See also: IRC log
<hsolbrig> https://chat.fhir.org/#narrow/stream/terminology/topic/Required.20Coding.20value.20sets
harold: We've started converting
required valuesets into shex. Some of them are in the form of
Coding or CodeableElement
... Looking at RDF rules needed for CodeableElement. Discoverd
that the Coding datatype omit the code system.
... The only thing in the example is the code itself -- no
coding system.
... Right now the shex validation rejects it, becuase the
structure def says that it should be a particular coding
system, but the instance data does not give the system.
https://chat.fhir.org/#narrow/stream/terminology/topic/Required.20Coding.20value.20sets
eric: I propose that shex validation should reject those cases.
dbooth: I agree.
eric: How did this come up?
harold: The valueset def gives a
set of code/system pairs. When it is required then the shex
needs to check it.
... CodeableElement says at least one of the entries has to be
a CodeSystem/Code combo, but you are allowed to have any
additional ones that you don't care about.
... This is a second case where we're confronted with partially
computable information.
... Do we have to characterize this problem so that people
understand that if validation fails it may not mean that they
are strictly in violation of FHIR.
... The coding rule could be expressed in a profile, i.e., that
system is required in this profile. That could also be done for
links, so that they are checkable (to ensure that they point to
the correct kind of entity).
... If we do this in profiles, how do we do the profile without
having to regenerate the shex? I have some ideas.
eric: This might harden the view of what inheritance means.
harold: Could also do runtime
validation with a set of modules that can be applied for a
given validation feature.
... I suggest we have one level of promiscuous validation that
reflects FHIR's requirements, and then handle additional rules
via profiles.
dbooth: I agree.
harold: ShEx consolidates several formalisms: slicing, valuesets, structure defs, profiles.
dbooth: proposed agreement: The baseline shex validation will reflect the (promiscuous) FHIR requirements. Additional validation rules can be triggered by profiles.
AGREED: The baseline shex validation will reflect the (promiscuous) FHIR requirements. Additional validation rules can be triggered by profiles.
<ericP> +1
eric: The linked FHIR profile could be equally useful to both RDF and JSON/XML worlds
<hsolbrig> http://hl7.org/fhir/2016Sep/datatypes.html#coding
<hsolbrig> http://hl7.org/fhir/2016Sep/observation.shex.html
harold: If you go to Turtle,
you'll notice that the Coding shape is different than the
datatype Coding.
... Neither fhir:concept nor fhir:index are listed in the
Turtle Template.
<scribe> ACTION: Harold to work with Grahame about fhir:index and fhir:concept missing from the datatype def [recorded in http://www.w3.org/2016/11/15-hcls-minutes.html#action01]
<trackbot> Created ACTION-72 - Work with grahame about fhir:index and fhir:concept missing from the datatype def [on Harold Solbrig - due 2016-11-22].
ADJOURNED
trackbot, start meeting
<trackbot> Meeting: Semantic Web Health Care and Life Sciences Interest Group Teleconference
<trackbot> Date: 15 November 2016
This is scribe.perl Revision: 1.148 of Date: 2016/10/11 12:55:14 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/FHIR/ShEx/ Succeeded: s/FHIR/ShEx/ Succeeded: s/reflects ShEx/reflects FHIR/ No ScribeNick specified. Guessing ScribeNick: dbooth Inferring Scribes: dbooth Present: David_Booth EricP Harold_Solbrig Found Date: 15 Nov 2016 Guessing minutes URL: http://www.w3.org/2016/11/15-hcls-minutes.html People with action items: harold[End of scribe.perl diagnostic output]