Draft FHIR RDF specification

http://w3c.github.io/hcls-fhir-rdf/spec/

<ericP> http://hl7-fhir.github.io/structuredefinition.html#resource

eric: profiles are structure definitions and some other stuff.
... Structure definitions are a name and snapshot + differential.
... THen click on ELementDefinition http://hl7-fhir.github.io/elementdefinition.html#ElementDefinition
... We work with the paths, because those are fully qualified, to give us the RDF property names.
... FHIR only has one level of disjunction, and it is in 'type', because there can be any number of them.
... Things like minValueDate are represented using the same names (like minValueDate) in RDF.
... This is a schema language represented in its own schema language.
... The other thing about 'type' is that it can be part of a code that is a controlled vocab.
... For closed valuesets, we enumerate them in shex.

<ericP> http://hl7-fhir.github.io/observation.shex.html

eric: This shows Observation.status and the corresponding observation-status enumeration: http://hl7-fhir.github.io/observation.shex.html
... I was thinking we should publish ASAP to help the ShEx work get published, but it looks like the ShEx work will get published anyway, so it would be better to wait on publishing this until ShEx is published, so that ShEx can be referenced.
... i want to implement this in scala, to help catch mistakes.
... The FHIR pub process produces the shex string, but not an abstract syntax.
... An implementation in scala would make it obvious how the mapping goes.
... If someone else wants to do that, I'd be happy to work with them.
... The scala would take structure definitions and output shex
... The current one is all string based, and its an integral part of the FHIR build process, in Java.

dbooth: is there more review needed of the draft doc?

eric: Not until we have the scala coded.

Next Meeting

dbooth: I will cancel the 5PM call -- too many key people out.

ADJOURNED