Hi Vipul,
On 12/20/2014 06:35 PM, Vipul Kashyap wrote:
Thanks for the feedback and clarification – Looks like we will have to
work at two different layers:
1.The syntactic translation from FHIR XML/JSON to RDF/OWL –
2.Enrichment of the RDF/OWL representation via ontologies and OWL
axioms.. which is where I believe would provide the value due to
inferencing in a wide variety of applications –e.g., CDS, Clinical
Documentation, Quality Metrics, etc.
Am I correct in assuming that we are focusing on (1) in this group –
Shouldn’t we try to do (2) as well?
I think it depends on what you mean in #2. I think fundamentally two
things are needed:
A. To define mappings that translate between FHIR XML/JSON instance
data and FHIR RDF instance data; and
B. To define a FHIR ontology that can be used in conjunction with FHIR
RDF instance data, to help interpret that data's meaning. For example,
this would define the classes and properties in which FHIR RDF instance
data is expressed.
These two things work hand-in-hand of course. However, B does not
necessarily -- out of the box -- need to include much in the way of
upper ontology concepts. (That could be a design choice.) Instead,
upper ontology concepts could be linked in separately for specific uses.
David Booth
