PIL OWL Ontology Meeting 2011-08-02

From Provenance WG Wiki
Jump to: navigation, search

Meeting Information

prov-wg - Modeling Task Force - OWL group telecon

Attendees

  • Tim
  • Satya
  • Khalid
  • Daniel
  • James

Background

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

Agenda

Discussions

Group organization

Tim: Jim McCusker is joining the group, but not in today's telecon

Wiki page

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

Reuse - time and location

Daniel: looking at OWL doc and html. Propose reusing wgs and owl time for Time and Location. James: use directly or recommend?

Time ontology has interval and instant.

TODO: Tim to dig up his owl time diagram.

TODO: refering from OWL class to the Conceptual Model. (e.g., why is "Time" in this ontology?)

Time: process execution and BOB

  • Beginning of a process, end of a process
  • use and generation
  • derivation has temporal

Khalid: this class is abstract class and people instantiate or extend it with their vocab?

Daniel: list recommendations. We don't constrain them to use one.

Tim: keep it general, but elaborate with OWL time and WGS (which people can grab and use). (RESOLVED)

Satya: http://www.w3.org/TR/owl-time/#calclock

Daniel: we have PIL types, but will offer subclasses that people will use.

James: pil:Time but not saying anything about it, then saying "if you want, equiv it to dateTimeDescription".

James: two levels, 1) "there are times, there are locations." 2) concrete "wgs and owl time"

Satya: Reference and domain.

topic: elaborated and unelaborated ontologies vs. a single elaborated ontology.

  • theoretical vs. practical application uses

Daniel: group agreed that we are NOT defining time or location.

Missing relationship between generation/use/derivation and Time/Location

 Missing relationship between generation/use/derivation and Time/Location. (Daniel G) 

pe_1 generated b_1 . pe_1 occuring_at_location loc_1 .

Daniel: do not rely on rules.

Daniel: modeling n-ary as classes makes it difficult to use inference.

Help avoid using owl:Classes:

  • generated_at_time
  • generated_at_location

Should these be pushed back to the conceptual model?


Roles

Roles are not represented yet

TODO: Daniel (to do): Role examples from Journalism scenario


Arities (i.e., cardinalities)

how many bobs are derived from another bob. how many bobs does a process 

TODO: Khalid to try to find some cardinaltity constraints


HTML document

How to get the HTML file? From the w3c repository.

http://www.w3.org/2011/prov/wiki/Mercurial_repository http://www.w3.org/2011/prov/wiki/MercurialRepository

RESOLVED: The wiki is the "starting point" and will be transitioned to the HTML document. RESOLVED: Things that don't make it to the HTML will be archived in the wiki page.

process execution

"continuous" interval of time.

http://dvcs.w3.org/hg/prov/raw-file/default/model/ProvenanceModel.html#concept-ProcessExecution

Satya concerned about "it doesn't occupy the entire interval".

(did we address this?)

Generation

http://dvcs.w3.org/hg/prov/raw-file/default/model/ProvenanceModel.html#concept-Generation

TODO: "role" - role of the generated BOB?

TODO: Tim to describe roles of input and params.

Diagrams

TODO: Tim to revisit his graphml output

http://dvcs.w3.org/hg/prov/raw-file/default/ontology/ProvenanceFormalModel.html


TODO: we want diagrams in section 2 http://dvcs.w3.org/hg/prov/raw-file/default/ontology/ProvenanceFormalModel.html#owl2-encoding-of-the-provenance-model

Entailments

http://dvcs.w3.org/hg/prov/raw-file/default/ontology/ProvenanceFormalModel.html#entailments-supported-by-the-provenance-model

Satya: start by describing it in HTML, then move to OWL when we have clarity on how it should be modeled.

TODO: James add something to section 3 (entailments)

James: They might not belong in the ontology but in the semantics document. Luc mentioned we should know what we are expecting or not expecting with entailments.

Satya: RDFS/OWL entailments. Other stuff needs to be described in the entailments section (as rules?).

Q: why is "is" used in the relationships, and not "was"? -Tim


References

  • OWL-Time,
  • RDF
  • OWL
  • BFO, 
  • DOLCE

Next meeting

Monday. We'll figure out time over email. 9am California  12 noon  5pm UK