RE: PROV-ISSUE-530: Data Model Section 5.7.2 (Table 6) [prov-dm]

Hi,

Thanks for your comments, Paolo.  You picked up on the core issue here - the use of user-defined attributes to extend the spec for the same semantic reason (e.g., expression of time) but using different methods to do so.

It would not surprise me if a user wanted to assign time, location, and/or role information to a broader set of relationships than those that currently take those attributes.  As I mentioned, Association and Delegation seem like logical candidates.  If the relationship definition does not support these attributes, it would be trivial for a PROV user to simply provide that information within user-defined attributes.

There may be cases where a given attribute doesn't make sense for a given relationship.  I believe those cases are rare, and in my opinion the advantage of pre-specifying common attributes and maintaining a higher level of interoperability outweighs the slight loosening of the spec.  It is hard to predict all possible use cases, so unless there are very strong reasons against it, I would lean toward including a set of "core qualifying attributes" in all relationships.

Thanks,
Bob


________________________________
From: public-prov-wg-request@listhub.w3.org [mailto:public-prov-wg-request@listhub.w3.org] On Behalf Of Paolo Missier
Sent: Tuesday, September 25, 2012 8:57 AM
To: public-prov-wg@w3.org
Cc: Provenance Working Group; Provenance Working Group Issue Tracker
Subject: Re: PROV-ISSUE-530: Data Model Section 5.7.2 (Table 6) [prov-dm]

Hi,

I wouldn't be so quickly dismissive of this comment. Given that we have a "loose" semantics for attributes, I find it hard to argue for or against using those same attributes in unanticipated ways.
I am wary of this comment:  "these restrictions could be circumvented using user-defined optional attributes if needed, at the expense of maintaining a single standard representation for the information".
This is the interop argument again -- the possibility that we see new attributes defined to mimick existing ones but with a different namespace and in a different context.
  So I would like to hear more comments from the group on this.

As a minor point, I don't think this is a good argument to mention:
"No other external comments made a request for allowing such attributes in more concepts."
I believe the merit of comments should not be based by their popularity amongs the public...

--Paolo



 On 9/24/12 7:46 AM, Luc Moreau wrote:
Dear all,

Find a proposed response for this in the wiki at:
http://www.w3.org/2011/prov/wiki/ResponsesToPublicComments#ISSUE-530_.28attributes.29
For convenience, it is copied below.
Comments, feedback?
Best regards,
Luc


ISSUE-530 (attributes)

  *   Original email: http://lists.w3.org/Archives/Public/public-prov-wg/2012Sep/0120.html
  *   Tracker: http://www.w3.org/2011/prov/track/issues/530
  *   Group Response
     *   The group has given careful considerations to attributes in prov-dm, specifically time, location and role.
     *   The group could not reach consensus to allow these attributes to more concepts of the data model.
     *   No other external comments made a request for allowing such attributes in more concepts.
     *   Role:
        *   We have already elaborated on roles in our response to ISSUE-532
     *   Location:
        *   While a notion of location is fairly intuitive for an activity or entity, it is less intuitive for associations for instance. In an association, the activity may have a location, and the agent may have a location. It is however unclear what the location of the association itself may be.
     *   Time:
        *   The same comments apply for time. However, in this case, the constraints document explains what kind of ordering constraints exist, between an agent and activity, for instance.
  *   So overall, the group could not find consensus to broaden these attributes to other relations in a meaningful manner. Given implementation, using the PROV extension mechanism, are however able to add similar attributes for their specific needs.
  *   References:
     *   Roles: http://www.w3.org/2011/prov/wiki/ResponsesToPublicComments#ISSUE-532_.28Role.29
     *   Resolution on roles: http://www.w3.org/2011/prov/meeting/2012-06-07#resolution_2
     *   Time constraints: http://lists.w3.org/Archives/Public/public-prov-wg/2012Sep/0120.html
  *   Proposed changes:none
  *   Original author's acknowledgement:



On 10/09/12 09:54, Provenance Working Group Issue Tracker wrote:

PROV-ISSUE-530: Data Model Section 5.7.2 (Table 6)   [prov-dm]

http://www.w3.org/2011/prov/track/issues/530

Raised by: Luc Moreau
On product: prov-dm



http://www.w3.org/2011/prov/wiki/LC_Feedback#Data_Model_Section_5.7.2_.28Table_6.29

ISSUE-463

The restrictions on when time, location and role can be used should be reviewed after the public feedback period closes and changes are made to the model. In particular, I think there is justification for allowing other relationships, such as Association and Delegation, to take these attributes. The model would be more flexible without these restrictions (which could be circumvented using user-defined optional attributes if needed, at the expense of maintaining a single standard representation for the information).









--
Professor Luc Moreau
Electronics and Computer Science   tel:   +44 23 8059 4487
University of Southampton          fax:   +44 23 8059 2865
Southampton SO17 1BJ               email: l.moreau@ecs.soton.ac.uk<mailto:l.moreau@ecs.soton.ac.uk>
United Kingdom                     http://www.ecs.soton.ac.uk/~lavm<http://www.ecs.soton.ac.uk/%7Elavm>





--
-----------  ~oo~  --------------
Paolo Missier - Paolo.Missier@newcastle.ac.uk<mailto:Paolo.Missier@newcastle.ac.uk>, pmissier@acm.org<mailto:pmissier@acm.org>
School of Computing Science, Newcastle University,  UK
http://www.cs.ncl.ac.uk/people/Paolo.Missier