April 8th, Templates Minutes

1. Roll Call
   1. Mark Shafarman (MS)
   2. Ken Kawamoto (KK)
   3. Russ Hamm (RH)
   4. Cecil Lynch (CL)
   5. Charlie McKay (CM)
2. Agenda Check
3. DAM REview from last week
   1. MS - Didn't see in model how the template nodes could be placed in a tree structure mapable into a HL7 version 3 model.
   2. MS - Attribute constrains do not clearly support HL7 datatypes. There may be other constraints.
   3. MS - Could not see how to constrain the node of one node based on the values on one or more nodes.
   4. MS - Couldn't see how to translate from this model into a LIM (version 3 serialized.)
   5. MS/CM - The nodes in the LIM can be named. Constraint Language needs to be able to name nodes.
   6. CM - Once you can name nodes you can reference the nodes from the instance of the templats. Use same node naming conventions. Node parts need to go down to the level of datatype components.
   7. MS - I think an algebraic or vector based language could be used (while you may not want to.)
   8. CL - A property of a class is an ObjectProperty in OWL. Need to get away from the idea of setting cardinality to 0 when it is unrealistic.
   9. CM - in openEHR the constraints are separate from the model structure, and path expressions are used to navigate to the nodes.
   10. MS - A semantically complete structure (linearized) with fully specified computable constraints assigned to it.
   11. MS - Prototype the identification metadata for a rim derived structure.
   12. CM - This looks more like a requirements model.
   13. MS - Going backwards from an existing structure into this DAM model. Key: How do we define the path along the nodes. A notation for the nodes is required, and hoe to express associations between the nodes. The notation is not important, but must be easy for people to understand.
   14. CL - if the intent of CL is to provide validation at the instance level, it works. the assumption is there are real events at the instance level. At the model level, validation is difficult where validation on cardinality of 0 is not realistic.
   15. RH - Does the fact that openEHR maintain the RM and the Constraints seprately address some of this?
   16. MS - The constraints are applied separately to the derived model.

1. New Call Time
   1. Next call will be one hour later.
2. San Antonio Agenda [1]
3. Next Agenda Topics
4. Adjourn

http://informatics.mayo.edu/wiki/index.php/April_8th%2C_Templates_Agenda