W3C home > Mailing lists > Public > www-webont-wg@w3.org > January 2003

Re: Review of Guide

From: Jeff Heflin <heflin@cse.lehigh.edu>
Date: Mon, 06 Jan 2003 10:43:34 -0500
Message-ID: <3E19A426.454666D8@cse.lehigh.edu>
To: "Smith, Michael K" <michael.smith@eds.com>
CC: WebOnt <www-webont-wg@w3.org>

Mike,

Thanks for your careful consideration of my somewhat copious comments.
Below are a few responses to specific issues/questions you raised.

Jeff

"Smith, Michael K" wrote:
> 
----------------------------------------------------------------------------
> ----
> > - Ontology headers, para. 3, "... the owl:Ontology tag simply identifies
> > a document as containing OWL syntax.":  I disagree! In order to
> > determine if a document contains OWL syntax, find out if it uses names
> > from the OWL namespace. The owl:Ontology tag makes an assertion that the
> > current document is an ontology. Now, this "ontology" may or may not
> > have classes and properties, and may just consist of instances, but the
> > author is stating that they think there is vocabulary worth reusing
> > here.
> 
> You are right.  This is funny. But an owl document does not have to use
> names from the OWL namespace.  I think of owl:Ontology as a statement
> of intent.  Perhaps  "... the owl:Ontology tag simply identifies
> a document as an OWL document."  ?

This is personal opinion, but I don't think that is correct either. If
we were just trying to identify an OWL document then the tag should be
owl:OWL and it should be mandatory in all OWL documents. I still believe
that the owl:Ontology tag means that the author is asserting that the
document is an ontology. Now of course this may mean that somebody is
calling something an ontology that many of us would not consider an
ontology (for example a document that contains only instances), and
that's perfectly fine, just like somebody may claim that they are the
King of England.


----------------------------------------------------------------------------
> ----
> > - Namespaces, para. 2, "The default namespace ... unprefixed elements
> > and empty URI references refer to the current ontology": I don't believe
> > that namespaces say anything about empty URI references. XML says they
> > are relative to the current document.
> 
> In XML they do.  If my namespace declarations include
> 
>     xmlns="http://www.example.org/wine#"
> 
> Then <Wine> expands to <http://www.example.org/wine#wine>.
> 

That looks like an unprefixed element to me, and not an empty URI
reference. I though an empty URI reference was something like
rdf:resource="", where an attribute is supposed to take a URI as a
value. Is this a misunderstanding of XML terminology on my part? If not,
maybe some clarification is needed here to prevent other people from
being similarly confused.

>  ----------------------------------------------------------------------------
> ----
> > - Functional Property: mention that FunctionalProperty is equivalent to
> > maxCardinality = 1 when used a restriction on owl:Thing
> 
> Huh.
> 

FunctionalProperty is basically like saying that there is a
maxCardinality of 1. I think it is useful to point this out. However,
there is one wrinkle: FunctionalProperty is global, while maxCardinality
is local to class. To make maxCardinality "global" you can use it as a
restriction on owl:Thing, the class of all instances.
Received on Monday, 6 January 2003 10:43:37 GMT

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