[Fwd: Re: ISSUE 5.14: Revised proposal to close ontology versioning]

Dan and I accidentally started this discussion off-list (my fault).
Here's Dan's response to my earlier message today.

-------- Original Message --------
Subject: Re: ISSUE 5.14: Revised proposal to close ontology versioning
Date: 03 Dec 2002 08:58:45 -0600
From: Dan Connolly <connolly@w3.org>
To: Jeff Heflin <heflin@cse.lehigh.edu>
CC: Dan Connolly <connolly@w3.org>

On Tue, 2002-12-03 at 08:32, Jeff Heflin wrote:
> Dan,
> I had originally responded to your suggestions in [1], perhaps you
> missed 
> that message?

Oops; just neglected to respond; sorry...

> In particular, regarding versionOf, I said:
> > While I think that something like this is an excellent idea for tech
> > reports, I think it could be dangerous for ontologies. First, of all
> > what is the generic ontology? Is it just the most recent version, as is
> > done with tech reports? If so, then if people start using this namespace
> > and importing this ontology, what happens when a modification is made
> > that isn't backward compatible?

Folks that link to the latest version accept that risk; the
offsetting benefit is that they get the latest improvements
to the ontology "for free".

> > Essentially, it would "break" existing
> > documents. I'd much rather document authors to be explicit about what
> > version of an ontology they commit to, and put the burden on them to
> > upgrade at whatever pace they feel comfortable with.
> Would you care to comment on this? Perhaps it would help if you gave
> some more details on how you think versionOf would be used in an
> ontology context.

It seems to me that the ontology context is completely parallel
to the spec context: folks that trust W3C to Do The Right Thing
just link to the latest version; folks that want to review
every change link to a specific dated version.

Oops; I just realized this is offlist. Please forward it
or whatever; I'm in the middle of other stuff.

Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Tuesday, 3 December 2002 10:10:42 UTC