PIL OWL Ontology Meeting 2012-04-23

From Provenance WG Wiki
Jump to: navigation, search

Meeting Information

prov-wg - Modeling Task Force - OWL group telecon


  • Tim
  • Khalid
  • Daniel
  • Jun
  • Stephan
  • Stian
  • Satya (may regret)


Around the room

Let's review what each of us did last week so that we are all up to speed. Please add points that I've forgotten.

  • Daniel:
    •   updated the expanded terms section.
  • Stian:
    • argued for hadQuoter
    • pulled out narrative code examples into eg-26
  • Satya:
  • Jun:
    • on travel, but still managed to
    • updated all the annotations to properties in the ontology. By the end of April 20, every term in the ontology was annotated using rdfs:comment property
    • write a SOTD section, which doesn't seem to have been integrated in the current draft yet
  • Khalid:ok
    • committed examples
  • Stephan:
    • Nothing yet
  • Tim:
    • Added prov-o rdfa to prov-o html
    • renamed Collection to Dictionary, added Invalidation




PROV-O HTML http://www.w3.org/2011/prov/track/products/10

[Are they all necessary for the upcoming public release? Are there anything that are absolutely important for the upcoming release?] @Jun

None are necessary for the upcoming release. This is looking beyond the public release. -Tim

What about the examples? To me, it would be nice to have them for the next release. - Jun

  • 349 from many, "cross section needs examples"
    • Do we work with one story for the examples or do people just do what they want? -Jun
    •   @Jun, I think that we should use the examples in section 3 as much as  possible, this way we minimize the number of examples that the reader  has to grasp
    • +1 Jun
    • 1/104 (0.96%) complete (now 10%)
    • announced at http://lists.w3.org/Archives/Public/public-prov-wg/2012Apr/0214.html
    • change files in http://dvcs.w3.org/hg/prov/file/tip/examples/eg-24-prov-o-html-examples/rdf/create/rdf
    • Jun: we should add as many examples as possible.
    • ... the rdfs:comments should.
    • Tim: change each file from "TODO" to "will be done by XXX (e.g. Khalid)
    • Daniel: we should review what is committed.
    • Stian: we want general guidance
    • Daniel: Expanded terms (ongoing). I would like someone to review the examples on the html doc. (will be reviewed by Jun tomorrow the 24th of April)
    • I'll do as much as I can today.(Dani) YES, qualified terms.
    • @Khalid, yes I wasn't planning on covering all :)
    • Stephan: willing to take an assignment
    • Tim: I'll make assignments and send them out after this meeting.
    • Tim TODO: review comments that Jun added.
  • 128 from Paul, "nongeographic locations"
    • Asked Paul what he needs to get it closed.
    • Daniel's example shows file path, Madrid, and a blog URL. (which should be a reasonable variety)
    • I never read them yet, so I can't comment - Jun
  • 334 from Jun, "qualification pattern not defined"
    • Suggestions from Satya?
    • Jun: give a definition. expect to see to express.
    • Stian: almost every PROV-O relation can also be expressed in a qualified pattern. Can add attriutues, time.
    • Jun: make it clear what the pattern is used form.
  • 338 from Tim, "how to explain prov:{activity,agent,dictionary,entity} vs. prov:had{Activity,Generation,Plan,Quoted,Quoter,Role,Usage} better"?
    • prov:involvee could be confused with prov:involved. Rename to prov:object ?
    • what is confusing?
    • Tim: reification's rdf:object
    • qualifiedGeneration [ atTime ] 
    • Stian: waht about prov:qualifiedGeneration [ a prov:Generation; prov:activity ]; .
    • Stian: waht about prov:qualifiedGeneration [ a prov:Generation; prov:qualifiedActivity ]; .
    • Khalid: but what does "qualified" mean?
    • Tim: 
    • Stian: My proposal used the same qualified* on both left hand side and right hand side - showing how they are related.. using only had* form on rhs might break this symmetry (slightly)
    • Daniel: but then, people will ask what is the difference between hadQualifiedActivity and hadActivity. No?
    • Daniel: Why don't we just say that the "had" is for optional things? (Like the hadActivity in derivation), while the "activity" means that the activity is not optional.
      • Stian: I think that would apply well to most terms we have already, +1

  • 344 from Jun, "definitions in cross reference section"
    • Jun added rdfs:comments, so mostly handled for now
    • prov:definitions are missing on properties, Tim needs to work cross-reference.py and annotations to avoid copy/paste syndrome.

ProvRDF http://www.w3.org/2011/prov/track/products/9

  • Satya to review and clean up (since most of them are his)

Ontology http://www.w3.org/2011/prov/track/products/3

  • 83 from many, "define inverses"
    • We agreed last week to maintain a separate file for inverses named.
    • Appendix listing inverses names will be added to next prov-o html draft.
    • Tim built automation around "prov:used prov:inverse "wasUsedBy"^^rdf:PlainLiteral"
    • REQUEST: someone to add these annotations to OWL file?
  • 89 from Luc, "entity attributes"
    • This is _very_ old (characterizing attributes in September), but Luc still thinks it needs to be addressed.
  • 267 from Tim, "justify all subproperty assertions with annotations"
  • 268 from Luc, "two layer ontology?"
  • 276 from Tim, "term coverage"
  • 281 from Tim, "http://www.w3.org/ns/prov# is still not dereferencable"
  • 336 from many, "incorporate review feedback"
  • 348 from many, "prov-o naming conventions"
    • This is the "do it" side of the "document it" aspect in 338
  • 363 from Tim, "prov:value (on KeyEntityPair) and prov:value (rdf:value)
    • naming clash. Rename prov:value to prov:keyValue?

seed issues


  • Raised by Jun: Can we talk about when or whether we will have snapshots for our ontology, like ProvenanceOntology-20120430.owl? Or achieve similar functionality via other mechanisms? I think it's important to have an ontology that is synchronized with each prov-o spec public release or even work draft.