15:02:43 <RRSAgent> RRSAgent has joined #hcls
15:02:44 <RRSAgent> logging to http://www.w3.org/2016/11/01-hcls-irc
15:02:45 <trackbot> RRSAgent, make logs world
15:02:46 <Zakim> Zakim has joined #hcls
15:02:47 <trackbot> Zakim, this will be HCLS
15:02:48 <Zakim> ok, trackbot
15:02:49 <trackbot> Meeting: Semantic Web Health Care and Life Sciences Interest Group Teleconference
15:02:49 <trackbot> Date: 01 November 2016
15:02:58 <dbooth> Chair: David Booth
15:04:11 <hsolbrig> hsolbrig has joined #hcls
15:04:12 <dbooth> Topic: Should we feature the JSON-LD format of FHIR more prominently?
15:07:31 <dbooth> FHIR now is now structured into modules: http://build.fhir.org/index.html
15:08:27 <dbooth> The Linked Data module: http://build.fhir.org/ontology-module.html
15:08:45 <dbooth> (Currently called the "Ontology Module", but will be rename to "Linked Data Module"
15:08:47 <dbooth> )
15:13:17 <dbooth> Suggestion for links to FHIR examples is to be listed "XML JSON Linked Data (JSON-LD or Turtle)"
15:14:19 <dbooth> that would appear in each example table like this one: http://build.fhir.org/allergyintolerance-examples.html
15:20:54 <dbooth> Eric: I prefer "RDF" there instead of "Linked Data", because it is more specific
15:21:34 <dbooth> AGREED: We'll see if we can change the example links to say "XML JSON RDF (JSON-LD or Turtle)"
15:21:50 <dbooth> Topic: Aligning custom resources with data dictionaries (EricP)
15:23:04 <dbooth> http://wiki.hl7.org/index.php?title=FHIR_Custom_Resources
15:23:09 <dbooth> http://fhir.hl7.org.au/fhir/rcpa/Colorectal-Colorectal-genexample-1.xml.html
15:24:17 <ericP> -> http://rdf-vocabulary.ddialliance.org/discovery.html#fig-vocabulary-overview DISCO <whoo!>
15:24:23 <dbooth> Topic: fhir.schema.org
15:24:31 <ericP> -> https://www.w3.org/2016/11/ddi-disco/ examples
15:25:55 <dbooth> harold: Next steps were to come up with use cases.  We're going to pick up that thread and show something end of Nov, to have a poster for it at SWAT4LS
15:26:20 <dbooth> Topic: Aligning custom resources with data dictionaries (EricP)
15:28:02 <dbooth> eric: With the excitement around FHIR, people are creating custom resources
15:28:05 <ericP> -> http://wiki.hl7.org/index.php?title=FHIR_Custom_Resources Custome Resources
15:28:16 <ericP> -> http://fhir.hl7.org.au/fhir/rcpa/prostate.html#table example
15:28:39 <dbooth> eric: People say "FHIR is great!  We'll do everything in FHIR!"
15:29:46 <dbooth> ... Custom resources are being used for aggregates of data plucked from other resources.
15:31:03 <dbooth> ... Can I turn existing data into custom resources?
15:31:34 <dbooth> ... Variables from a study become attributes in a FHIR resource.
15:31:57 <dbooth> ... We are spitting out FHIR structure definitions.
15:32:16 <dbooth> ... Is this worthwhile?
15:33:40 <dbooth> dbooth: That allows the FHIR protocols to be reused, but others will not understand the custom resources.
15:35:54 <dbooth> eric: Harmonization between custom resources: two studies use some of the same questions, using the same URLs.
15:36:14 <ericP> https://www.w3.org/2016/11/ddi-disco/
15:42:25 <dbooth> eric: Goal here is, DDI is well ahead of FHIR in terms of these aggregations.  If we provide a mechnical transformation to do a schema-schema transformation to produce a FHIR struture def from a DDI def, then ittakes advantage of what the DDI folks have done and brings it into FHIR.
15:49:04 <dbooth> dbooth: So one main benefit is that this causes DDI to coordinate consistent semantics between same-named/same-URI FHIR elements that appear in different custom FHIR resources?   And the other benefit is the auto-generation of the custom FHIR structure defs?
15:50:34 <dbooth> eric: Not quite.  The sameness is ensured by generating from the same DDI dictionary.
15:53:07 <dbooth> dbooth: How will someone seeing only custom resource A and custom resource B know that element X in A is the same as element Y in B?
15:53:23 <dbooth> eric: Additional triples are needed to say that.
15:54:58 <dbooth> eric: FHIR does not have any way to indicate that two elements in different resources have the same meaning.
15:56:02 <dbooth> dbooth: Does this process at least ensure that the FHIR local names are the same?
15:56:50 <dbooth> eric: No, it depends on whether it is called the same variable in different studies.
15:57:14 <dbooth> eric: Lloyd has on his to-do list to figure out how to express sameness across FHIR elements
15:58:21 <dbooth> ADJOURNED
17:32:49 <Zakim> Zakim has left #hcls
17:39:25 <dbooth> rrsagent, draft minutes
17:39:25 <RRSAgent> I have made the request to generate http://www.w3.org/2016/11/01-hcls-minutes.html dbooth
17:50:16 <dbooth> Present: EricP, David Booth, Harold Solbrig
17:50:19 <dbooth> rrsagent, draft minutes
17:50:19 <RRSAgent> I have made the request to generate http://www.w3.org/2016/11/01-hcls-minutes.html dbooth