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

Re: Abstract Syntax and Semantics: review comments

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Thu, 26 Dec 2002 14:26:25 -0500 (EST)
Message-Id: <20021226.142625.87605459.pfps@research.bell-labs.com>
To: herman.ter.horst@philips.com
Cc: www-webont-wg@w3.org

From: herman.ter.horst@philips.com
Subject: Re: Abstract Syntax and Semantics: review comments
Date: Tue, 24 Dec 2002 08:46:45 +0100

> > Re: Abstract Syntax and Semantics: review comments
> > From: Peter F. Patel-Schneider (pfps@research.bell-labs.com)
> > Date: Fri, Dec 20 2002 

[...]

> > > As long as such characterizations remain unknown, 
> > > a sequence of simple sufficient conditions for entailment (i.e., simple 
> > > inference rules, formally stated, with proofs of validity) could 
> > > form a useful informative addition to the document in order to 
> > > familiarize people with the various kinds of OWL entailment. 
> > 
> > Again, the generation of such rules is very difficult and prone to error.
> 
> It would be the responsibility of authors (and reviewers) to get it 
> error-free, after which users would have the benefit.

But who would do this?  I'm not going to, as I don't have the time it would
require, and I don't see any advantage from it.

> I am thinking of simpler rules, for example like the ones used implicitly
> in the short example appendix to the OWL Semantics document.

What sort of rules would be ``simpler''?

> Or: several rules for entailment (or non-entailment) appearing in the test 
> document
> could be "lifted" to somewhat more abstract statements in the semantics 
> document. 
> Such rules would not need to form a complete set.

Well, if all you want is *some* rules, then the situation is much simpler.
However, I don't see any advantage to an incomplete set of rules, at least
not in a semantics document. 

> They could help to bridge the gap between the basic theoretical 
> definitions of 
> entailment and practical situations.

Well, perhaps, but I don't see how.

[...]

> > >  - It would also be nice to know how an OWL DL KB is characterized in 
> > > triple form.  That is, what is an "OWL DL graph" as a special kind of 
> > > RDF graph?
> > 
> > This is the subject of section 4 of the document.
> 
> Section 4 defines "OWL DL graphs" (without naming them in this way)
> as the outcome of a translation process from abstract syntax to RDF.
> Since RDF graphs are the exchange syntax, it would be interesting to also 
> have an 
> alternative definition of such OWL DL graphs purely in terms of RDF 
> triples.

I attempted this, and ran into problems.  It would take quite an effort.
Again, anyone is willing to try, but I don't have the time.

> > > - Section 2.1: A small simplification is possible: use the two rules 
> for 
> > > <annotation> also in <directive>, instead of their content. 
> >
> > The rules are somewhat different, so reuse is not easily possible.
> 
> The only difference that I see is that the word "annotation" starts in one 
> case
> with a capital and in the other case with a small letter.

Which makes them different.  Extracting their similarities would, in my
opinion, be more complex than leaving them independant.

> > > - 5.2: In the tables defining the cardinality restrictions: 
> > > instead of card({v:<u,v>...) make it completely explicit with the set 
> > > from which v can be taken:  card({v elementOf ...:<u,v>...) 
> > 
> > This is not needed as card({v:<u,v> in EXT(...)}) delimits the set
> > appropriately.
> 
> I understand what you mean, by means of information that I can guess from 
> the document.  There is a small, formal point, however.
> I propose three very simple additions to the table in the following way:
> {v elementOf IOT : <u,v> in ...} instead of {v: <u,v>...}.
> In this way it is made completely explicit which elements are to be 
> counted
> (in this description of the formal semantics of minCardinality, 
> maxCardinality
> and cardinality), and it is made clear that what you take the cardinality 
> of is 
> actually a set.

OK, OK. 


peter
Received on Thursday, 26 December 2002 14:26:52 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:56 GMT