very soon.


I am thus providing a proposal for filling in these missing sections.  


As I see it, there are the following documents underway in the working
group:


		Current OWL Documents

Name	Description		Author(s)/Editor(s)		

1 Requirements document		Heflin
	Listing of requirements for OWL

2 Issues document		Michael Smith
	Listing of issues for the WG

3 OWL Feature synopsis 		McGuinness et al
	High-level, informal description of OWL features

4 OWL Reference Description	Dean
	Informal description of OWL KBs in triple form

5 OWL Formal specification 	Patel-Schneider et al.
	Formal specification of abstract syntax of OWL
	Informal description of meaning of abstract syntax

6 OWL XML Presentation Syntax	Patel-Schneider

7 OWL UML Presentation Syntax	Schreiber


These leave three portions of OWL unspecified:

M1 What is an OWL KB in triple form, and, more importantly, what
   collections of triples are not OWL KBs?

M2 What is the translation from the abstract syntax to triples?

M3 What is the formal meaning of an OWL KB?


Jeremy Carroll has made a proposal that includes the idea that not all
collections of triples are OWL KBs.  I strongly support this idea and,
moreover, *propose* that the definition of just what constitutes an OWL KB
in triple form be specified as the result of a mapping from the abstract
syntax, something like the mapping in

	http://www-db.research.bell-labs.com/user/pfps/owl/translation.html

This document provides both a mapping from the abstract syntax to triples,
filling in M2, and (at the very end) a definition of what collections of
triples constitute an OWL KB, filling in M1.


As far as M2 (the formal meaning of an OWL KB) goes, I have previously sent
out a semantics for OWL KBs in abstract syntax form.  This document is
available at

	http://www-db.research.bell-labs.com/user/pfps/owl/semantics.html

I *propose* that the meaning of OWL KBs be specified from the
abstract syntax.  OWL KBs in other forms would be given meaning by
translation back to the abstract syntax.


These two additions would, I think, complete the set of documents needed
for a definition of OWL.  Of course, much work needs to be done to update,
revise, and harmonize these documents, but I feel that the planning session
of the upcoming face-to-face would be a good time to at least start
discussion on how to proceed on filling in the holes in the definition of
OWL.


Peter F. Patel-Schneider


PS: Note that I have not said anything about layering issues.  One way of
    handling layering would be to make some of the triples in an OWL
    ontology RDF-dark.

Received on Wednesday, 19 June 2002 10:05:48 UTC