Notes from 2016-02-16 Web Conference

Attendees:

  • Bridget Almas
  • David Dubin
  • Ming Fang
  • Andrew Treloar
  • Ajinkya Prabhune
  • Nicholas Car

Action Items:

  • Ajinkya has requested feedback from any members who have experience with ProvONE and have tips or things to be aware of that they are willing to share.
  • Any members attending P7 are asked to keep the group informed of activities, discussions of interest related to provenance.
  • Members should begin thinking about meaningful activiites of the group for the months leading up to P8.

Notes: 

  • Dave is continuing work on documenting how to contribute to the group vocabulary . Goal is for this to be made available to the DFT group by the start of P7 next week. He will also ask the chairs of the Reproducibility group to test the contribution approach.
  • Of the members present on the web conference, only Andrew will be attending P7. He will try to keep an eye out for discussions and other activities that are of interest to the PROV IG while there.
  • We agree that as soon as P7 is over, we should begin in earnest planning activities with P8 in mind.
  • Ajinkya reported on a project he has been working on to build a provenance system for nanoscopy research. He will post additional details to the group wiki soon, but here are some of the highlights form my motes:
    • they evaluated OPM, Prov and ProvONE. They chose ProvONE for its ability to capture both the prospective and retrospective provenance.
    • they have mapped all of their workflows into ProvONE and integrated this data with the contextual metadata about the workflow outputs.
    • they used the representation of the provenance graphs to show the scientific workflows and execution plans
    • this enabled them to reduce the amount of duplication of workflows because it was easier for the researchers to see what had already been tried and the results of those workflows clearly
    • the provenance data is kept in ArangoDB because it accomodated storage of the native provenance graphs as well as the contextual metadata, which isn't represented as a graph.
      • (on the call we had some discussion about why the contextual metadata couldn't be modeled using an ontology - there were arguments both for and against that. they chose not to because the workflows are undergoing a great deal of change, and there was resistance to trying to model it before they are more stabilized. )
    • Ajinka is interested in receiving feedback from members of this group about whether there are any pitfalls they should be aware of in using ProvONE
  • Nicholas reports that he hopes to attend Provenance Week in Washington in June. If any other group members are going, let him know.
  • We discussed a bit about the item from last meeting, on how we could better surface the work members of the group have been doing to map their domain-specific workflows and ontologies onto standard vocabs like PROV and make that available as a resource to members of community.  Nicholas suggested that one useful way of thinking of this would be as a set of Design Patterns for use of PROV and other ontologies.  I'd like to explore at our next meeting if we could make some concrete steps towards such an activity.
  • The next call will be scheduled for shortly after P7 and will be at a time that is more amenable for our European colleagues.