RE: LANG: Proposal to close ontology versioning (ISSUE 5.14)

> Perhaps what you mean is that we can't say that a
> new version is incompatible with a prior version.

Yes.  And why would "priorVersion" imply incompatibility?  Not obvious from
the label.  It is not the case that 'A backCompatWith B' and 'A priorVersion
B' are contradictory, right?  

> I consider this to be
> the default: If backward-compatibility is not explicitly stated, then
> incompatibility should be assumed.

Why assume?  Maybe that's just the careful thing to do, but I would like to
be able to be explicit.  I would be happy if priorVersion explicitly stated
that compatibility is intended not to hold.

A new version can have 2 relation to the earlier one.  It is either upwardly
compatible or not.  All of the old entailments hold or they do not. 

Extends = priorVersion and backCompatWith. The new version is compatible.
All of the old entailments hold.

Replaces = priorVersion and not backCompatwith. In the new version it is not
the case that all of the old entailments hold.

Of course these are just statements of intention, with no logical force. 

- Mike

-----Original Message-----
From: Jeff Heflin [mailto:heflin@cse.lehigh.edu]
Sent: Tuesday, November 26, 2002 9:50 AM
To: Smith, Michael K
Cc: www-webont-wg@w3.org
Subject: Re: LANG: Proposal to close ontology versioning (ISSUE 5.14)


Mike,

I'm not sure what you mean by extends vs. replaces. I see imports as a
way of "extending an ontology." A new version on the other hand is a
kind of "replacement." However, since we are in a distributed
environment we don't want to technically replace an ontology, beause
that may break dependencies on the prior version. So the best we can do
is say such-and-such is the latest version, and whether or not it is
backward-compatibile. Perhaps what you mean is that we can't say that a
new version is incompatible with a prior version. I consider this to be
the default: If backward-compatibility is not explicitly stated, then
incompatibility should be assumed.

If I have completely missed your point, could you explain what you mean
by extends and replaces?

Jeff

"Smith, Michael K" wrote:
> 
> Jeff,
> 
> I am pretty neutral on this, but it doesn't seem like the first two get to
> the important
> relations, which are whether one ontology is a COMPATIBLE extension or
> and INCOMPATIBLE extension.  I.e EXTENDS vs. REPLACES.
> 
>         I propose to add the following identifiers to the OWL namespace:
>         priorVersion
>         backCompatWith
>         deprecatedClass
>         deprecatedProperty
> 
> We can't say REPLACES.  The combination of priorVersion +
backCompatibleWith
> 
> implies EXTENDS.
> 
> An interesting feature is that an ontology can say that it is a
priorVersion
> of some
> other one.  Right?
> 
> - Mike
> 
> Michael K. Smith, Ph.D., P.E.
> EDS - Austin Innovation Centre
> 98 San Jacinto, #500
> Austin, TX  78701
> 
> * phone: +01-512-404-6683
> * mailto:michael.smith@eds.com

Received on Tuesday, 26 November 2002 17:36:41 UTC