Notes from today's call are here:
http://www.w3.org/2014/09/23-hcls-minutes.html
and also below in plain text.
----------------------------------------------
[1]W3C
[1] http://www.w3.org/
- DRAFT -
HCLS
23 Sep 2014
See also: [2]IRC log
[2] http://www.w3.org/2014/09/23-hcls-irc
Attendees
Present
+1.415.740.aaaa, +1.978.794.aabb, DBooth, ericP, Tony,
+1.510.418.aacc, +1.301.825.aadd, +44.792.050.aaee,
HartC, Ingeborg, +1.310.266.aaff, Claude, Guoqian,
+1.323.444.aagg, Mehmet, +1.323.444.aahh
Regrets
Chair
EricP
Scribe
ericP
Contents
* [3]Topics
1. [4]FDA Therapeutic area wrapup
2. [5]HL7 meetings and proposed WG on RDF for Semantic
Interoperability
* [6]Summary of Action Items
__________________________________________________________
<egombocz> aaaa is egombocz
<dbooth> zakim is doing weird things -- not accepting codes but
hanging up
-> [7]http://www.w3.org/2014/Talks/0924-FDA_TA-egp/ slides
[7] http://www.w3.org/2014/Talks/0924-FDA_TA-egp/
FDA Therapeutic area wrapup
<HartC> yes.
<Marc_AgfaHealthcare> Hi this is Marc (Agfa Healthcare-Belgium)
joining
<dbooth> _: Using LOINC or SNOMED CT codes?
<dbooth> Eric: Will get to that
<dbooth> Peter: When the RIM first came out, it was based on
entities and roles particiatping in acts. But groups kept
adding to it and it got so big that nobody wanted to deal with
it. hence FHIR. But the basic idea of entities and roles
participating in acts.
<dbooth> ... Top of ontology should at least start with
Entities, Roles, Participations and Actions. Probably also
Substances and Procedures.
<dbooth> ... In Protege OWL model, everything is was directly
under owl:Thing originally.
<dbooth> Eric: I see data entities under owl:Thing.
<dbooth> Peter: Yes, that's what I was objecting to. That's why
I suggested having those other classes directly under
owl:Thing.
<dbooth> Peter: Observation is under Act, because it is an act
of observation.
<dbooth> ... They're participating in an observation.
<dbooth> Claude: We're also doing a CDS model on top of FHIR,
and distinguishh between an act and an observation, which is
what you observed.
<dbooth> Eric: The HL7 use case are typically not interested in
the physiological processes, but more interested in the speech
acts involved.
<dbooth> Peter: The actual pathology should be in LOINC or
SNOMED CT etc.
<dbooth> Eric: There were pictures that says "a procedure event
has an act relationship to an observation event".
<dbooth> Peter: All of the things I saw in the FHIR model could
have been under one of the 5 main classes.
<egombocz> agree with Peter strongly
<Marc_AgfaHealthcare> There are couple of confusing definitions
regarding the observations/observables and
acts/activities/procedure. Mainly due to different
terminologies: SNOMED CT, BRIDG, ETC. Here we need to stick on
one definition OR mapping
<dbooth> Peter: BRIDG will be adding those 5 top level classes.
<dbooth> Claude: In the CDS we're building a more layered ont
on top of FHIR, which aligns fairly well with the RIM.
<Marc_AgfaHealthcare> will BRIDG be influenced by SNOMED CT in
creating those 5 classes?
<dbooth> ... For CDS we need to be able to have core concepts
for FHIR. I could present what we've done so far, for feedback.
<dbooth> Peter: I'm no longer defending that it has to be the
RIM, because the RIM is too heavy to survive.
<dbooth> ... But they can use the basic notions of Entities and
Roles participating in Acts.
<dbooth> Claude: Could you map RIM to BRIDG to the same effect?
<dbooth> Peter: You can give me any one thing and I can tell
you which of the 5 classes is belongs in.
-> [8]http://www.bridgmodel.org/owl/3.2 bridgmodel3.2.owl
[8] http://www.bridgmodel.org/owl/3.2
<dbooth> Peter: Ont for RIM is big. They took a given subclass
of Act and made lists of codes that were legitimate for the
slots and put them all in.
<dbooth> ... Lloyd put thousands of vocabulary terms.
<dbooth> Marc asks: ill BRIDG be influenced by SNOMED CT in
creating those 5 classes?
<dbooth> Peter: No, SNOMED CT does not talk about them at all.
<dbooth> ... They touch each other but they are not mixed up.
<Marc_AgfaHealthcare> Then it will be somehow in line with RIM?
<dbooth> (back to EricP slide 7)
<dbooth> slide 9
<dbooth>
[9]http://www.w3.org/2014/Talks/0924-FDA_TA-egp/#%289%29
[9] http://www.w3.org/2014/Talks/0924-FDA_TA-egp/#%289%29
<dbooth> DBooth: Is Organizer class good or bad?
<dbooth> EricP: Seems necessary for organizing.
<dbooth> Guoqian: Is this ont available somewhere?
<dbooth> Eric: It's linked at the front of the slides
<dbooth> Guoqian: Nice how you're handling the BRIDG
complexity. How do you capture the domain expertise from domain
experts?
->
[10]http://www.w3.org/2013/12/FDA-TA/doc/RenalTransplantation-c
map.html Rtrans cmap
[10]
http://www.w3.org/2013/12/FDA-TA/doc/RenalTransplantation-cmap.html
<dbooth> Eric: This CMAP links into BioPortal
<dbooth> Eric: We make a CMAP with the domain expert and then
turn it into an ont.
<dbooth> ... I hacked the CMAP SVG to add the links to
BioPortal, after exporting the SVG.
<dbooth> Guoqian: BRIDG is so complex. How do you make the
domain experts understand it?
<dbooth> Eric: We used a part of BRIDG
<dbooth> Guoqian: Any connection between this work and CIMI?
<dbooth> Eric: No, and that's unfortunate.
<dbooth> ... But Harold Solbrig is working on modeling CIMI
using Shape Expressions.
<dbooth> Peter: I don't find CIMI reference model very useful
because it has classes like Item, Collection, Element, etc.,
which are like boxes within boxes.
<dbooth> ... The names of the classes don't have much meaning
associated.
<dbooth> Guoqian: The advantage to use CIMI is that if other
models like FHIR, SNOMED CT, etc. were in CIMI that we could
related different models.
<dbooth> Peter: I don't see that you gain anything. Much better
off putting them in the 5 backbone classes.
HL7 meetings and proposed WG on RDF for Semantic Interoperability
<scribe> scribenick: ericP
claude nanjo and i attended
<dbooth> I spoke to: Grahame Grieve, FHIR architect
<dbooth> Charlie McKay
<dbooth> Cecil Lynch
<dbooth> Paul Knapp
<dbooth> Woody Beeler
<dbooth> Bernd Blobel
<dbooth> Ken McCaslin
<dbooth> John Quinn
dbooth: claude nanjo and i attended
<dbooth> Charter draft:
[11]http://dbooth.org/2014/hl7/rdf-semantic-interop-wg-v7.doc
[11] http://dbooth.org/2014/hl7/rdf-semantic-interop-wg-v7.doc
dbooth: i flew out to talk about starting the HL7 RDF for
Semantic Interoperability WG
... we prepared a few slides.
<egombocz> need to leave, thank you
<HartC> sorry, got to go.
<dbooth> ITS work group:
[12]http://www.hl7.org/Special/committees/xml/overview.cfm
[12] http://www.hl7.org/Special/committees/xml/overview.cfm
<dbooth> AID work group:
[13]http://www.hl7.org/Special/committees/java/overview.cfm
[13] http://www.hl7.org/Special/committees/java/overview.cfm
<dbooth> Peter: When you're doing things at HL7, are you
thinking of using RDF as data stores? Vs to communicate between
black boxes? Or as a way to define models?
<dbooth> Peter: RDF triples don't stand by themselves. They
need the definitions, which are about the modeling.
<dbooth> ... There is interest in doing the modeling in RDF and
OWL.
<dbooth> ... In Kaiser we have K-Con. We start with OWL and
underneath it using Protege and OWL we come up with everything
that would be needed to assess that disease.
<dbooth> Peter: RDF Fox sits on top of triplestore -- a fast
OWL RL reasoner from Ian Horrock's group.
<dbooth> ... Take a FHIR resource, start with OWL ontology with
5 upper classes, fill out the OWL ont first of what the model
should look like, then figure out how to parse FHIR into that.
<dbooth> ... You (David) could present this proposal at our
committee.
<dbooth> [14]http://dbooth.org/2014/hl7/rdf-for-interop.pdf
[14] http://dbooth.org/2014/hl7/rdf-for-interop.pdf
<dbooth> Claude: More support for having this effort be a part
of the ITS group than a separate WG.
<dbooth> ... This would have cross-cutting concerns, connecting
to a number of WGs. One of the big goals also is to bridge W3C
and HL7 communities.
<dbooth> ... If we were to go forward as a subgroup, which WG
would make most sense? The people that we spoke to at HL7
thought ITS. What do you (peter) see as pros/cons of AID group?
<dbooth> Peter: Our group started as the Java SIG, when the RIM
was new.
<dbooth> ... Gunther was my co-chair, and he was main author of
RIM.
<dbooth> ... Format could create Java classes from the RIM. But
people said they used other languages. So we changed our name
to RIMBA.
<dbooth> ... so you can use it for generating tables, etc. too.
You can take CDA messages, parse them into RIM and put them
out.
<dbooth> ... Then FHIR came along. Others in HL7 make the
stnadards, and we use them, so we changed our name again.
<dbooth> ... If it's "how do you do X" then it's our group. If
it's how to serialize in XML, then it's ITS.
<dbooth> ... If it's "how to think about things in a different
way" then it's our group. If it's "how to make the syntax so
that everyone agrees" then it's ITS.
<dbooth> David: Does AID have regular teleconferences?
<dbooth> Peter: No. When we meet we have pre-invited people to
present their approaches.
<dbooth> ... No weekly meetings.
<drjava> wiki.hl7.org
<dbooth> ACTION: David to try to schedule a presentation to ITS
and AID HL7 WGs to discuss the proposed "RDF for Semantic
Interoperability" WG. [recorded in
[15]http://www.w3.org/2014/09/23-hcls-minutes.html#action01]
<dbooth> ADJOURNED
Summary of Action Items
[NEW] ACTION: David to try to schedule a presentation to ITS
and AID HL7 WGs to discuss the proposed "RDF for Semantic
Interoperability" WG. [recorded in
[16]http://www.w3.org/2014/09/23-hcls-minutes.html#action01]
[End of minutes]
__________________________________________________________
Minutes formatted by David Booth's [17]scribe.perl version
1.138 ([18]CVS log)
$Date: 2014-09-23 17:01:20 $
__________________________________________________________
[17] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm
[18] http://dev.w3.org/cvsweb/2002/scribe/
Scribe.perl diagnostic output
[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.138 of Date: 2013-04-25 13:59:11
Check for newer version at [19]http://dev.w3.org/cvsweb/~checkout~/2002/
scribe/
[19] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/
Guessing input format: RRSAgent_Text_Format (score 1.00)
Succeeded: s/Quo/Guo/
Succeeded: s/This/This CMAP/
Succeeded: s/HL7 meetings/HL7 meetings and proposed WG on RDF for Semant
ic Interoperability/
Found ScribeNick: ericP
Inferring Scribes: ericP
Default Present: +1.415.740.aaaa, +1.978.794.aabb, DBooth, ericP, Tony,
+1.510.418.aacc, +1.301.825.aadd, +44.792.050.aaee, HartC, Ingeborg, +1.
310.266.aaff, Claude, Guoqian, +1.323.444.aagg, Mehmet, +1.323.444.aahh
Present: +1.415.740.aaaa +1.978.794.aabb DBooth ericP Tony +1.510.418.aa
cc +1.301.825.aadd +44.792.050.aaee HartC Ingeborg +1.310.266.aaff Claud
e Guoqian +1.323.444.aagg Mehmet +1.323.444.aahh
Got date from IRC log name: 23 Sep 2014
Guessing minutes URL: [20]http://www.w3.org/2014/09/23-hcls-minutes.html
People with action items: david
[20] http://www.w3.org/2014/09/23-hcls-minutes.html
[End of [21]scribe.perl diagnostic output]
[21] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm