See also: IRC log
<Claude> 1. Please join my meeting.
<Claude> https://global.gotomeeting.com/join/275068589
<Claude> 2. Use your microphone and speakers (VoIP) - a headset is recommended. Or, call in using your telephone.
<Claude> Dial +1 (213) 929-4130
<Claude> Access Code: 275-068-589
<Claude> Audio PIN: Shown after joining the meeting
<Claude> Meeting ID: 275-068-589
<Claude> GoToMeeting®
<Claude> Online Meetings Made Easy®
<sundaram> hi
<ericP> https://global.gotomeeting.com/join/275068589
<ericP> nested FHIR extensions in XML
are we using the GoToMeeting phone line, or the W3C phone line?
<ericP> nested FHIR extensions in RDF
Tony and I are on the GTM phone line
<ericP> dbooth, w3c phone line
ok, we'll switch
<ericP> 4257#
<Claude> https://global.gotomeeting.com/join/275068589
Tony: Worked on healthcare
interop for 15 years. Worked on eHealth exchange.
... Started looking on how to express FHIR payload in OWL DL.
How to express instance exchange in FHIR, in DL?
... Lots of interest in bringing SNOMED CT, ITSD, and others
into DL to blend them.
... Looks like it sort of works in Protege.
Jay: Jay Sundaram, want to find
out state of the art. Spent last 14 years as bioinformatics
eng. Worked at institute for Genomic research.
... In 2003 decided we'd use community defined standards.
Bioinformatics sequence language. GMOD community provided
database schema.
... Key piece was controlled vocabularies in OBO. Also taking
courses in Hopkins and took Sem Web course, piqued interest in
representing datasets in RDF. In AstraZeneca, working on
biomarkers. Want to enhance data models to use RDF to see what
inferences can be made by leveraging ont.
Eric: This call today is mostly phenotype side. There's another call focused on the genotype side.
<Claude> http://www.hl7.org/implement/standards/fhir/extensibility.html
<Zakim> dbooth, you wanted to note also the dual use of the "status" predicate (for both datatype and object type) and to say I've been thinking of doing non-monotonic extensions in
Eric: Danger of usign a named graph for non-monotonic extensions with naive representation is that they might accidentally merge the graphs and make wrong conclusions.
David: yes, that's a danger. My motivations in suggestng the use of named graph for non-monotonic: (a) mostly the same software might be used, so easier if the representation is the same; (b) easier for human to grok if the representatino is the same.
Tony: I took a totally different
approach.
... based on classes.
... For example, "Active" would be a class, so everything with
an Active status would be a member of that class.
... I believe that terms in a controlled vocabulary should be
viewed as types in RDF.
Eric: If Active is a class then you can do restriction classes. You can make the class representation from Claude's way of doing it using restriction classes (anonymous classes) with equivalences. This would give Tony the effect that he wants without a named class.
Tony: I said the observation instance is a member of the class and a bunch of other classes also.
David: A key use case for me is to make the resulting RDF simple to read and use, so that people who are not RDF heads can look at it without being intimidated, and understand it.
Claude: Right now at HL7 we have a model/terminology connundrum because they have different granularity. But if they are all represented in teh same medium (RDF) then that distinction can be overcome.
<ericP> terminfo prob
Claude: Can get from one ontology to another using transformations, or a common core.
Eric: Charlie and I worked on
eliminating that boundary between terminology and models.
... SNOMED has pre and post coordinated terms that need to be
reconciled.
... Need to microparse: have an info model and extra data that
needs to be parsed into the model.
<ericP> semantics in terminology
<ericP> in information model
<sundaram> thank you all. interesting discussion. i'll check in again next week. have to get ready for another meeting.
Eric: Can SNOMED precoordinated terms have a mapping to semantics? E.g., that have both diagnosis and severity?
David: What is the status of the SNOMED ontology work?
Tony: The work out of ITSIDU is sanctioned. You can run an export and create RDF from it.
Eric: Kent's script started getting distributed with SNOMED.
David: I heard Tony mention the problem of "custom languages", and I agree!
<ericP> C-CDA in RDF
Eric: There's a representation of
the MIF in OWL that Lloyd McKenzie did. Example of C-CDA in
RDF.
... That example takes CCDA doc and serializes it according to
the ORIM.
eric, we need a way of capturing and linking in screenshots of interactive demos and such, during presentations. do know of any easy way to do that? I can take screenshots easily enough, but would need an easy place to upload them. Then a link would need to be generated and pasted into IRC.
I think i could automate the upload process easily enough, but still would need a place to upload them, and then a way to get the generated URL into IRC. Not sure of an easy way to automate that part.
I've never done IRC hacking.
This is scribe.perl Revision: 1.138 of Date: 2013-04-25 13:59:11 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/ by Tony Malia/s/ Succeeded: s/phenotype side/genotype side/ No ScribeNick specified. Guessing ScribeNick: dbooth Inferring Scribes: dbooth Default Present: ericP, +1.301.398.aaaa, +1.469.226.aabb, jay, Claude, Neda, DBooth, +1.978.794.aacc, Tony Present: ericP +1.301.398.aaaa +1.469.226.aabb jay Claude Neda DBooth +1.978.794.aacc Tony Got date from IRC log name: 01 Apr 2014 Guessing minutes URL: http://www.w3.org/2014/04/01-hcls-minutes.html People with action items:[End of scribe.perl diagnostic output]