W3C home > Mailing lists > Public > public-prov-wg@w3.org > October 2012

RE: PROV-ISSUE-521: Data Model Section 5.3.3 [prov-dm]

From: Freimuth, Robert, Ph.D. <Freimuth.Robert@mayo.edu>
Date: Wed, 3 Oct 2012 18:37:21 +0000
To: "public-prov-wg@w3.org" <public-prov-wg@w3.org>
Message-ID: <76A706C559A90249BA321EE35470B85701E202@MSGPEXCEI12A.mfad.mfroot.org>
Hi,

I'm trying to catch up on the flurry of PROV email over the last week.  The first two questions in this issue have been answered.  I would like clarification on the third.

Question:
> Can associations be valid for a window of time for a given activity? For example, several agents are responsible for the activity of preparing and serving food in a restaurant (server, cook, bartender), but not all of them are responsible for the entire time the activity is active. To support this, start and end times should be added to Associations.
Response:
> PROV associations are not temporal relations. Instead, prov-constraints define ordering constraints that are implied by associations. The agent in an association is expected to have some overlap with the activity. Likewise, for attribution, the agent exist before this entity was generated.
Follow up:

How would PROV be used for the example above (preparing and serving food)?  In that case, three entities are responsible for specific portions of an activity.  Is there a way to say that a person acting as a bartender was responsible for only a portion of the activity?  (Defining the association on the basis of role (e.g., cook vs. bartender) seems straightforward, but the temporal aspect does not.)

A different example might involve oversight of (responsibility for) the development of a widget (an activity).  Supervisor A might initiate the development process, but then transfer responsibility to B at a shift change.  If the widget is found to be defective, one might look to the provenance of the widget's production to determine who was responsible at a particular point in time.  How is this expressed with PROV?

If temporal relations are not supported, is it possible to detect logical inconsistencies in provenance?  Validating provenance assertions is likely going to be a common use case, and I would expect temporal validation to be heavily used.  This is similar to temporal restrictions on entities (e.g., cannot be used after it is terminated).

Thanks,
Bob


________________________________
From: public-prov-wg-request@listhub.w3.org [mailto:public-prov-wg-request@listhub.w3.org] On Behalf Of Luc Moreau
Sent: Tuesday, September 25, 2012 8:27 AM
To: public-prov-wg@w3.org
Subject: Re: PROV-ISSUE-521: Data Model Section 5.3.3 [prov-dm]


Hi all,

Find below a draft response to this issue, available on the wiki at:
http://www.w3.org/2011/prov/wiki/ResponsesToPublicComments#ISSUE-521_.28Responsibility_Activity.29

Feedback appreciated,
Regards,
Luc

ISSUE-521 (Responsibility Activity)

  *   Original email: http://lists.w3.org/Archives/Public/public-prov-wg/2012Sep/0111.html
  *   Tracker: http://www.w3.org/2011/prov/track/issues/521
  *   Group Response
     *   PROV agents bear responsibility for activities taking place, entities being generated, and other agents.
     *   PROV agents MAY be entities or activities
     *   Given this, it is legal to write the following, in which the type of a2 and a1 is inferred to be agent.

activity(a1)
activity(a2)
actedOnBehalfOf(a2,a1)


  *
     *   The group has provided its answer to ISSUE-503
     *   PROV associations are not temporal relations. Instead, prov-constraints define ordering constraints that are implied by associations. The agent in an association is expected to have some overlap with the activity. Likewise, for attribution, the agent exist before this entity was generated.
     *   If an application, it is necessary to express that an activity is associated with agent ag1 during interval (evt1-evt2) and then with agent ag2 during interval (evt2-evt3), the approach is to model this with two activities, one for the first interval, or one for the second interval.
  *   References:
     *   Type inference: http://www.w3.org/TR/2012/WD-prov-constraints-20120911/#typing
     *   Adopt plan response: http://www.w3.org/2011/prov/wiki/ResponsesToPublicComments#ISSUE-503_.28adopt_plan.29
     *   Association ordering constraints: http://www.w3.org/TR/2012/WD-prov-constraints-20120911/#wasAssociatedWith-ordering_text
  *   Changes to the document: http://dvcs.w3.org/hg/prov/rev/e89828e2a81c
  *   Original author's acknowledgement:


On 10/09/2012 09:48, Provenance Working Group Issue Tracker wrote:

PROV-ISSUE-521: Data Model Section 5.3.3   [prov-dm]

http://www.w3.org/2011/prov/track/issues/521

Raised by: Luc Moreau
On product: prov-dm


http://www.w3.org/2011/prov/wiki/LC_Feedback#Data_Model_Section_5.3.3

ISSUE-463

By definition, agents can be both entities and activities (section 2.1.3). Can activities be responsible for other activities, or can only entities be assigned responsibility?

Similarly, can activities adopt a plan when acting as an agent, or can only entity agents adopt plans?

Can associations be valid for a window of time for a given activity? For example, several agents are responsible for the activity of preparing and serving food in a restaurant (server, cook, bartender), but not all of them are responsible for the entire time the activity is active. To support this, start and end times should be added to Associations.







--
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
Received on Wednesday, 3 October 2012 18:37:46 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:58:19 UTC