PIL OWL Ontology Meeting 2012-03-26

From Provenance WG Wiki
Revision as of 17:21, 26 March 2012 by Tlebo (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Meeting Information

prov-wg - Modeling Task Force - OWL group telecon

Attendees

  • Tim
  • Khalid
  • Paul
  • Paolo
  • Stephan
  • Satya
  • Mike
  • Jun
  • Stian
  • Daniel

Summary

  • We reviewed the assignments for sections in the PROV-O HTML, whose draft is due Friday. As sections get added this week, we'll announce within group for some internal review.
  • We discussed the ProvRDF coverage automation, how the ASN signatures are extracted from DM, and how they should match. Any desire to have a different signature should be an ISSUE against DM.
  • We discussed collections in the telecon, and Stephen proposed some changes that Paolo will be considering:
    • simple insertions are "bulk" insertions of one element (and thus simple should fade away)
    • collections' members should be assertable directly (without showing derivation to empty sets)
    • how to reconcile prov collections with RDF collections (which may be deprecated)?

Minutes

    Timeline

schedule - https://docs.google.com/a/revelytix.com/spreadsheet/ccc?key=0An15kLxkaMA3dFVCWm9aREZFemNOYjlGQjdPRkdFZXc#gid=0

  • 3/30 (4/2) - draft WD2 due (both ontology and html)
  • 4/9 - feedback from WG due
  • 4/16 - final WD2 due for vote (both ontology and html)
  • 4/19 - WG votes
  • draft should be ready for publishing, so we can devote the final two weeks (3/30-4/15) to closing new issues that get raised and taking care of administrative issues

    HTML

Section assignments:

  • Section 1: Introduction
    • Mike
  • Section 2: Glance
    • Automatic
  • Section 3: The PROV-O Ontology Description
    • Jun can update the diagram and examples and overview text
    • TODO: Jun: did your changes make it in?
    • the text is fine, (except for "Terms for expressing Qualified relationship" should be "Terms for expressing qualified relationship") but the diagrams are not shown in the html yet. 
  • Section 3.1 Prov-O starting points:
    • Khalid (Description of the diagram + example)
      • will include Jun's and Khalid's updates - look for it tomorrow.
  • Section 3.2 Prov-O expanded:
    •  ???? (depends on 3.1? -- wait until 3.1 is done?) I agree
    • Daniel, Satya? TODO (keep it simple. give an introduction)
  • Section 3.3 Qualifying Relationships in Prov-O:
    • Tim (delayed to Tue afternoon)
    • A lot have been outlined in the structure-v3.docx, including a good selection of examples and two diagrams
  • Section 3.4 Collections: (lowest priority section)
    • Stian (should be able to start on the 27th)
    • done: Tim remind Stian.
  • Section 4: Cross reference for PROV-O classes and properties
    • Automatic

reviewers (less of a concern for within team - could be released to wg):

  • khalid - thursday
  • paul
  • Jun - Friday morning (UK time)

Authoring methodology options:

ProvRDF

  • provRDF-coverage - need to have all terms from DM represented in OWL and vice versa
    • http://aquarius.tw.rpi.edu/prov-wg/provrdf-owl-coverage
    • Stian - any comments?
    • Stian to update "had" properties
    • Tim to update ProvRDF page to use correct syntax for auto-generating coverage page
    • Satya: I am not clear about including "id a prov:Agent" on the RHS for Person, Organization, and SoftwareAgent - it is not present in DM. If we include inferred triples, do we consider prov:Entity also for these three cases?
      • mention not only the direct, but also the inferred on the RHS. Why not include prov:Entity?
    • done: Tim to give Paolo command to give list of ASN expressions. ()
curl -H "Content-Type: text/turtle" -d "<http://dvcs.w3.org/hg/prov/raw-file/default/model/releases/WD-prov-dm-20120309/prov-dm.html> a <http://xmlns.com/foaf/0.1/Document> ." http://sparql.tw.rpi.edu/services/datafaqs/util/wikitable-fol

ontology

  • IN PROV-N
1) derivedByBulkInsertionFrom(c1, c, {("k1", v1), ("k2", v2)})  
2) derivedByInsertionFrom(c2, c1, "k3", v3) 
3) derivedByBulkInsertionFrom(c2, c1, {("k3", v3)})

I [stephan] think 2) == 3)

paolo saw the bulk as a convenience for the individuals.

when we see a prov-n of single insertion, we can consider it as a bulk of one.


    • containment

http://lists.w3.org/Archives/Public/public-prov-wg/2012Mar/0459.html

in PROV-N

containedBulk(c3, {("k4", v4), ("k5", v5)})  

Why not this? (NOT PROV-N)

collection(c3, {("k4", v4), ("k5", v5)})
 
 :c3 a prov:Collection ;
        prov:containment [ a prov:Containment ;  # This is a prov:Involvement ?
                                           prov:collection :c3 ;  # we omit this, subject is implied?
                                           prov:keyValuePair [ a prov:KeyValuePair ;
                                                                                   prov:key "k3"^^xsd:string ;
                                                                                   prov:value :v3 ] ] ;
        prov:containment [ a prov:Containment ;
                                           prov:collection :c4 ;
                                           prov:keyValuePair [ a prov:KeyValuePair ;
                                                                                   prov:key "k4"^^xsd:string ;
                                                                                   prov:value :v4 ] ] .

paolo: this is set membership, not spatial containment. (spatial containment is already defined elsewhere (e.g. geonames))