RE: Regarding the definition of IVP OF

To respond to a couple of the recent emails at once:

 

I think I agree with Khalid that being precise about the property-level
meaning of IVP is going to be tricky and that ultimately it will have to
be an assertion rather than something we can infer. Although I'd like to
see some guidance on how to use IVP that includes talking about the idea
of invariant properties, I would be OK if those were not formal parts of
the definition.

 

Regarding a bi-directional IVP of relationship - although I think it
will be rare, I think it does occur and would like to see it be a
possibility. Even in the document/file case, I think I could talk about
the document lifecycle where at one point a particular file was a view
of it with a particular content/state, while I could also think of a
temp.txt file lifecycle that at different points in its lifecycle had
different documents that were views on the file content/state. 

 

To me, the general problem is that of crossing ontologies - there are
different ways of viewing the world that 'line -up' at certain points so
you can relate instances in one ontology to those in another. In
practice, I think a lot of the ontologies people want to connect
(especially in provenance) are hierarchical (in the sense of
representing an 'X' and the concept of X-in-state/context-Y') or pseudo
hierarchical (looks hierarchical but there are some processes that break
the correspondence and make it clear that one concept is not a strict
state of another - files aligning with versions of a document, but you
could edit the file to produce a different document so you can't infer
from an 'edit' operation that a new file is also an IVP of the same
document.). And then, if you push far enough, I think you get to cases
where the objects from both ontologies have separate, interesting
lifecycles that we want to record - I tried to cast the document/file
example this way above, but I think it is even clearer in some of the
discussions of an egg versus the set of molecules in the egg, the
physical and legal ship, etc.

 

So - personal opinion again - I'd like to see IVP be something that we
allow to be asserted in both directions, but I would be happy to see
that downplayed because I think it is a relatively rare and fairly
confusing use case but an important one in some areas (the times when we
might really have lots of provenance in the two ontologies being mapped
and have important questions that depend on them might be rare or
specialized but I suspect that a lot of scientific experiments really
have some of this at their heart and I'd like to keep the power to
express them in the model.)

 

  Jim

 

From: public-prov-wg-request@w3.org
[mailto:public-prov-wg-request@w3.org] On Behalf Of Cresswell, Stephen
Sent: Friday, July 08, 2011 11:18 AM
To: Khalid Belhajjame; public-prov-wg@w3.org
Subject: RE: Regarding the definition of IVP OF

 

 

I have another issue with the current definition of "IPV of".

As it currently stands, I believe that it does not exclude the
possibility that two bobs may be mutually "IVP of" each other -
i.e. you could have bobs A, B such that (B IVPof A) & (A IVPof B), and
this is surely not intended.

This could arise if, for bobs A, B :
- A and B both represent the same entity
- A and B share some immutable properties, and they have corresponding
values.
- B has some immutable properties which correspond to mutable properties
of A
- A has some immutable properties which correspond to mutable properties
of B

Possibly the asserter-defined test (included in "IPV of" definition)
that real world states modelled by A and B are "consistent" may disallow
"IPV of" in this situation.  However, unless that is guaranteed, I think
that the definition of "B IPV of A" (if it is still to have a
definition) should additionally require that:
"A has no immutable properties which correspond to mutable properties of
B"

Stephen

-----Original Message-----
From: public-prov-wg-request@w3.org on behalf of Khalid Belhajjame
Sent: Fri 08/07/2011 16:01
To: public-prov-wg@w3.org
Subject: Regarding the definition of IVP OF


During the F2F meeting, there was a discussion in the second day
regarding "IVP of". The definition that was suggested during the F2F can
be found in [1]. In my opinion, the definition of "IVP of" should be
simplified. Specifically, I would prefer a definition that states that
"IVP of" is an asserted relationship between two entity states. I list
in what follows the reasons:

(i) In the definition of "IVP of" [1], the conditions on the properties
of the two entity states A and B (i.e., that the properties the entity
states share must have corresponding values, and that some mutable
properties of A correspond to some immutable properties of B), are not
enough in order to infer that B is an IVP of A. This is because there is
a third condition that is not included, because it is hard to formally
specify, viz. A and B are consistent.

(ii) A consequence of (i), is that we will not be able to automatically
infer that an entity state B is an IVP of another entity state B. All we
can safely do, is identify cases in which an entity state B cannot be an
IVP of another entity state of A.

(iii) Even if we find a means for formally specifying that two entity
states A and B are consistent, e.g., using assertions, it will be
difficult to use the definition of IVP of to make inference. This is
because the definition of IVP of requires correspondences between the
properties of two entity states to be specified. These correspondences
can be complex many-to-many mappings that may turn out to be hard to
encode using existing semantic web technologies.

Thanks, khalid



________________________________________________________________________
This e-mail has been scanned for all viruses by Star. The
service is powered by MessageLabs. For more information on a proactive
anti-virus service working around the clock, around the globe, visit:
http://www.star.net.uk
________________________________________________________________________

************************************************************************
***********************

This email, including any attachment, is confidential and may be legally
privileged. If you are not the intended recipient or if you have
received this email in error, please inform the sender immediately by
reply and delete all copies from your system. Do not retain, copy,
disclose, distribute or otherwise use any of its contents. 

 

Whilst we have taken reasonable precautions to ensure that this email
has been swept for computer viruses, we cannot guarantee that this email
does not contain such material and we therefore advise you to carry out
your own virus checks. We do not accept liability for any damage or
losses sustained as a result of such material.

 

Please note that incoming and outgoing email communications passing
through our IT systems may be monitored and/or intercepted by us solely
to determine whether the content is business related and compliant with
company standards.

************************************************************************
***********************

The Stationery Office Limited is registered in England No. 3049649 at 10
Eastbourne Terrace, London, W2 6LG