W3C home > Mailing lists > Public > www-webont-wg@w3.org > October 2002

Re: outline of semantics document

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Tue, 29 Oct 2002 07:26:34 -0500 (EST)
Message-Id: <20021029.072634.92197510.pfps@research.bell-labs.com>
To: www-webont-wg@w3.org
Cc: jjc@hpl.hp.com

From: Jeremy Carroll <jjc@hpl.hp.com>
Subject: [Moderator Action] Re: outline of semantics document
Date: Mon, 28 Oct 2002 14:29:41 -0500 (EST)

> > The outline and, in fact, the contents, can be found at 
> 
> >   http://www-db.research.bell-labs.com/user/pfps/owl/semantics.html
> 
> > or by following the links in that document.
> 
> 
> Hi
> 
> some quick high level comments ... I don't expect to look at the detail for a 
> week or two.
> 
> 1: I am surprised that the proof in the appendix is "normative" - a quick 
> glance did not reveal quite what that was intended to mean.

Hmm.  I now think that the appendix should not be normative.  I'll change it.

> 2: I was surprised at the use of N-triples (in section 4)

What surprised you?  This section (the mapping from the abstract syntax to
the RDF syntax) has been around for quite some time.  I'm not interested in
doing the mental conniptions to translate from the abstract syntax to
RDF/XML, particularly as all the RDF information in an RDF/XML document is
captured in the n-triples.

> I note that: the use of q-names is not actually allowed, (in the OWL test 
> cases I think I specifically use N-triple augmented with q-names), and that 
> you omit discussion of all the literal escaping stuff.

If the QNames bit is all that is worrying you, I can easily go from
n-triples to n-triples+QNames. The literal stuff is, of course, incomplete,
as RDF Core has not finished with datatyping.

> I would more have expected the use of triples and the graph abstract syntax.
> I believe that the phrasing in the RDF Model Theory has avoided these pitfalls 
> by being somehow 'clearly' at a more abstract level.
> i.e. it would be mistaken to respond to these comments by fixing ntriple, or 
> avoiding qnames in the rules, or by defining escaping algorithms - all these 
> isses are uninteresting at the OWL level.

> Are you looking for detailed review at this stage? If so I think I will be 
> able to for 11th Nov.

I have no idea whether detailed reviews are being solicited yet.  I expect
that this will be discussed at the teleconference this week.  I am, of
course, always interested in (non-destructive) comments.

> Jeremy

peter
Received on Tuesday, 29 October 2002 07:26:44 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:53 GMT