Thursday, February 6, 2020, 10 AM US Eastern Time

Connection Info

To join the online meeting:


  • Mike Conlon
  • Brian Lowe
  • Anna Kasprzik
  • Ralph O’Flinn


Ontology Interest Group Google Folder

Google Doc for meeting notes:


  1. Updates
    1. CNI abstract submitted.  Will notify when I hear back
  2. February 20th Virtual workshop
    1. One track on events?
    2. Facilitator?
    3. Ready for announcement?
  3. Possible other ontology events
    1. US2TS, March 9-11, North Carolina State University, Raleigh.  Chris Mungall, Juan Sequada, Jim Hendler
    2. Graph conference, May 5-7, Columbia University
    3. CRIS2020, June 17-20, Limassol, Cyprus
    4. US VIVO event, October? no update yet
    5. Force11 October 19-21, San Sebastien, Spain. Simon will be there.
    6. SWIB, November 25-27, Bonn
    7. Fall CNI, December 14-15, Washington
    8. Virtual meeting – an all-day meeting.  
  4. Language Ontology draft. This draft was created in keeping with Early Thoughts on Representing Languages and Language Capabilities
    1. Overview
    2. Next steps?


  1. Event on Feb 20.  Possibly about events.  We might ask attendees about future topics such as Arts and Humanities, the W3C Time Ontology.
  2. Typical pain points for VIVO are representation of publications and the associations of publications with people and organizations and 2) grants and projects and their various associations.
    1. How can tooling help?  We might be able to separate the concerns of the data managers from the concerns of the ontologists.  RMLMapper can encapsulate the ontology and transform JSON, RDB, CSV and other formats to assertions.  This can also be done by Karma.  It behooves the ontology group to explore these tools and demonstrate how the ontology can be backgrounded.  Ontologists need to understand the ontology.
    2. Can there be shortcuts?  That is, one might say "a wrote e" Rather than "a has role b; b realized in c; c part d; d has output e".  If we have tooling, we should be able to produce shortcuts as well as the ontologically correct set of assertions.  Tooling for short cuts could be Dead Simple Ontology Patterns.  Used in the OBO to add assertions.  Supported in robot.  Needs more exploration.
    3. How is the ontological set of assertions useful?  Two ways: 1) It provides the appropriate places to have data properties.  Just as in any other modeling, relations have properties.  A position has a time interval of employment.  We can not simply says "z works_at x".  The latter can be inferred from a set of assertions which involve a relation, the position.  2) The ontological assertions allow for data sharing across domains.  When VIVO uses date times, they should be the same dat times used by others.  When VIVO asserts a publication is an information content entity, this has meaning to people who work with other information content entities.  We should be creating linked open data, useful by others, not just within the domain in which it was created.
  3. Tooling might have two steps:
    1. Represent as part of the ingest process.  Looks familiar. Tooling creates robust ontology structure. Then you don’t need shortcut properties.  The short cut is the JSON structure. Perhaps a two step process -- RMLMapper to make assertions, robot to reason and reduce.  This should work fine for small incremental loads. Isolate the practitioners from the creation of the assertions.
    2. The full load use case?  Could be a problem -- the reasoning gets big.  Need to experiment more with tooling.
  4. Why all the talk of tooling?  Ontologists in other domains are able to focus on ontology.  VIVO ontologists are always in tooling conversations.
    1. Other domains use ontological URI as annotations in existing data.
    2. Other domains use ontology for inference, not data
    3. Only now are large knowledge graphs of actual data being created by companies and others.  The thinking at Cornell was more than a decade advanced
  5. Speaking of tooling, can we get linked open data when we need it?
    1. Micro servers to provide linked open data -- dates, pubs, grants.  Could imagine setting up shim servers that translate data from open APIs to linked open data using VIVO Ontology.  PubMed could be treated this way.  VIVO stores the URI of the pub.  The shim provides data from pubmed about the publication on demand.
    2. DBPedia to provide data on countries, other domains?

To Do

  • Mike will ask about the fit between W3C Time Ontology and BFO.
  • Gain familiarity with RMLMapper
  • Gain familiarity with robot
  • Gain familiarity with opportunities for shim servers
  • No labels