Skip Navigation
NIH | National Cancer Institute | NCI Wiki   New Account Help Tips
Page tree
Skip to end of metadata
Go to start of metadata

Date/Time: Thursday, November 17, 2011; 10-11 AM ET

Attendees

Brian Davis, Sherri De Coronado, Stuart Turner, Rick Kiefer, Harold Solbrig, Peter Yim, Grace Stafford, Larry Wright, Stuart Turner, Hua Min, Gilberto Fragoso, Michael Gruninger, Riki Ohira

Agenda

  1. Introduction of Peter Yim and the Open Ontology Repository group
  2. Overivew of OOR -Peter Yim
  3. Questions and Answers
  4. Review of Previous Meeting Action Items (time permitting)

Assigned To

Description

Due Date

Stuart Turner

Send teleconference information for ORWG weekly meeting to Peter Yim and colleagues

11/10/2011

The Group

Add questions for Peter Yim regarding OOR to the wiki site: https://wiki.nci.nih.gov/x/SLJiAw

11/15/2011

Riki Ohira

Send meeting invite for next week's ORWG to Gilberto Fragoso since he is the POC for Ontolog group

11/11/2011

Stuart Turner

Continue to research possible issues with bringin OMV to standardization with ISO

11/16/2011

Trish Whetzel

Speak with Ray (BioPortal) development team about adding this functionality to the features list and determine whether this is an easy add or would need to be included in later releases

11/16/2011

Brian Davis

Put together project plan for Roadshow and outline of content for presentation

11/16/2011

The Group

Provide input into project plan

11/15/2011

The Group

Provide comments on the Citation Table for OMV references

11/23/2011

Brian Davis

From previous meeting: Incorporate timeline for OMV standardization with goal of June 31st meeting

11/15/2011

Notes

  • Peter Yim gave an overview of the Open Ontology Repository group history and activities
  • History of the group
    • lineage: DARPA DAML Library (MikeDean/BBN), MITRE study on Ontology repository and registry (LeoObrst/MITRE), and the Collaborative Ontology Development Service and infrastrcucture (CODS) effort (PeterYim/CIM3 with support from Mark Musen/Stanford created using client-server Protege running on a tier-1 open infrastructure with the hope to provide an open ontology editing and repository platform.)
    • the OOR discussion started on the Ontolog forum, and was then proposed as the theme for the 2008 Ontology Summit.
    • Ontology Summit is an annual series of activities spanning over a period of 3-months, co-organized by Ontolog, NIST, NCOR, NCBO, IAOA and most recently National Coordination Office of NITRD (the OSTP body that coordinates IT research and development within US government agencies.)
    • 2008 Communique represented culmination of that Ontology Summit's work (when the group was really moved forward)
  • Goal is to promote global collaboration
  • The OOR initiative can be looked at as:
    • an open source software development project – the initiative and model themselves after Apache,
    • an infrastructure project – with a commitment to provide (at least one instance) of an ontology repository and registry for public use,
    • a value-added service to the stewards as well as the users of ontologies hosted on OOR
    • a social project – fostering collaboration and interoperability (through global federation) of ontology repository operators
  • They have discussed the impact of IP restrictions and licensings
    • At the core, they want whatever is developed to be open and unencumbered. chose, as their default, the simplified (two clause) BSD license for codes, and the CC-By-3.0 license for content.
  • The current OOR Sandbox instance is running pure BioPortal except for very minor customization (like their logo)
  • There are individuals/institutions contributing code and others contributing content
  • Q: Do you have implementation that has passed sandbox or using extensions to the BioPortal software?
  • A: We are just going through the effort to put up a sandbox instance, and a devbox instance. The latest version of BioPortal just released (as ncbo-appliance-v0.4) will be run in the OOR sandbox. OOR is expected to be supporting non-biomedical content. The development box (devbox) will be standing up extensions as they get developed. Past the devbox, they plan to have another box which they are calling the production OOR. Difference between sandbox and production is a gate keeper, which is software that would enforce policies, e.g. restrict the uploading of ontologies (there needs to have standard names, metadata, etc.). It would ensure proper access to the ontologies, metadata, etc.
  • Q: Do you have a sense of the direction of BioPortal development and OOR. Would the forked OOR contribute back to BioPortal; do you see BioPortal and OOR becoming the same and merging into a single resource?
  • A: Development work in BioPortal and OOR would be separate, and the contribution would be bi-directional, but not necessarily symmetric. OOR will, most likely, want to incorporate enhancements in BioPortal, but BioPortal may not always need OOR enhancement. To support Spatial Ontologies, or to support Common Logic, for example, may involve extensions that may or may not be relevant for Biomedical ontologies. Anyhow, any enhancement would become available to those running BioPortal instances on their own if they find it useful. Then again, there are OOR  developments, like the federation, that BioPortal would want to leverage. Another joint NCBO/OOR session is coming up (Nov-29) to review architecture with an aim towards better synergies in development efforts from both teams.
  • This group has been focused on representations of ontologies, we have been looking at OMV, ISO 19763, and CTS2
  • Harold Solbrig and colleagues at Mayo have done a cross map of different models using RxNorm as pilot
  • CTS2 created a wrapper and plan to have native CTS2 view into BioPortal content
  • Comment: The OOR team has been  discussing extending the OMV to better cover relationships among and within ontologies. This is not quite mapping (as mapping usually looks at the term-level), but this is an effort to look at relationships of entire ontologies. Are two ontologies the same, overlapping, extensions of each other, etc?
  • The OMV mapping, there were just a few that didn't map exactly, but most did. They still need to put tags
  • Peter and Michael answered the questions the group came up with earlier:
    • A1: There is an informal commitment that they would adopt OMV as core, but would need to add relationships (not extensions in terms of properties)
    • Extensions of OMV, trying to get a better idea of what the extensions are and how they are being used "in the wild"
    • OOR sees themselves as users of metadata and don't really want to be the drivers of determining metadata, for the most part properties, OMV is supplying that role
    • OMV Analysis: Identified properties that might be better to reuse like Dublin Core and seeing what properties should be harmonized. And discussed process for modifications, updates to OMV with OMV.
    • Q: Have you been adding new classes, properties, or extending existing values of properties? We looked at each class, property to make sure they made sense to us as a group and try to improve definition or naming, etc. Looking at gaps like licensing and versioning. We've also agreed where there are mappings with other standards and we will provide mapping where it is known.
    • Dublin Core rendering is one thing we would want to include for CTS2. We want to make sure "has creator" would include Dublin Core creator element as well
    • Q: How are ontologies and artifacts different from software and artifacts? It helps to focus what is special about ontologies.
    • Versioning, OOR is still struggling and it may be worth having a larger group sessions/discussion on that. Harold is interested in participating with that as is NCI.
    • Repository versus Registry: Differences, OOR would be both repository and registry. As for ontology and ontology instance data we are not sure yet (still discussing).
    • As for mirrors, in terms of meeting infrastructure robustness requirements it will definitely be implemented. As for functional needs, the guess is that it will be the preference of the implementer, and be left as being configurable by the operator of an OOR instance.
    • Q: There has been discussion with OMV to move it forward for standardization (looking at ISO, extension of ISO 19763), or may look at other ISO 17347 as a more relevant option. We want to see OMV more formalized since so many people use it. Any thoughts about this?
    • A: OOR just hosted a session on this. Michael and Peter are (as is Harold Solbrig) involved with the ISO NP 17347 project (on US and Canadian expert group for this ISO WG). OOR would want to talk about how to annotate ontology for a particular vocabulary, but not getting into the content. The OOR team hasn't engaged with Raul Parma so far, but have been contacting Peter Haase and, more recently, Mari Carmen Suárez-Figueroa (who used to be on Raul's team).
  • Q: Is Raul Parma still involved in OMV?
  • A: They have by default moved on because the NeOn project has ended, but they are still devoted to it and they are still working on it, but there is no governance body and specific funding.
  • IAOA has set up a standardization coordination committee that is (not endorsing standards), but will foster interactions between the ontology and standards communities, and to facilitate the application of ontologies and ontological analysis to existing and emerging standards. Co-chair is Michael Gruninger and John Bateman (Bremen University). There is always hope that the right people need to be brought together for collaboration, individuals are willing to "meet" and collaborate to move these types of things forward.
  • Peter encouraged the group to join the mailing list and attend meetings; sometimes they have special events or key open meetings to discuss activities
  • Join the [oor-forum] mailing list: oor-forum-join@ontolog.cim3.net
  • See: http://OpenOntologyRepository.org
  • Come to our OOR-team weekly conference calls - Tuesdays 16:30 UTC (8:30am PST)

Action Items

Assigned To

Description

Due Date

Stuart Turner

Send link to Harold and Sherri's ;mapping to Peter Yim and colleagues

11/22/2011

  • No labels