See also: IRC log
please join: https://www.w3.org/community/hclscg/
http://wiki.hl7.org/index.php?title=RDF_for_Semantic_Interoperability#Deliverables_and_Editors
grahame's editing policy: http://wiki.hl7.org/index.php?title=RDF_for_Semantic_Interoperability#Edits_to_the_FHIR_RDF.2Fontology_pages
eric worked on this page: http://w3c.github.io/hcls-fhir-rdf/spec/rdf.html
as a potential new version of https://hl7-fhir.github.io/rdf
eric: I propose renaming the ontology page to Linked Data.
dbooth: The proposal: 1. rename
ontology-module.html to linked-data-module.html 2. change the
title to "FHIR Linked Data Module".
... 3. On the FHIR home page, put links to RDF and Ontology
pages in the Linked Data module box
eric: Another proposal: rename
the title of page https://hl7-fhir.github.io/rdf
to match the XMl and JSON pages: "RDF Representation of
Resources"
... Regarding deliverable 4. Formal specification of FHIR/RDF
http://w3c.github.io/hcls-fhir-rdf/spec/
ADJOURNED
<trackbot> Meeting: Semantic Web Health Care and Life Sciences Interest Group Teleconference
<trackbot> Date: 23 August 2016
https://github.com/w3c/hcls-fhir-rdf/issues/18
dbooth: potential options: blank
node, relative URI, or skolem URI
... A relative URI looks like <>
eric: i favor <> (the base URI)
<ericP> GET http://a.example/some/path -> <> a fhir:Foo; fhir:identifier [ ... ]
grahame: That would allow only one unidentified URI in scope at at time?
<ericP> <http://a.example/some/path> rdf:type fhir:Foo .
<ericP> <http://a.example/some/path> fhir:identifier [ ... ].
grahame: If I have two unidentified resources in the same graph then this would not work.
<ericP> GET http://a.example/some/path -> <other/path> a fhir:Foo; fhir:identifier [ ... ]
<ericP> <http://a.example/some/other/path> rdf:type fhir:Foo .
<ericP> <http://a.example/some/other/path> fhir:identifier [ ... ].
dbooth: I'm against relative URI because they don't exist at the RDF level. In RDF *every* URI is absolute.
grahame: this example is different: http://hl7-fhir.github.io/parameters-example.ttl.html
<Grahame> ....Eric made me do it....
dbooth: i think a blank node would be best aligned with the RDF model. A relative URI would become absolute when interpreted as RDF, and I'm concerned that it would not round trip well
grahame: XML and JSON forms do not include the address of the content. The RDF form does. That's why we say that the Turtle is for reasoning, rather than interop.
dbooth: We haven't nailed down what constitutes round tripping.
grahame: For xml and json we say that the canonical form must be identical.
ISSUE: Define canonical FHIR RDF https://github.com/w3c/hcls-fhir-rdf/issues/34
<trackbot> Created ISSUE-39 - Define canonical fhir rdf https://github.com/w3c/hcls-fhir-rdf/issues/34. Please complete additional details at <http://www.w3.org/2014/HCLS/track/issues/39/edit>.
eric: Do we need to specify how the root node should be named?
grahame: If we say that it is a URI, but might not be unique, then if people merge data that could be a problem.
dbooth: I.e., if a relative URI is used <> then different data might accidentally use the same URI and users would not realize it.
eric: An advantage of a relative URI is that they can do sparql queries using that node and its name is stable.
grahame: We could put a warning in the spec.
dbooth: in round tripping RDF-->XML/JSON-->RDF, I don't want my URIs disappearing.
grahame: There's no place to put that URI in the XML or JSON
dbooth: We might agree to let the
user decide whether to use relative, abs or blank node.
... What should we put in the examples?
eric: Relative URIs
dbooth: still against relative
URIs in the examples.
... Prefer blank node used in the examples.
Eric agrees to blank node
AGREED: In the examples, we will use a blank node for the root node of the resource
<scribe> ACTION: Eric to draft warning text about using relative URIs for the root node [recorded in http://www.w3.org/2016/08/23-hcls-minutes.html#action01]
<trackbot> Created ACTION-67 - Draft warning text about using relative uris for the root node [on Eric Prud'hommeaux - due 2016-08-30].
ADJOURNED
This is scribe.perl Revision: 1.144 of Date: 2015/11/17 08:39:34 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: i/I have made the request/Topic: ================== 5pm Call ================= Succeeded: s/ls// No ScribeNick specified. Guessing ScribeNick: dbooth Inferring Scribes: dbooth WARNING: Replacing previous Present list. (Old list: Rob_Hausam, David_Booth, Eric_P, Harold_Solbrig) Use 'Present+ ... ' if you meant to add people without replacing the list, such as: <dbooth> Present+ EricP, David_Booth, Harold_Solbrig, Rob_Hausam Present: EricP David_Booth Harold_Solbrig Rob_Hausam (5pm_call) Thomas_Lukasik David_Goodall Grahame_Grieve Brian Scheller Found Date: 23 Aug 2016 Guessing minutes URL: http://www.w3.org/2016/08/23-hcls-minutes.html People with action items: eric[End of scribe.perl diagnostic output]