PIL OWL Ontology Meeting 2012-07-09

From Provenance WG Wiki
Jump to: navigation, search

Meeting Information

prov-wg - Modeling Task Force - OWL group telecon

Attendees

  • Tim 
  • Satya
  • David Corsar
  • Stian
  • Stephan
  • Sam Coppens (regrets)
  • Jun (regrets)
  • Khalid (regrets)
  • Daniel (regrets)

Agenda

For the issues that you are assigned:

  • describe the original concern
  • describe any perspectives already expressed
  • recommend next step, or propose a solution

Voting on Thursday to release PROV-O to public

ISSUES

All

Jun

    • extra paragraph for "now that we've qualified, what can we do?" (right at the beginning of the section, after or before the cheat-sheet tables)
      • activity, entity, agent, dictionary point to the objects being qualified
      • hadActivity, hadGeneration, hadUsage hadPlan and hadRole provide the additional statements about the can-be-qualifeid properties, via their corresponding qualify classes or an involvement class
      • TODO: Tim to make sure all are covered.
    • rename prov:activity, prov:entity, prov:agent -> influencerAgent, influencingEntity, influencingActivity ?
      • Tim: anybody have suggestions on rename?
      • Stephan / Satya: too late to rename.
      • Tim: postpone to list.
    • cheat-sheet tables for "what annotation can be put on {Delegation,Derivation,etc} - can you mock this up?

Stian

  • no go on Altruism?

Satya

  • http://www.w3.org/2011/prov/track/issues/377 RL++ justifications appendix
    • currently at http://aquarius.tw.rpi.edu/prov-wg/prov-o#prov-o-owl-profile
    • ... Stian: some domains are "doubled" one for RL and one for non-RL.
    • ... Khalid: why are we defining the superclasses in addition to the union.
    • ... Tim: hadActivity has two domains: Influence and (Delegation or Derivation or Start) --- Note that the latter is a subclass of the former---
    • Luc's LC feedback: "hadRole: why is domain defined as intersection of Influence and six of its subclasses. Why not the subclasses directly?"
    • TODO: include these considerations?
    • Satya: hadActivity's domain: Delegation OR Derivation OR Start AND Influence
    • Tim: The RIGHT is for the RL reasoner, the LEFT is handled okay by the DL reasoner.
    • Satya: it will be different
    • Stian: that's the point, but to be the same, we'd need intermediate classes. RL reasoners skip the union.
    • Satya: given that Influence has so many subclasses. Suggest removing the AND Influence.
    • Stian: adding information that is not really true,
    • Satya: if you say JUST Influence, they you get in trouble.
    • Stian: just by stating domain owl:Thing, that doesn't mean that it can be any arbitrary class that you want.
    • TODO: Stian to first cut and send to Satya.

Tim

  • Stephan review's
    • Plan: We do not mention in Section 3.3 that Plan is a subclass of Entity (and we do mention that Bundle and Collection are subclasses of Entity).  
      • Stephan: we give all but plan, but do not mention it in 3.2. One could infer that Plan is not an entity.
      • ... we don't say why Plan is an Entity. One could describe the provenance of a Plan. So it's not always a qulfieid term.
      • TODO: Tim add the one-liner for Plan in 3.2, add to Figure 2. "one can describe provenancne of a plan".
    • We may want to introduce Plan as a specialization of Entity in Section 3.2 and include it in Figure 2.
    • Role: I think including role information in the Qualified Association example in section 3.3 would help the narrative.  Otherwise, fine.
      • Stephan: dind't see Role in section 3.3. Along with HadPlan, add a Role.
      • TODO: Tim to add hadRole to the example, plus the narrative of the example.
  • review of prov-o july 3 2012 for last call (and its 104 raised feedback points)
    • qualifiedXXX: shouldn't they be inverseFunctional? (luc's comment)
      • Otherwise, this would allow for a given Influence instance, to be a qualified Influence for multiple subjects. This is not intended.
      •   The qualified pattern is prov-o specific. It was inverse functional before, but I think this characteristic was incorrectly removed.
      • Stephan: not against it, but goal is to refelct the DM, so the DM should assert that it is inverse functional.
    • influencer: should it be functional: there is only one influencer per qualified pattern instance, isn't there.
    • hadPlan: is functional; hadUsage: is functional; hadGeneration: is functional; hadActivity: is functional
      • TODO: Satya to help skim through, Stephan to skim but busy, David.
    • cross references links to DM, constraints, and N. (remove all but DM?)
      • satya: remove to prov-n
      • ... for constraints
      • stephan: fine with removign to prov-n, since we're not trying to model the constraints
      • david: sounds reasonable, constraints
    • diagrams for examples (what is PROV conventions?)
      • TODO: David to give it a shot.
  • Luc's scan
  • RDFa is bunk

Stephan

Khalid

  • highlighting focus terms in prov-o examples
    • with strong HTML elements

Daniel

  • The only 2 things that I have detected is that in section 3.2 there is no example for mention and asInBundle (even though the example has bundles) and there is no wasInfluencedBy in the example (it could be easily added as an inference of wasAttributedTo).
  • Rename chart maker, make consistent.
    • - I wonder if a better example organization instead of ex:chartgen would be "National Newspaper"
    • - I think the team was already looking at the consistency of the examples. It was Chart Generators and now in this example it's Chart Generators Inc
  • it was suggested for prov-dm that examples should be described in past tense. It should be done here too.

Backburner

AOB