See also: IRC log
trackbot, start meeting
<trackbot> Date: 01 September 2015
Paul: Head up Rosetta effort for
PCD domain. Medical devices. LOINC project mapped 600 most
popular terms.
... Also heavy on valueset constraints. Want rigorous testing
of EMR data from devices.
http://wiki.hl7.org/index.php?title=File:Vocabulary_and_Terminology_in_RDFv2.pdf
lloyd: line 50,
allergy-intolerance-status needs to be a blank node, because
there can be extensions on the URI
... I could have the same system in 3 different places with 3
different extensions, so it needs to be a blank node.
tony: ok
lines 93-99
lloyd: Why have a class for the
CodeSystem?
... Why not have 2 owl:hasValues?
tony: cleaner this way.
lloyd: We'll have instances w no
predefined knowledge of the codeSystems or their versions.
E.g., string values for them.
... I'm concerned that this approach may cause challenges.
tony: That's ok, but the reasoner won't infer any type for it.
david: Should we look at both options?
<scribe> ACTION: Tony to draft a new version that shows both approaches (two owl:hasValues that Lloyd suggested, versus Tony's existing draft) [recorded in http://www.w3.org/2015/09/01-hcls-minutes.html#action01]
<trackbot> Error finding 'Tony'. You can review and register nicknames at <http://www.w3.org/2014/HCLS/track/users>.
lloyd: where is the OID coming
from?
... Resource instances won't have OIDs.
... Namespaces I thought we wouldn't expose in RDF.
tony: CodeSystems have OIDs
lloyd: I wouldn't expect to see any of that in RDF?
eric: they are in the FHIR examples
tony: We don't need them.
lloyd: If they're in the
examples, the should all be URIs, as "urn:..."
... lines 105-109 comes from where, what FHIR artifact? Needs
to come from a structure def, valueset, or something. What XML
file?
... NamingSystem is used to define both CodeSystems and
identifier systems.
tony: We can leave out lines
103-110
... but where does lines 117-121 come from?
... lines 123-128, Authority is only for housekeeping, saying
what comes from where. Comes from the namespace of the
ontology.
lloyd: FHIR ont is generated purely from FHIR artifacts -- no knowledge of LOINC or other ont.
tony: Want to know how the ont pieces are assembled to a combined ontology that allows the reasoner to work.
lloyd: We'll show for a couple of terminologies how to link them in.
david: Need to see an example of how Authority is used.
line 130
lloyd: Concepts can have more than one code.
eric: Two concepts that are equivalent, or a single concept with multiple codes?
lloyd: same concept can be
identified by different combination of code system, code and
version.
... But ValueSet can prohibit certain codes (even though they
may identify the same concept)
david: This could be modeled either way: as a single concept, with multiple codes; or as one concept per code, and those concepts are equivalent.
eric: Do i want to be able to validate ValueSets in OWL, versus ShEx or something else?
lloyd: OWL
... Should change line 130 to say "one or more code
representations within that code system"
<scribe> ACTION: Tony to change line 130 to say "one or more code representations within that code system" [recorded in http://www.w3.org/2015/09/01-hcls-minutes.html#action02]
<trackbot> Error finding 'Tony'. You can review and register nicknames at <http://www.w3.org/2014/HCLS/track/users>.
lloyd: Also need to bring in
version.
... Question of whether a single concept can span
versions.
... Same code and system may be bound to different concepts for
different versions, or sometimes the same concept.
... Some code systems reuse code strings after 10 years,
meanign completely different things.
... And some change the hierarchy and properties between
versions, even though the display name didn't change.
... There are also fine-grained changes between versions.
tony: I'll address version later.
lines 133-150
lloyd: Abstract Concepts are declared wrt a particular intended use.
lines 179-189
lloyd: Is there any chance that someone else would have a concept with that snomed URI, <http://snomed.info/sct> ?
tony: We can change that to be a owl:hasValue against that string, instead.
lloyd: HL7 doesn't own that namespace URI, so HL7 cannot define what it means.
david: that would be URI squatting.
<scribe> ACTION: Tony to change line 182 to use owl:hasValue restriction instead of URI squatting [recorded in http://www.w3.org/2015/09/01-hcls-minutes.html#action03]
<trackbot> Error finding 'Tony'. You can review and register nicknames at <http://www.w3.org/2014/HCLS/track/users>.
eric: For orgs that cannot manage their URIs well, HL7 could provide URIs for them. But some orgs can manage them well and ensure that their URIs are uniquely defined.
tony: If we use owl:hasValue then
we don't have to worry about collisions.
... But if we use that URI directly as an RDF concept, then we
risk URI collision.
eric: Can we get them to give us a safe URI?
lloyd: in theory, but that's probably 1% of the vocabularies.
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/Topic: Next Steps// No ScribeNick specified. Guessing ScribeNick: dbooth Inferring Scribes: dbooth Present: Paul____ Lloyd EricP Tony David Rob_Hausam Found Date: 01 Sep 2015 Guessing minutes URL: http://www.w3.org/2015/09/01-hcls-minutes.html People with action items: tony[End of scribe.perl diagnostic output]