W3C

- DRAFT -

Semantic Web Health Care and Life Sciences Interest Group Teleconference

03 Mar 2015

See also: IRC log

Attendees

Present
Claude_Nanjo, David_Booth, Dale_Nelson, Darrell, Guoqian_Jiang, Lloyd_McKenzie, Ingeborg, Marc_Twagirumukiza, Paul_Knapp, Tony_Mallia, Brian_Pech, Erich_Gombocz, EricP, Michel_Dumontier, mscottm, Rob_Hausam, Rafael_Richards, (IRC, only)
Regrets
Chair
David Booth and Paul Knapp
Scribe
rhausam, dbooth

Contents


<trackbot> Date: 03 March 2015

<Lloyd> Will be late connecting to audio

Approve Minutes of previous meetings

<dbooth> Feb 24 minutes: http://wiki.hl7.org/index.php?title=ITS_RDF_Concall_Minutes_20150224

<dbooth> http://wiki.hl7.org/index.php?title=ITS_RDF_Concall_Minutes_20150224

<dbooth> RESOLVED: Feb 24 minutes approved!

FHIR RDF and Validation/Translation Task force kickoff (Fridays 11am Boston timezone)

<inserted> ScribeNick: rhausam

work on mechanics of transformation and validation in RDF

tool is shape expressions

extensions which allow creating artifact like a new document

transform FHIR RDF to FHIR XML

Dale: provide method of validating CCDA via RDF?

discussion of CCDA validation

Eric: validation language may only handle transformation bits

David: new call for transformation on Friday

<inserted> ScribeNick: dbooth

Eric: There is XSLT that translates FHIR XML into FHIR RDF. That XSLT is generated by a script, from the FHIR spec.
... That code could also generate the FHIR ontology.

Tony: Assuming the property resources are the formal declaration, looking at mapping to OWL ontology. Is your tooling robust enough to work?

Eric: It's a proving ground. It's like using chopsticks with chopsticks.
... We can take this approach, and see if there are pieces we want to change.
... The script is python.
... Manual transformations are XSLT.
... With a hand-tooled footer for the XSLT.

<Zakim> dbooth, you wanted to note that another potential route we're exploring is to use JSON-LD, which would mean that any JSON-LD RDF parser could turn FHIR JSON(-LD) into RDF.

Tony: Wise to do the parallel exploration.

David: Task Force call will start this Wednesday 11am Eastern.

Eric: Timezone chaos in the next 2 weeks. All calls are relative to Boston time.

FHIR JSON-LD update (David)

http://dbooth.org/2015/fhir/json-ld/fhir-in-json-ld.pdf

David: FHIR is a closed content model. That's very helpful for the JSON-LD.

Tony: Worth investigating whether the structure represents graphs or only tree structure.

<mscottm> Daylight savings time in the U.S. starts Sunday March 8, in EU, it starts Sunday March 29. In practice, this means people calling in from Europe have to translate their normal call times by -1 hours to keep in sync with Boston, i.e., U.S. will have a *smaller* time difference with EU in the chaotic period. So, this weeks Task Force call will be at 5PM CET as usual for EU participants but next week's will be at 4PM CET and so on until March 29.

<mscottm> yes

<mscottm> Wed is difficult for me.

<Marc_Twagirumukiza> it will be ok for me

<rhausam> Wed works for me

FHIR examples for driving ontology developments (SKIPPED)

Valuesets in FHIR (Lloyd)

<mscottm> Keep in mind that in 3 weeks the call will be at 5PM CET for the long run.

https://global.gotomeeting.com/join/157514853

Access Code: 157-514-853

<Alejandra> Hello all

<Lloyd> http://hl7-fhir.github.io/datatypes.html

<Alejandra> sorry, cannot join the call but wanted to express that Wed would be good for me, thanks!

lloyd: Giving background on how FHIR handles terminology, and how we may use RDF to represent FHIR and link to other ontologies such as SNOMED.
... First look at what datatypes we have: http://hl7-fhir.github.io/datatypes.html
... Then talk about FHIR's notion of bindings, then valuesets.
... At the simplest level is a Code datatype -- a string. Namespace and allowed choices. We use it for status codes, gender, etc.
... Reasonable to use when the set of codes will not change without a major new release of FHIR.

<Alejandra> Great, tx Eric

lloyd: Second type is Coding and CodeableConcept.
... Coding represents a single code: both the code and the code system (URI). Maybe also a version -- important to some systems, if meaning changes bewteen versions.
... Code + System + Version = meaning of the context.

http://hl7-fhir.github.io/datatypes.html#coding

<Marc_Twagirumukiza> -Marc

scribe: The notion of a system being a URI will map well to URI, but version string will not.
... Code system defines what the version string looks like.
... Also the Code itself is not part of the URI.
... We'll avoid spaces in codes for codes we define, but other code systems might have them or other funny chars.

Tony: Is a code unique within a Coding system?
... There's an expectation that we do not have homogeny (same code having differnet meanings). Some systems do that (UCUM). Code "M" could mean "meter" or "milli", but milli would not stand by itself so that isn't a problem.
... But FHIR does not prohibit code systems that do that.
... We have System, Code, Version, and need to be alble to reconstitute them when going between FHIR representations. They can have slashes, hashes, etc., that will make them very difficult to concatenate into a URI.
... Which I think means that RDF will need to retain them as separate properties, and we'll need to use some kind of equivalence to a full URI.

<ericP> example mapping between Coding and URL

Tony: Could have both version code and URI.

lloyd: But the conversion must use a single algorithm that is independent of code system.
... Might have a companion ontology that links FHIR ont to specific coding system.

eric: Could define a URI transformation, but would be nice if the coding system URI is dereferenceable.

lloyd: Can we make an algorithm that would work for them all? Maybe, but scary.

<Lloyd> http://hl7-fhir.github.io/datatypes.html#codeableconcept

ISSUE: How to represent coding system, code and version as round-trippable URIs?

<trackbot> Created ISSUE-3 - How to represent coding system, code and version as round-trippable uris?. Please complete additional details at <http://www.w3.org/2014/HCLS/track/issues/3/edit>.

lloyd: Common with coded data to have more than one code from different coding systems, which are more or less equivalent. That's purpose of CodeableConcept.
... Code from SNOMED might be the first or 5th. People can pass instances that meet a variety of consumers without separate instances for each target.

Eric: If we get a document with a Mayo code, SNOMED code and CPT code, and we spit it back out with all 3 codes in different order, is that ok for round tripping?

lloyd: FHIR rule is that it is an array, and the ordering is significant.
... We don't assign significance in this case, but it would break a hash of the message.

Claude: Can label one as preferred?

lloyd: One can be labeled as Primary. That means "This is the one the user chose". If someone codes after the fact (reading the text), then none would be primary.
... If you're translating, you want to start with Primary.

<scribe> ACTION: David to sent calendar invite for Wed [recorded in http://www.w3.org/2015/03/03-hcls-minutes.html#action01]

<trackbot> 'David' is an ambiguous username. Please try a different identifier, such as family name or username (e.g., dderour, dhansen2, dhau2, dnewman, dshotton, dwild).

ADJOURNED

<scribe> scribenick: rhausam

Summary of Action Items

[NEW] ACTION: David to sent calendar invite for Wed [recorded in http://www.w3.org/2015/03/03-hcls-minutes.html#action01]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015-03-03 17:47:25 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
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/Dale -/Dale:/
Succeeded: s/David -/David:/
Succeeded: s/Friday/Wednesday/
Succeeded: s/zakim aaaa is Tony//
WARNING: Bad s/// command: s/Topic: FHIR RDF and Validation/Translation Task force kickoff (Fridays 11am Boston timezone)//g
WARNING: Bad s/// command: s/Topic: Back to/Topic: FHIR RDF and Validation/Translation Task force kickoff (Fridays 11am Boston timezone)/
Succeeded: s|Topic: FHIR RDF and Validation/Translation Task force kickoff (Fridays 11am Boston timezone)||g
Succeeded: s|Topic: Back to|Topic: FHIR RDF and Validation/Translation Task force kickoff (Fridays 11am Boston timezone)|g
Succeeded: s|FHIR RDF and Validation/Translation Task force kickoff (Fridays 11am Boston timezone) FHIR|FHIR|g
Succeeded: i/work on mechanics/ScribeNick: rhausam
Succeeded: s/Eric -/Eric:/g
Succeeded: i/Eric: There is XSLT/ScribeNick: dbooth
Succeeded: s|s/Topic: FHIR RDF and Validation/Translation Task force kickoff (Fridays 11am Boston timezone)//||
Succeeded: s|s///g||
Found ScribeNick: rhausam
Found ScribeNick: dbooth
Found ScribeNick: rhausam
WARNING: No scribe lines found matching ScribeNick pattern: <rhausam> ...
Inferring Scribes: rhausam, dbooth
Scribes: rhausam, dbooth
ScribeNicks: rhausam, dbooth
Default Present: +1.978.794.aaaa, ericP, DBooth, egombocz, +1.678.999.aabb, Tony, bpech, +1.512.680.aacc, Darrell, +1.604.250.aadd, Pknapp, +1.916.783.aaee, DNelson, +1.801.949.aaff, rhausam, +1.202.528.aagg, Ingeborg, +1.323.444.aahh, +1.507.261.aaii, Marc_Twagirumukiza, +1.507.261.aajj, +31.62.427.aakk, mscottm, +1.310.266.aall, claude, michel, [IPcaller], lloyd
Present: Claude_Nanjo David_Booth Dale_Nelson Darrell Guoqian_Jiang Lloyd_McKenzie Ingeborg Marc_Twagirumukiza Paul_Knapp Tony_Mallia Brian_Pech Erich_Gombocz EricP Michel_Dumontier mscottm Rob_Hausam Rafael_Richards (IRC only)
Found Date: 03 Mar 2015
Guessing minutes URL: http://www.w3.org/2015/03/03-hcls-minutes.html
People with action items: david

[End of scribe.perl diagnostic output]