OBOE Ontology Extensions Meeting Agenda
=======================================
MINUTES
-------
+-------------------+----------------------------------------------------------+
|DATE               | June 10, 2011  1:00 pm PDT                               |
+-------------------+----------------------------------------------------------+
|TYPE OF MEETING    | Ontology Development                                     |
+-------------------+----------------------------------------------------------+
|CALL IN            | EVO_ (Password: oboe)                                    |
+-------------------+----------------------------------------------------------+
|NOTES              | Chris Jones                                              |
+-------------------+----------------------------------------------------------+
|ATTENDEES          | Shawn Bowers, Ben Leinfelder, Chris Jones, Matt Jones,   |
|                   | Margaret O'Brien, Mark Schildhauer                       |
+-------------------+----------------------------------------------------------+

**Initial Agenda**

1. **Discussion of Spatial and Temporal class choices**

- See: Spatial-Temporal-Components.pdf_
- Ontology extensions need consistent classes for spatial and temporal concepts
- International standards provide low-level building blocks of these concepts
  ISO, OGC, FGDC, etc.
- Domains are expected to create application schemas that incorporate these   
  concepts into higher level concepts
- None provide OWL-based implementations NASA SWEET provides both, although less 
  comprehensively

2. **Which is more appropriate for integration into OBOE extensions?**

- SWEET
    - More domain relevant 
    - More complex relationships
    - Less consistent modeling
    - OWL syntax

- GML
    - Less domain relevant 
    - Less complex relationships
    - More consistent modeling
    - XML Schema syntax

Adopting concepts from GML or SWEET namespaces may be fuzzy (copy to new ns, use separately), or defined as equivalent to the concept.

Versioning of OBOE may be problematic if SWEET is adopted because of their versioning policy

Annotations are defined in RDF Schema: 'isDefinedBy': used to provide a definition of a term (a special form of "seeAlso"). Also 'similarTo' may be useful.

Certain areas of SWEET look to be more well defined, and others not.  It would be good to work with the ESIP community to help SWEET to evolve.

OBOE classes could include annotations that refer to SWEET classes that 'didn't quite fit' and why.  This may be a reasonable paper topic.

The problem of ontology alignment is not new, and we should try to address it within our domains of interest.

What about partial alignment? It's not easy in the OWL framework.  The problem is where you have agreement among two communities of the concept of a 'watershed', but we don't have agreement on how that concept fits in to the larger concept map.  Part of the issue is the pragmattic expression of the differences in OWL.

Less formal version of "equivalentClass" could be "inPrincipalEquivalentTo"

relatedTo
similarTo
sameConceptAs (different than sameClassAs)  ... maybe need a better name for this?
  -- e.g., equivalentInPrinciple
  -- or sameScopeAs

Annotations will go into oboe.owl
Space and Time classes will go into oboe.owl

We could potentially subclass SWEET classes, but we will still inherit superclass concepts.
  -- e.g., if we define B, which is "isNotQuiteEquivalent" to C
  -- One option: have B subclass C, but would make B inherit superclasses of C
  -- Could also define a more general class A and have B and C subclass A 
  -- Could go further, and assert overlap 
    -- Could share an instance
    -- Could define an "overlap" class D (subclass of B intersect C)

.. _Spatial-Temporal-Components.pdf: https://code.ecoinformatics.org/code/jmx/private/meetings/20110610-Conference-Call/Spatial-Temporal-Components.pdf

.. _EVO: http://evo.caltech.edu/evoNext/koala.jnlp?meeting=MIM9Ms2D2iDIDl999aDa9I