W3C home > Mailing lists > Public > www-webont-wg@w3.org > December 2002

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

From: Smith, Michael K <michael.smith@eds.com>
Date: Tue, 3 Dec 2002 14:27:31 -0600
Message-ID: <B8E84F4D9F65D411803500508BE322141241AE5B@USPLM207>
To: Jim Hendler <hendler@cs.umd.edu>, Jeff Heflin <heflin@cse.lehigh.edu>
Cc: www-webont-wg@w3.org

Jim,

> Mike - as logical as this may seem, I'm not in favor because the 
> intended meaning isn't clear (I don't mean formal semantics, I mean 
> operational or whatever) -- that is, if I have an ontology which 
> defines 100 classes, and the I create a new version and say
> "" incompatibleWith B
> do I mean none of the classes are the same?  Do I mean 1 class may 
> have changed?  Do I mean 96 classes have changed?

Huh?  I thought we wanted a simple system.  I think its important
to say that a new version is not compatible with a prior version.
It would be nice to say why, but that is a step up in complexity.

> In Jeff's proposal, I assume the following - if I have an old 
> ontology and a new one - if I say
> "" backCompatibleWith B
> then I mean that none of the 100 have changed.

That's great.  I buy that one.

> If I say
> "" PriorVersion A.
> 
> :moose a owl:deprecatedClass.
> :cow a owl:deprecatedCass.

As I have been following the conversation, if B has a prior version A,
and B contains deprectatedClass assertions, this does not mean that it
is incompatible with A.  Quite the contrary, it suggests that :moose
is still compatible with its definition in A, and that IN SOME LATER
VERSION of the ontology (C), it will not be.  Or do I have that wrong?

- Mike


-----Original Message-----
From: Jim Hendler [mailto:hendler@cs.umd.edu]
Sent: Tuesday, December 03, 2002 12:59 PM
To: Smith, Michael K; Jeff Heflin
Cc: www-webont-wg@w3.org
Subject: RE: LANG: Proposal to close ontology versioning (ISSUE 5.14)


At 11:19 AM -0600 12/3/02, Smith, Michael K wrote:
>Normally I think of 'A and B' as being consistent with 'A'. 
>
>While I know we are not providing a formal semantics for versioning, I find
>your interpretation extremely counterintuitive.  Why make it so confusing?
>
>I think there should be a priorVersion, a backCompatibleWith (which implies
>priorVersion), and an incompatibleWith (which also implies priorVersion).
>Something like:
>
>A newVersionOf B
>A compatibleWith B   ( -> A newVersionOf B)
>A incompatibleWith B ( -> A newVersionOf B)
>
>- Mike

Mike - as logical as this may seem, I'm not in favor because the 
intended meaning isn't clear (I don't mean formal semantics, I mean 
operational or whatever) -- that is, if I have an ontology which 
defines 100 classes, and the I create a new version and say
"" incompatibleWith B
do I mean none of the classes are the same?  Do I mean 1 class may 
have changed?  Do I mean 96 classes have changed?

In Jeff's proposal, I assume the following - if I have an old 
ontology and a new one - if I say
"" backCompatibleWith B
then I mean that none of the 100 have changed.

If I say
"" PriorVersion A.

:moose a owl:deprecatedClass.
:cow a owl:deprecatedCass.

then I know exactly what the inconsistencies are and know if I don't 
use these two classes all is fine.  Question - is this backcompat? 
incompat?

so from te experience we've had with SHOE and with DAML, my intuition 
(and that's all it is) is that "incompat" would be much harder to get 
people to use consistently than "backcompatible" would
  -JH

p.s. Jeff - I'd sure like us to use backCompatibleWith and not 
backCompat - as our policy to date has been complete names and we 
should have only one idiom - why introduce a new one just for this


>-----Original Message-----
>From: Jeff Heflin [mailto:heflin@cse.lehigh.edu]
>Sent: Tuesday, December 03, 2002 11:03 AM
>To: Smith, Michael K
>Cc: www-webont-wg@w3.org
>Subject: Re: LANG: Proposal to close ontology versioning (ISSUE 5.14)
>
>
>"Smith, Michael K" wrote:
>>
>>  > As you point out above, my proposal can
>>  > already express these two notions.
>>
>>  How do I say 'A not backCompatWith B' using your terms?
>
>In my terms, A priorVersion B without A backCompatWith B is the same as
>A not backCompatWith B. While A priorVersion B and A backCompatWith B
>means they are compatible. Since a version can only either be compatible
>or incompatible, I only see the need for 1 bit of information.
>
>However, if you think it would make things clearer to have the
>"complement" of backCompatWith, that's fine. It certainly doesn't break
>anything, it just adds another symbol to our vocabulary. What do you
>suggest we call it? incompatWith? Why don't you propose specific wording
>for an addendum to my proposal, and the WG can vote on whether or not to
>add it?
>
>Jeff
>
>>  You stated that if
>>
>>  A priorVersion B
>>
>>  then you should not rely on compatibility.  I don't want to
>>  suggest it, I want to assert that they are definitely not compatible.
>>
>>  While I am fairly neutral regarding the terms you pick,
>>  'priorVersion' suggests a temporal relation between versions
>>  and nothing more.  At least to me.
>>
>>  - Mike
>>
>>  -----Original Message-----
>>  From: Jeff Heflin [mailto:heflin@cse.lehigh.edu]
>>  Sent: Tuesday, December 03, 2002 8:41 AM
>>  To: Smith, Michael K
>>  Cc: www-webont-wg@w3.org
>>  Subject: Re: LANG: Proposal to close ontology versioning (ISSUE 5.14)
>>
>>  "Smith, Michael K" wrote:
>>  >
>>  [snip]
>>  >
>>  > 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, are you suggesting that we use Extends and Replaces instead of
>>  priorVersion and backCompatWith? As you point out above, my proposal can
>  > already express these two notions. I think Extends could be confusing,
>>  because imports already provides a form of ontology extension, but I
>>  would be interested in hearing the opinions of others.
>>
>>  Jeff


-- 
Professor James Hendler				  hendler@cs.umd.edu
Director, Semantic Web and Agent Technologies	  301-405-2696
Maryland Information and Network Dynamics Lab.	  301-405-6707 (Fax)
Univ of Maryland, College Park, MD 20742	  240-731-3822 (Cell)
http://www.cs.umd.edu/users/hendler
Received on Tuesday, 3 December 2002 15:27:57 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:55 GMT