Semantic Web Health Care and Life Sciences Interest Group Teleconference

22 Sep 2015

See also: IRC log


Tony_Mallia, David_Booth, Bill_Kleinebecker, Darrell_Woelk, Harold_Solbrig, Lloyd_McKenzie, Marc_T, Paul_Knapp, Rob_Hausam
David Booth and Paul Knapp


<trackbot> Date: 22 September 2015




tony: two construction options

lloyd: don't frame it as two approaches. there are 6-7 different construction options depending on how the valueset is defined.

tony: yes, one approach multiple options.

lloyd: essentially only one option: this is the specific set of codingbase individuals that are permitted. Then there are many mechanisms for determining those individuals: enumeration, hierarchy, expression, etc.
... But it always resolves to an enumeration of CodingBase individuals.

tony: right

lloyd: but you're implying that there's a decision by using the word "option"
... It isn't an option for the person generating the RDF.
... all VSs are named classes representing restrictions of CodingBase individuals, and a number of different ways of expressing those restrictions.

example is from: http://hl7-fhir.github.io/valueset-allergy-intolerance-status.html

tony: this example is a structured hierarchy
... direct RDF is not useful. direct RDF equivalent means: instance becomes instance, type becomes type, etc.

lloyd: we MUST treat ValueSets the same as everything else.
... Need to be able to round trip to XML and JSON
... Other requirements: want to be able to take conformance resources and expose that content in OWL for validation, reasoning, etc.
... That's a separate serialization that only exists for some resources.
... These resources will have an EXTRA serialization, but the basic serialization will stilll be there. Only the RDF will be over the wire. The OWL will be a different beast.

tony: So there are two forms? A direct RDF representation, and and OWL one?

lloyd: Portions of the VS def we can expose in ways that OWL tooling knows how to handle.
... But not necessarily everything. Full meaning will be in the instance.


paul: Reproducability for round tripping is important, but we need to correctly represent the model to accomplish that.

lloyd: Two separate things happening: turning an arbitrary FHIR instance to RDF; versus taking a structure def or VS def and expose them in OWL to do reasoning on them. Need to do both.

<Marc_Twagirumukiza> We will need to take into account also what the reasoning engines can handle

lloyd: Difference is that a structure def says "FHIR instances look like this", but instance data doesn't do that.

<Marc_Twagirumukiza> So I support to have both

lloyd: There are specific FHIR artifacts where an OWL representation makes sense. But most others make no sense to have OWL rep.

paul: Could just as easily say that Patient should have an OWL rep, but not Observation.

lloyd: A structure def says what elements can exist and cardinality, etc. So that makes sense to map to OWL. But not a patient instance.

tony: Difference between data instance vs schema.

lloyd: There are certain FHIR resources that act as schemas for other FHIR resources. That's a use that is mappable to OWL.

paul: But if you render everything to OWL ...

lloyd: But you can't. You cannot map Patient to OWL (schema level).
... Because there's no notion of gender, DOB, etc in OWL.
... There are certain resources that map to OWL (like structure def, VS def, etc).

line 159-167

tony: lines 171-194 shows unaligned (where VS is not equiv to the code system)

lloyd: should be done by a single algorithm

tony: line 203

lloyd: A code system is not a class, it's an instance. There are concepts in them, but should not try to expose code systems as class.
... Not a VS

tony: code system is a superclass of the concepts that are in it.

lloyd: Don't need to say equivalentClass. You can say they are all instances that have blah.
... From my perspective FHIR VS and code system are disjoint.

tony: They cannot be disjoint because codingBase instance is a member of both VS and code system.

lloyd: CodeBase is not a class member of the code system.

tony: No, it's CodingBase, not CodeBase.

harold: Every instance of Confirmed is an instance of allergy intolerance Code System? Not good.

lloyd: A code system is not a class.

tony: code system is a namespace.

lloyd: There's a linkage between code system and an CodingBase.

marc: We should avoid using subclass

harold: Line 159 says every instance of allergyintolerancestatus is also an instance of a restriction on a Coding system that has value allergy-intolerance-status.
... But the equivalence isn't making sense to me. Concepts are described in coding systems, but are not specializations of them.
... Suggest take one of these options and create a data instance and try to classify it.
... That will lead to surprising results.

Approval of Minutes


Minutes of Aug 11 approved!

Atlanta HL7 face-to-face meeting preparation

RDF subgroup meets Wed 9:00-10:30am

wed oct 7

lloyd: Won't make it.

tony: won't be there, but could call in.

rob: can bring teleconference equipment, but may not be able to join the session.

paul: I will get the equipment from rob.

rob: i'll join for half the meeting.

david: will cancel tomorrow's TF call
... flying tomorrow

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015/09/22 16:40:06 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
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/.../lloyd:/
Succeeded: s/approval minute/Approval of Minutes/
No ScribeNick specified.  Guessing ScribeNick: dbooth
Inferring Scribes: dbooth
Present: Tony_Mallia David_Booth Bill_Kleinebecker Darrell_Woelk Harold_Solbrig Lloyd_McKenzie Marc_T Paul_Knapp Rob_Hausam
Found Date: 22 Sep 2015
Guessing minutes URL: http://www.w3.org/2015/09/22-hcls-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]