<trackbot> Date: 29 September 2015

Report from last week's CTS Ontology Workshop

http://ncorwiki.buffalo.edu/index.php/CTS_Ontology_Workshop_2015

dbooth: good workshop; made steps in building bridges

lloyd: what do we want to achieve in this bridge building?

dbooth: ideally, get Barry working with us
... Chris Stoeckert wanted to ensure that the FHIR ont is compatible with BFO
... I encouraged him to join the calls and help.

Approval of minutes

http://wiki.hl7.org/index.php?title=ITS_RDF_Concall_Minutes_20150922

Minutes of Sep 22 approved!

Atlanta HL7 face-to-face meeting preparation

Rob: Good to do a brief overview

Paul: Suggest 10 minutes overview

FHIR ValueSets -- draft by Tony Mallia

Tony: Problems with named classes for coding systems and named classes for concepts
... An OWL restriction statement is a subclass of OWL class, so everytime we put one in square brackets, we're declaring a class.
... In the next version i made them named classes.
... Concerned that we may be negating that approach, because it's more efficient than declaring the same anonymous class multiple times.

http://wiki.hl7.org/index.php?title=File:ValueSet_approachesv3.pdf

Tony: Line 97 has a code system represented as a class.

<ericP> CONSTRUCT { ?x a fhircs:allergy

<ericP> -

<ericP> intolerance

<ericP> -

lloyd: CodeSystem should not be a class -- it's not merely the collection of codings.

tony: a restriction is always a class.

lloyd: the problem is line 105: rdfs:subClassOf fhir:CodeSystem .
... It's defining a set of legal CodingBase instances

tony: agreed -- line 105 should be rdfs:subClassOf fhir:CodingBase .

lloyd: agreed

<ericP> CONSTRUCT {

<ericP> ?x a fhircs:allergy-intolerance-status ;

<ericP> a fhir:CodingBase

<ericP> } WHERE {

<ericP> owl:onProperty fhir:CodingBase.system [

<ericP> fhir:value <http://hl7.org/fhir/cs/allergy-intolerance-status>

<ericP> ]

<ericP> }

<ericP> CONSTRUCT {

<ericP> ?x a fhircs:allergy-intolerance-status ;

<ericP> a fhir:CodingBase

<ericP> } WHERE {

<ericP> ?x fhir:CodingBase.system [

<ericP> fhir:value <http://hl7.org/fhir/cs/allergy-intolerance-status>

<ericP> ]

<ericP> }

<ericP> s/"http://hl7.org/fhir/cs/allergy-intolerance-status"/"http://hl7.org/fhir/cs/allergy-intolerance-status"/

tony: Line 24 uses this allergy-intolerance-status class

dbooth: Reason for representing these all as classes is so that when we express a fhir:Observation for a patient, we can say that that observation is an instance of class allergy-intolerance-status:confirmed, and potentially some other classes too.

lloyd: Nervous about having these as classes, because they are ripe for abuse. People will think of them as semantic classes and try to do equivalent testing on them.
... Someone might say that these two codes are equivalent, and say that they are equivalent classes and that will lead to wrong semantics.
... These classes are not semantic classes, they are just about the strings.
... Alternative is to be verbose: don't define named classes for things that specify both code system and code.

tony: But that's what SNOMED is doing.

<ericP> CONSTRUCT {

<ericP> ?x a allergy-intolerance-status:confirmed ;

<ericP> a allergy-intolerance-status:active ;

<ericP> a fhircs:allergy-intolerance-status

<ericP> } WHERE {

<ericP> ?x a fhircs:allergy-intolerance-status ;

<ericP> fhir:CodingBase.code [

lloyd: But that's outside of our stuff. Bridging ont will have to do that.

<ericP> fhir:value "confirmed"^^xsd:token ] }

eric: Another way to look at this. If you give people names for these, they might make erroneous statements about them.
... Safer thing is to name them, but document them warning not to make semantic statements about them.

<Zakim> dbooth, you wanted to say Maybe we can mitigate Lloyd's concern by how we name these syntactic classes

<ericP> let's name them with names that use unicode characters that make their eyes bleed

dbooth: For example, in programming, internal entities are often named starting with underscore: _foo

eric: everywhere you would otherwise write the class name, you write the whole anonymous expression.

lloyd: When we map to an external ont, we're mapping using the concepts, not the codes that signify the concepts.
... I think we can do this having names for the concepts but not for the CodingBases.

ADJOURNED

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/____/Stoeckert/
WARNING: Bad s/// command: s/<http://hl7.org/fhir/cs/allergy-intolerance-status>/"http://hl7.org/fhir/cs/allergy-intolerance-status"/
Succeeded: s|<http://hl7.org/fhir/cs/allergy-intolerance-status>|"http://hl7.org/fhir/cs/allergy-intolerance-status"|
No ScribeNick specified.  Guessing ScribeNick: dbooth
Inferring Scribes: dbooth

WARNING: Dash separator lines found.  If you intended them to mark
the start of a new topic, you need the -dashTopics option.
For example:
        <Philippe> ---
        <Philippe> Review of Action Items

Present: Adam_(webex_only) Bill_Kleinebecker Rob_Hausam Darrell_Woelk Lloyd_McKenzie Paul_Knapp Steve_(webex_only) Tony_Mallia
Found Date: 29 Sep 2015
Guessing minutes URL: http://www.w3.org/2015/09/29-hcls-minutes.html
People with action items: