- From: Paolo Missier <Paolo.Missier@ncl.ac.uk>
- Date: Wed, 27 Jul 2011 18:32:52 +0100
- To: public-prov-wg@w3.org
Paul participation/control -> Interesting distinction. I agree it's an important point. Two comments: - this seems to cut through an axis that is different from "agent as human / agent as machine" etc. , and probably orthogonal to it - I vaguely recall a recent talk somewhere where the focus was on provenance analysis in which these roles (and more) were explicit, aimed at reconstructing the hierarchy (business, chain of responsibility etc.) of the organization that was collectively responsible for the production of a document. Anyone recalls this? -Paolo On 7/27/11 4:45 PM, Paul Groth wrote: > Hi Luc, > > I think the examples in the documents are control. Each of the agents > clearly have some control over the process execution (David controls the > editing of crime.txt). > > I think participation is weaker where you can say that David > participated in the meeting but Alice controlled the meeting. > > Somehow, control is saying that a process execution wouldn't have > happened without an Agent. Whereas participation implies that the > process execution could still have happened but the participation of an > Agent changed how that process execution happened. > > Thanks, > Paul > > Luc Moreau wrote: >> Hi Paul, >> >> Can you try and elaborate on this. It's quite an important point. >> >> Currently, with the definition outlined here, participates is a kind of >> abbreviation, >> for core concepts. It's useful since it facilitates the writing of >> provenance assertions, >> when full details are not known by the asserter. >> >> However, you seem to indicate there is a stronger meaning for control. >> Do you >> see the examples of control in the document as control, or as participation? >> >> Regards, >> Luc >> >> On 07/27/2011 08:44 AM, Paul Groth wrote: >>> Looking at the definition of isControlledBy this is what I think participates means. >>> >>> I would think that controlled by has a stronger meaning than its current definition but I can't articulate it :-) so I'll leave it at that. >>> >>> Paul >>> >>> >>> On Jul 27, 2011, at 9:23 AM, Luc Moreau<L.Moreau@ecs.soton.ac.uk> wrote: >>> >>> >>>> Hi, >>>> >>>> I have a few questions/comments about this notion of participation. >>>> >>>> 1. Are we then saying the following? >>>> participates(pe,x) if uses(pe,x) or isControlledBy(pe,x) >>>> >>>> 2. Paul, what do you mean by more general version of control? Do you >>>> mean more >>>> than the above? If yes, can you clarify? >>>> >>>> 3. Satya, what do you mean by 'plays some other role in a Process >>>> execution'? >>>> is it control? something else? is it role as defined in the >>>> specification draft? >>>> >>>> 4. Do you want 'participates' to be something that is inferred, or >>>> something that >>>> is asserted, or both? >>>> >>>> 5. Jim Myers also suggested something along the lines of: >>>> >>>> if participates(pe,x) and ivpOf(x,y) then participates(pe,y) >>>> >>>> Is it also in line with what you think participation is. >>>> >>>> Regards, >>>> Luc >>>> >>>> On 07/27/2011 07:53 AM, Paul Groth wrote: >>>> >>>>> +1 for participation being with respect to agents and being a more general version of control. >>>>> >>>>> Paul >>>>> >>>>> On Jul 27, 2011, at 4:19, Stephan Zednik<zednis@rpi.edu> wrote: >>>>> >>>>> >>>>> >>>>>> I would argue against Participation as a parent property of >>>>>> isGeneratedBy. I do not think that being the result of a process >>>>>> execution entails participation. Does a cake participate in getting >>>>>> baked? Does 'book now owned by person b' participate in getting sold >>>>>> from person A to person B? >>>>>> >>>>>> For that matter, I am not sure I am happy with Participation as a parent >>>>>> property of isUsedBy. When I think of participation I think of agents. >>>>>> Agents participate, non-agents get used. If a thing with no agency is >>>>>> acted upon in a process execution, is it participating? >>>>>> >>>>>> I would agree that Participation is a parent property to >>>>>> isControlledBy. It would seem logical to argue that an agent that >>>>>> controls a process execution, participates in the process execution. >>>>>> >>>>>> --Stephan >>>>>> >>>>>> On 7/26/2011 7:29 PM, Provenance Working Group Issue Tracker wrote: >>>>>> >>>>>> >>>>>>> PROV-ISSUE-49 (Participation): Suggested definition for Participation [Conceptual Model] >>>>>>> >>>>>>> http://www.w3.org/2011/prov/track/issues/49 >>>>>>> >>>>>>> Raised by: Satya Sahoo >>>>>>> On product: Conceptual Model >>>>>>> >>>>>>> Participation is a property linking BOB with Process execution,where BOB isUsedBy or isGeneratedBy or plays some other role in a Process execution. >>>>>>> >>>>>>> This definition makes participation a parent property of both isUsedBy and isGeneratedBy properties. Also, this helps cover scenarios where BOB helps in execution of a process but is not consumed by or generated by process execution (it can be re-used in another process execution) >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>> -- >>>> 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 >>>> United Kingdom http://www.ecs.soton.ac.uk/~lavm >>>> >>>> >>>> -- ----------- ~oo~ -------------- Paolo Missier - Paolo.Missier@newcastle.ac.uk, pmissier@acm.org School of Computing Science, Newcastle University, UK http://www.cs.ncl.ac.uk/people/Paolo.Missier
Received on Wednesday, 27 July 2011 17:33:33 UTC