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

Re: Review of Guide

From: Jeff Heflin <heflin@cse.lehigh.edu>
Date: Wed, 22 Jan 2003 15:31:56 -0500
Message-ID: <3E2EFFBB.2340536E@cse.lehigh.edu>
To: "Smith, Michael K" <michael.smith@eds.com>
CC: WebOnt <www-webont-wg@w3.org>

"Smith, Michael K" wrote:
> 
> Thanks Jeff.
> 
> Responses below.
> 
> - Mike
> 
> > 1) We need an example of an OWL ontology that consists soley of
> > instances. These will be the most numerous OWL "ontologies" on the Web,
> > so we should at least show people how to do them. I think the easiest
> > way to do this is to take all of the Wine and Winery instances from
> > wine.owl and move them into a third file called winelist.owl. Then in
> > section 3.2.1 (Defining Individuals), we can add something like
> > "Ontologies that describe classes and properties may include
> > descriptions of individuals. However, many ontologies will consist soley
> > of individuals. For example, the wine list of a particular restaurant
> > may be represented as an ontology that describes all of the wines
> > available. In order to include the descriptions of the various styles of
> > wine, this ontology must import wine.owl. An example of this is
> > presented in winelist.owl."
> 
> This was a great suggestion that Mike Dean made earlier, that I meant
> to follow up on and have not had time for.  I can't devote much more time
> to the Guide for the rest of this month, so this change currently has
> low priortiy.
> 

I suppose its not important enough to delay last call. However, I do
think its necessary for Proposed Rec, and will bug you about it again
when we get to that stage. ;-)


> 
> > 4) Replace the first paragraph in the Ontology Versioning section
> > (currently 3.5) with the following two paragraphs. (and note the
> > correction of owl:Ontology capitalization in the 2nd parahraph)
> >
> > Ontologies are like software, they will be maintained and thus will
> > change over time. However, once an ontology has been released, other
> > documents on the Web may come to depend on it (for example, by importing
> > it), and any changes may have significant impacts on these dependent
> > documents. Therefore, when a deployed ontology needs to be updated, the
> > original file should not be modified. Instead, the changes should be
> > made to a copy of the ontology which is assigned a different URL. OWL
> > provides some basic properties to describe relationships between
> > different versions of an ontology. Note that these properties need not
> > be used when changing an ontology that is in its design phase and has
> > not been relased yet. In those situations, there are either no
> > dependencies or the dependencies can be managed internally, so standard
> > document or software version management techniques may be used.
> 
> Seems a little long.  Will come up with revised wording.
> 

I noticed in the "last call" version you decided to omit this paragraph.
I think we need to explain why versioning is important in distributed
systems, and also to describe the general philosophy of versioning in
RDF/OWL (e.g., give each new version of a schema/ontology its own URI).
Once again, this isn't critical enough to hold up last call, but I will
bring it up again come time for Proposed Rec.

Jeff
Received on Wednesday, 22 January 2003 15:32:00 GMT

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