Harold and Grahame will be at Dev Days next week Wed-Fri.
<scribe> ACTION: Harold will sched teleconference with Grahame at Dev Days and notify the rest of us about time/day
<trackbot> Created ACTION-91 - Will sched teleconference with grahame at dev days and notify the rest of us about time/day [on Harold Solbrig - due 2017-11-14].
(Harold presents slides)
harold: In our case OWL consists
of a pile of RDF with an OWL header
... Also need to get code systems and valuesets. They have a
FHIR RDF format, but we need it as an OWL ont.
... I.e., using rdfs:subClassOf etc.
... Also FHIR people treat Turtle and RDF as synonymous, but
some may need to server other formats.
... We may need to differentiate OWL from FHIR/RDF.
eric: It helps putting the OWL
declaration in the footer. People do not complain as
much.
... When people put a valueset in OWL, they decide whether to
use individuals or classes. We're invested in teh classes
approach, both for inference and for the type connections in
FHIR.
harold: We do care about it, but
we need to bring it up with the terminology group. Rector and
others did a lot of work on that.
... Some things break if you use the class approach.
... There's an lengthy ___ about how valuesets should be
handled. We need to refer to that before proceeding.
ISSUE: Need to follow Rector's guidance on handling valuesets in OWL
<trackbot> Created ISSUE-45 - Need to follow rector's guidance on handling valuesets in owl. Please complete additional details at <http://www.w3.org/2014/HCLS/track/issues/45/edit>.
dbooth: Why need to differentiate OWL from FHIR/RDF?
harold: The footer that declares
it as an ontology.
... Two representations of code systems: the structure def and
the RDF that defines the OWL semantics.
eric: FHIR code knows how to
handle codes, systems, etc. But when we put in in Turtle, it is
hard to discriminate the same formate but expressed in OWL. In
order to build the class hierarchy.
... But the same media type text/turtle is currently used for
both.
harold: In the FHIR spec need to
define what it means to be an RDF compliant FHIR service.
... Also need to define a test suite for that.
ISSUE: Is mime type application/json-ld still used? If not, need to take it out of the FHIR spec
<trackbot> Created ISSUE-46 - Is mime type application/json-ld still used? if not, need to take it out of the fhir spec. Please complete additional details at <http://www.w3.org/2014/HCLS/track/issues/46/edit>.
https://www.w3.org/TR/owl-ref/
https://www.w3.org/TR/n-triples/#n-triples-mediatype
<ericP> application/n-triples per https://www.w3.org/TR/n-triples/#n-triples-mediatype
dbooth: Maybe use a format= or semantics= query parameter to distinguish RDF from OWL
eric: purists would want a different URL for them, since they carry different information
harold: it might also go in the URL after the version
eric: Am I supposed to get back a Location header that says more precisely which version of the resource was serverd?
harold: It's in the fhir resource meta field inside the resource itself. They do say to return the precise URL in the Location header.
(slide: "One more thing")
harold: The FHIR build server is
not very usable for testing, because anything that redirects
goes back to the main FHIR server.
... In our ont header we need to add the owl:versionIRI if
there is one.
eric: If I have observation/1234/version/1 does it have the same ont as observation/1234/version/2 ?
harold: We could add /owl to the URL to indicate the OWL format.
<scribe> ACTION: Harold to clean up slides on FHIR RDF conneg and give to David to post online
<trackbot> Created ACTION-92 - Clean up slides on fhir rdf conneg and give to david to post online [on Harold Solbrig - due 2017-11-14].
ADJOURNED
This is scribe.perl Revision: 1.152 of Date: 2017/02/06 11:04:15 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00) Succeeded: s|https://hangouts.google.com/_/elUi/chat-redirect?dest=https%3A%2F%2Fwww.w3.org%2FTR%2Fowl-ref%2F|| Present: Luis_Marco EricP Harold_Solbrig David_Booth No ScribeNick specified. Guessing ScribeNick: dbooth Inferring Scribes: dbooth Found Date: 07 Nov 2017 People with action items: harold WARNING: IRC log location not specified! (You can ignore this warning if you do not want the generated minutes to contain a link to the original IRC log.)[End of scribe.perl diagnostic output]