See also: IRC log
<trackbot> Date: 06 May 2015
<Lloyd> Call running long. Should be on in ~2min
ok
My followup with Bijan Parsia: https://lists.w3.org/Archives/Public/public-owl-wg/2015May/0001.html
dbooth: Could also leave the
blank nodes in, and blindly skolemize *all* blank nodes.
... Grahame is concerned about the burden that turning these
into URIs would place on RDF generation.
<rhausam> can you state again what Bijan's proposal was? where the functional property was to be used?
Rob, Bijan's reply: https://lists.w3.org/Archives/Public/public-owl-wg/2015May/0001.html
http://hl7-fhir.github.io/rdf.html
http://hl7-fhir.github.io/rdf.html#1.17.3.3.5
dbooth: We want to be able to
infer that the blank node instance of fhir:Observation.code is
a loinc:54411-4;
... based on fhir:Coding.system [ fhir:uri.value <http://loinc.org>] ; and
fhir:Coding.code [ fhir:code.value "54411-4" ];
lloyd: Linkage to a concept is a
combination of fhir:Coding.system and fhir:Coding.code and
sometimes fhir:Coding.version
... The version may or may not be indicated and may or may not
be relevant. Orthogonal issues.
... Some say what the version is, some don't.
... Plus version is a string, and there is not consistency in
the string representation.
... In some code systems the meaning can change between
versions, and it is subjective whether you care about the
meaning change between versions.
... Ideally you would have the concept inferred from system,
version and code.
... And you'd have relationships between the concepts, and the
version would have a fixed set of values.
... If you have code x and version y, the version must be one
of ten possibilities, so that if the concept is equivalent in
all 10 versions, then you know what it is.
... Some will assert LOINC changes meaning between versions;
others say no. Depends on how much you care about the
details.
Rob: SNOMED is pretty well behaved. ICD is not.
dbooth: Postpone deciding how to handle these blank nodes until eric has tried function and inverse function a bit more.
<scribe> ACTION: Eric to try functional and inverse functional properties to see if they fix the blank node problem. [recorded in http://www.w3.org/2015/05/06-hcls-minutes.html#action01]
<trackbot> Created ACTION-25 - Try functional and inverse functional properties to see if they fix the blank node problem. [on Eric Prud'hommeaux - due 2015-05-13].
lloyd: Define a common concept for the three, then define three constraints.
dbooth: Need something in the instance for round tripping.
lloyd: Could be indicated in the type.
dbooth: We need to see three FHIR XML examples, with their corresponding RDF.
http://hl7-fhir.github.io/datatypes-examples.html#1.18.1.6
<diagnosis>
<coding>
<system value="http://snomed.info/sct" />
<code value="128045006:{363698007=56459004}" />
</coding>
<text value="Cellulitis of the foot" />
</diagnosis>
<Lloyd> http://hl7.org/fhir/2015May/auditevent.html
lloyd: click UML and look at
Event. It has type CodeableConcept, subtype CodeableConcept,
outcome code, and purposeOfEvent Coding.
... When the type is only code, the system is already
fixed.
http://hl7.org/fhir/2015May/auditevent-example.xml.html
that example always seeems to use system and code.
<ericP> AuditEvent/event/type is a codable coding
lloyd: starting with AuditEvent go to AuditEvent / event / type (CodeableConcept)
<ericP> AuditEvent/event/outcome is a code
<ericP> AuditEvent/source/type is a coding
<ericP> fhir-rdf pipeline
<scribe> ACTION: Lloyd to propose RDF versions of the above 3 FHIR XML examples [recorded in http://www.w3.org/2015/05/06-hcls-minutes.html#action02]
<trackbot> Error finding 'Lloyd'. You can review and register nicknames at <http://www.w3.org/2014/HCLS/track/users>.
<ericP> example of ShEx extension
eric: This page is about tooling
that reads FHIR spec and writes XSLT, which translates FHIR XML
to FHIR RDF.
... And it generates ShEx also, for going back to FHIR
XML.
... The teal colored material in the last link is the ShEx
extension
<ericP> -> https://github.com/w3c/hcls-fhir-rdf/blob/master/generic/fhir-shapes.shex#L2274 GenX for FHIR/RDF -> FHIR/XML
eric: that converts ORIM RDF to
FHIR RDF.
... Here's what it looks like in ShEx for FHIR RDF.
... GenX instructions generate FHIR XML.
<ericP> demo of GenX
eric: I've tried it on small pieces and it works.
eric: in this example you have
the schema with GenX instructions for creating DOM tree.
... Then RDF in the middle. Then validation info on the
right.
... Click popup there, under Validation messages.
... ShEx approach is concise, though we might use the reference
library instead of ShEx.
lloyd: Reference libraries will
end up being used by most supporters. If we include anything
else other than the ref libraries then implementers won't touch
it.
... We also want a single authoritative implementation that
we've tested the hell out of.
... But we could also have this out there.
... Fine to let many flowers bloom, though we may not maintain
them all, depending on bandwidth.
dbooth: May be value in using
ShEx initially to enable quicker trials and iteration.
... The quickest way to try it the better.
lloyd: I suspect that grahame has
done some work on the ref impl, but not sure what.
... Once we're comfortable that the conversions work and we
like them, we can push it into the ref impls.
dbooth: should we start looking at a test suite?
eric: I'd like others to help on ShEx also, because it would give me more feedback on how useful it is.
dbooth: I'd like to help on ShEx approach.
my github: dbooth-boston
eric: Send me your github ID if you want to hack on the ShEx approach.
This is scribe.perl Revision: 1.140 of Date: 2014-11-06 18:16:30 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/type is a concept/type is a coding/ Succeeded: s/concept/coding/ No ScribeNick specified. Guessing ScribeNick: dbooth Inferring Scribes: dbooth Default Present: DBooth, +1.801.810.aaaa, rhausam, ericP, lloyd Present: dbooth ericP lloyd rhausam Found Date: 06 May 2015 Guessing minutes URL: http://www.w3.org/2015/05/06-hcls-minutes.html People with action items: eric lloyd[End of scribe.perl diagnostic output]