Re: ISSUE 5.6 - daml:imports as magic syntax

From: Jeff Heflin <heflin@cse.lehigh.edu>
Subject: Re: ISSUE 5.6 - daml:imports as magic syntax
Date: Tue, 10 Sep 2002 16:22:04 -0400

> Peter,
> 
> Please see my responses inline below...
> 
> "Peter F. Patel-Schneider" wrote:
> > 
> > A very good summary of daml/owl:imports.
> > 
> > From: Jeff Heflin <heflin@cse.lehigh.edu>

[...]

> > A treatment of imports can be done completely syntactically, by replacing
> > imports foo, where foo is a URI (or whatever) by the contents of the
> > document pointed at by foo.   This is the way I would handle it in the
> > abstract syntax and direct semantics.
> 
> Although imports could be treated syntactically (kind of like a
> "#include" directive), I think that would be a big mistake.  The point
> of imports is that knowledge from another source applies to the resource
> in which it is expressed. The Semantic Web is fundamentally about
> distributed ontologies and data sources, and as such its semantics
> should discuss these things explicitly. A syntactic fix obscures one of
> the things that differentiates the Semantic Web from traditional logic
> approaches. Ontologies and the interrelationships between them are
> important; they aren't just things to be swept under the rug.

Oh agreed, they are important.  However, making ontologies elements of the
domain of discourse opens up a large number if issues, including very
dangerous things like self-reference.  N3 has (or had) something along
these lines, and it is easy then to create the liar's paradox and other,
nastier, problems.

> > > In Pat Hayes' recent OWL model theory, it should
> > > be sufficient to say that owl:imports is not an instance of owl:Object,
> > > owl:DatatypeValue, owl:Class, owl:ObjectProperty, or
> > > owl:DataTypeProperty. However, it would still be an rdfs:Property, but
> > > any RDF graph that uses it as a subject or object would not be
> > > well-formed OWL/RDF. Note, owl:imports doesn't fit into any of the
> > > classifications in the table in section 2 of the Hayes model theory, so
> > > perhaps a new classification called metaproperty or such is needed.
> > 
> > Even in this separated treatment, imports is problematic.  It would require
> > a completely new kind of semantic rule, and one that brings new kinds of
> > things into the semantics.
> 
> Sure this semantic rule is a little different from the kind that we
> already have considered, but I don't think it would break things or make
> them more confusing. In fact, I think it would illucidate how knowledge
> sharing actually works on the Semantic Web.

> > > Semantics:
> > > ----------
> > > It is important that the semantics of owl:imports be added to any
> > > semantics documents. To do this, we need to be able to refer to sets of
> > > OWL statements (such as a web page, a database with an OWL interface,
> > > etc.) This could probably be called a resource, but that term is also
> > > used to describe RDF instances, so for lack of a better term, I will
> > > choose the term graph for the time being. Let graph be a function from a
> > > URI (URL?) to an RDF/OWL graph. Each OWL graph has a set of entailments
> > > that are determined by the model theory. The semantics of a statement:
> > >
> > > A owl:imports B.
> > >
> > > are:
> > >
> > > if graph(B) |= X then graph(A) |= X
> > >
> > > (Note: Here, "|=" is the OWL entailment relation)
> > 
> > This adds in a lot of error-prone machinery.
> 
> What is error prone about it? This simply says that anything entailed by
> the ontology denoted by B is also entailed by the ontology denoted by A.

> > > One question is whether we need something more specific in the model
> > > theory that, for example, uses ICEXT and IEXT.
> > 
> > Something more is needed.  In particular, a relationship between the
> > current graph and its name is needed.  Otherwise how can the condition
> > be discharged in the conditional above?
> 
> The relationship can be implmented by using the HTTP GET function to
> fetch the contents at the URL and then parsing it based on OWL
> semantics. I don't see what the problem is.

Well you now have to incorporate the meaning of the HTTP GET function in
the semantics itself.  How are you planning on doing this?  This is the
sort of machinery that is hard to specify and easy to pervert.

[...]

peter

Received on Tuesday, 10 September 2002 16:41:41 UTC