CS API based on collapsed models
Jump to navigation Jump to search
- If one were to create a programming API for RIM based models, one could base it on collapsed/flattened models instead of on a CIM. Code generation based on collapsed/flattened models results in smaller (more manageable) code.
- Open question: could one use collapsed/flattened models in the CO and CP cells?
Lyssa K. Neel (CIHI), on the RFP for a Canadian tool based on this mechanism:
- The "collapsed" MIF is described here Message_reshaping_rules
- Grahame: More specifically, as implemented in the pan-Canadian schemas / Instance editor.
- It does the following:
- Hide all fixed values from the application
- Where possible, expose data elements using business names, not HL7 names
- Simplify the object model by eliminating unnecessary nesting, 1..1 relationships and similar constructs
- This had been balloted as a "new ITS" but was rejected.
- Mohawk College has a working version of such a system (called the "Everest Framework") available on their website: http://everest.marc-hi.ca/
- During the Phoenix WGM (January 2010) Mohawk reported that initially their API was based on a mixture of collapsed-models and full v3 models. This only lead to confusion on the part of the implementers. Their conclusion: either hide all complexity (and use a simple DSL), or use the fullblown v3 models. Don't mix the two.