Warning:
This wiki has been archived and is now read-only.

PIL OWL Ontology Meeting 2012-01-16

From Provenance WG Wiki
Jump to: navigation, search

Meeting Information

prov-wg - Modeling Task Force - OWL group telecon

Agenda

  • blogging PROV-O
  • 1. Remaining issues to map updates in PROV-DM second PWD to PROV-O
  • 2. Best practices document

Attendees

  • Satya
  • Daniel
  • Khalid
  • Tim
  • Stephan
  • Stian
  • James (joined 1730gmt)

Discussions

Blogging about PROV-O

  • marketing the DM/PAQ/PROV-O/primer
  • marketing the PROV-O (specifically)

Describe simple examples in individual blogs? So people can see simple examples without digging into the docs.

blogs not part of the HTML doc, they stand alone.

Post to official w3c blog?

bite sized portions of PROV-O in a single blog, which references the full prov-wg docs.

Who was giving the feedback?

Pick individual morsels (e.g., one construct) - make an example and describe it. Then point to the wg documents.

http://www.w3.org/2011/prov/wiki/PROV_OWL_ontology_component_examples

Tim would rather have something better to advertise than to be distracted by advertising.

  Rest o' the llst

http://www.w3.org/2011/prov/wiki/PIL_OWL_Ontology

http://www.w3.org/2011/prov/wiki/PIL_OWL_Ontology#Outstanding_issues_as_of_Dec_2011

•  alternateOf and specializationOf properties are missing in prov-o. They replace wasComplementOf.  •  ProvenanceContainer should be renamed RecordContainer. Also,  according to prov-dm a record container should not be defined as a  prov-dm record (i.e, we should not include it in the ontology), because  otherwise it could appear arbitrarily nested inside accounts.   •  I think we should remove this term (Daniel G)  • •  tracedTo relationship missing.  •  wasControlledBy, wasEventuallyDerivedFrom, dependedOn,  hadParticipant and wasScheduledAfter properties should be removed from  prov-o. 


Issue: Provenance Container=

Daniel: Accounts to replace provenance container Stian: Need both since they 

  • recordContainer ::= 'container' namespaceDeclarations ( record ) + 'endContainer '
  • "they start, specify some namespaces, specify some records, and then they stop"

RecordContainer is analogous to GraphSerialization (RDF 1.1 WG's term)

  • This is more mechanical and concrete (analogous to frbr:Manifestation)
  • It is the basis (starting point) for a machine to begin processing.

Account is analogous to (RDF Abstract) Graph (RDF 1.1 WG's term)

  • This is more abstract (analogous to frbr:Expressions)

A record container is a  house-keeping construct of PROV-DM, also capable of bundling PROV-DM  records. A record container is the root of a provenance record and can  be exploited to package up prov-dm records in response to a request for  the provenance of something ([PROV-PAQ]). Given that a record container is the root of a provenance record, it is not defined as a PROV-DM record (production record), since otherwise it could appear arbitrarily nested inside accounts.

Khalid: Account has records that are consistent, whereas record container may have records that may not be internally consistent Daniel: Can we have an orphan account (where there is no explicit asserter)

RESOLVED: RecordContainer does not need to be modeled in PROV-O. It is the AWWW's "resource representation".

write on your own line!

WrittenPagesOnPulp rdfs:subClassOf prov:Account .
WrittenPagesOnPulp owl:disjointWith rdf:Graph .

accountRecord ::=   account  (  identifier  ,  asserter  ,  record optional-attribute-values)  

The constraint for having an asserter for an account needs to be relaxed in PROV-O (open world assumption)

Issue: tracedTo relationship=

Khalid: replacement for dependedOn

"A traceability record states the  existence of  a  "dependency path" between two entities, indicating that  one entity can be shown to be in the lineage of another, and may have  influenced it, or may bear some responsibility for it, in some way. A  traceability record subsumes derivation, activity association, and  responsibility, and is defined to be transitive.         A traceability record, written tracedTo(id,e2,e1,attrs) in PROV-ASN: • id:  an optional identifier  id identifying the traceability record; • entity:  an identifier e2 identifying an entity;  • ancestor: an identifier e1 identifying an ancestor entity in the lineage of  e2;  • attributes: an optional set attrs of attribute-value pairs to further describe this record."

Sweet - I get credit for waht Daniel does!   Get that phd, Daniel!  :D :D

So need a new hadQualifiedTrace -> Trace 

  1. We call it Generation etc. # I like it "Trace"
  2. Can we put some Dick Tracy references in here? Yes.
entity_2

 a prov:Entity;  prov:tracedTo :entity_1  prov:hadQualifiedTraceTo [         a prov:Trace;         # qualifies the trace from :entity_2 to the entity :entity_1         prov:qualifiedEntity :entity_1;         :attribute "value";         :attribute2 "value2";      ]; .

  1. entity_2 is playing the role of the activity.. so you are right, :entity_1 is the qualifiedEntity

                                                        ^^^ in the "analogy"; QualifiedInvolvement was generalized to include between two entities.

Outcome: agree to replace dependedOn in Prov-O

Issue to raise: Are the attributes in tracedTo transitive? - Raised by Stephan as Issue-218 Stian's view: Attributes should not be transitive - they are like making a subproperty of traceTo and might apply only for the big trace - but yes, this should be explicit. I agree with your view Stian. In case, tracedTo is inferred, then the attribute list attached to the inferred tracedTo is empty/unknown. (which is the same). I agree too. The attributes are not transitive. The property is.


=wasControlledBy, wasEventuallyDerivedFrom, dependedOn, hadParticipant  and wasScheduledAfter properties should be removed from prov-o.

Agree to replace these points

Mechanism (i.e. prov:Plan):

  • Check out from Mercurial
  • Edit relevant sections of HTML, 
  • Update affected examples (or make new)
  • Update affected figures (or make new) - raise as new issue to reassign this
  • Add to changelog in end of HTML
  • Check in (solve any merge first)
  • ( OWL file updated later after changes in HTML have been reviewed by group. )


Agreed upon points (assigned)

1. Replace Recipe with Plan [Stian] *Done*

2. wasControlledBy, wasEventuallyDerivedFrom, dependedOn,  hadParticipant  and wasScheduledAfter properties should be removed from  prov-o.  [james - will delete after renaming is handled; will wait for a ping]

3. Person, Organization and SoftwareAgent as (disjoint) subtypes of Agent. Add equivalence between prov:Person and foaf:Person [khalid] *Done*

4. wasAssociatedWith to be added [Tim] *Done(ish)*

5. alternateOf and specializationOf to be added instead of wasComplementOf [khalid] *Done*<

6. wasStartedBy and wasEndedBy relationships to be added. [Satya]

7. actedOnBehalfOf. [Daniel]

Timeline: Within Wednesday 20:00 Eastern time (Thursday 01:00 GMT) each of us email the changes - then we can report it during Thursday telcon.

Tim: Hoping for a review process, perhaps use the HTML only before changing the OWL


Outstanding issues: 1. OUTSTANDING: hasAnnotation can be used for recording meta-provenance?   2. OUTSTANDING issue: If we consider this (actedOnBehalfOf) as an n-ary property, it needs to  be discussed. If it can be a binary property, we can easily add it to  PROV-O Stian: This is probably true for every unaddressed point.  3. OUTSTANDING: Satya's concern about activity versus agent.  4. Attributes of TracedTo obtained using transitivity (Stephan raised issue 218) 5. wasStartedBy and wasEndedBy relationships to be added + qualification