26 Apr 2012

See also: IRC log




<BobF> I keep getting disconnected on the voice line - anyone else having trouble?

<matthias_samwald> i made it after the 2nd attempt.

<matthias_samwald> scott, could you scribe?

<matthias_samwald> OSEHRA

OSEHRA - Open Source Electr. Health Rec. Agent

<matthias_samwald> http://www.osehra.org/

<JoanneLuciano> OSEHRA.org

<matthias_samwald> http://www.osehra.org/blog/some-background-material-semantic-web-topics-architecture-conference-call-feb-28

<ericP> JoanneLuciano: OSEHRA is the VA's standard EHR

<ericP> ... could W3C produce a standard EHR?

<ericP> ... after we propose one, could we make it a W3C Recommendation?

<ericP> matthias_samwald: osehra appears to be discussing using RDF and SPARQL

<ericP> mscottm2: we took inventory of this a couple years ago:

<ericP> ... .. OpenEHF

<ericP> ... .. OpenEHR

<boycer> was it OpenVista?

<ericP> ... .. OpenVista

<matthias_samwald> Matthias: the one that I looked into was "Tolven"


<ericP> ... .. Tolven, from collabrx

<ericP> ... in terms of making an RDF EHR, it depends on the information model you're using

<ericP> ... e.g. support all of OpenVista or HL7 V3

<boycer> maybe OpenMRS <openmrs.org>

<ericP> ... you could make RDF versions of the given systems

<ericP> JoanneLuciano: if we provide mappings between them...


(above was just something I dug up related to the conversation)

Joanne: by mapping to main EHR players and building applications on top of that


<ericP> mscottm2: fits our goals, but my tactic is to work with someone who wants to do this, e.g. a clinic participating in EUREKA (connecting clinical records and radiology,

<ericP> ... otherwise we have to create all of the data model

<ericP> ... before there's value in mapping EHRs to each other, we have to show value in working in RDF

<ericP> matthias_samwald: before we build standards, we need to connect with the people working on e.g. FHIR to see where we can help them

<matthias_samwald> ericP: has weekly telecons with FHIR participant.

<matthias_samwald> ericP: FHIR developer enthusiastic about RDF for agent development.

ericP: Cecil Lynch tells people that you can do agile development with a triplestore and ontologies
... "It's a good way for my company to facilitate collaborative work."

HL7 folks created a UML model called RIM some years ago (10?) but it didn't really take off. The notion was that everyone would use RIM to encode things and get them to connect. But adoption was limited.

<matthias_samwald> reminds me of Barry's HL7 criticism on http://hl7-watch.blogspot.com/

scribe: So, they've created a new project called Fire that uses XML and I (ericP) created RDF versions of it.

ericP: I'm not sure that they realize that committing to RDF means a certain amount of work to settle on the definitions.

Working with Lloyd MacKenzie.

ericP: There's a question of how much good modeling we require during the development vs. having someone else come along and figure it out post facto .
... I can bring more of this to the task force if you like.

<JoanneLuciano> Scott: Dan Brickley has been interested in geneology and clinical will need to represent that info from Morman's in an interoperable way - no way now to identify a patient mutation in EHR

<JoanneLuciano> Scott: is there synergy there with FHIR

<JoanneLuciano> Back to Agenda!

Mormon geneology format is the de facto standard at the moment: GEDCOM

<JoanneLuciano> MS: looked into how to use OWL for creating recommendations for clinical pharmacogenomics based on extensions of TMO

<JoanneLuciano> MS: Many of there rules can be expressed by simple SPARQL queries

<matthias_samwald> http://purl.org/zen/ssr.ttl

<JoanneLuciano> Eric: How close is this to SPIN?

<JoanneLuciano> MS: SPIN has many features - expected to check novel or stop iterating or magic properties, etc...

<JoanneLuciano> SPIN - simple sparql - rules, and some alternative to OWL reasoning based on SPARQL (object oriented) - but more complex than what Matthias was looking for.

Scott asked: What the difference between chaining SPARQL Constructs (as in SWObjects) and SPIN, and diff with Matthias's simple sparql rules?

EricP: SWObjects uses query rewriting and will not chain deeply. SPIN will chain to a predetermined depth.

Matthias: With each iteration, SPARQL rules add triples from consequent to the triplestore.

Scott: Does SWObjects query rewriting chain then?

EricP: Yes, between SPARQL endpoints but not within a single map file (of SPARQL Constructs).

<matthias_samwald> Decision support table (including Simple SPARQL rules on the far right): https://docs.google.com/spreadsheet/pub?hl=en_US&hl=en_US&key=0AiGT-vnkGcoLdFFVMEdqcFdYaDFqS0xHTnlUT0N3cEE&single=true&gid=3&output=html

Nice! You can click on each rule to view the definition.


<boycer> Matthias, great work! I have to go but will send you some thoughts by email.

<ericP> G1:{ :Laborador ⊆ :Doggie . :Doggie ⊆ :Mammal . :Mammal ⊆ Animal } | { ?a ⊆ ?b . ?b ⊆ ?c } ∴ { ?a ⊆ ?c } = G1 + { :Laborador ⊆ :Mammal . :Doggie ⊆ Animal }

<matthias_samwald> https://docs.google.com/spreadsheet/ccc?key=0AiGT-vnkGcoLdGlPS3lSR25OWXNLUWl6SWp3LUJLN2c#gid=0

<ericP> G1:{ :Laborador ⊆ :Doggie . :Doggie ⊆ :Mammal . :Mammal ⊆ Animal } | { ?a ⊆ ?b . ?b ⊆ ?c } ∴ { ?a ⊆ ?c } | { ?a ⊆ ?b . ?b ⊆ ?c } ∴ { ?a ⊆ ?c } = G1 + { :Laborador ⊆ :Mammal . :Doggie ⊆ Animal . :Laborador ⊆ :Animal }

<matthias_samwald> https://docs.google.com/spreadsheet/ccc?key=0AiGT-vnkGcoLdGlPS3lSR25OWXNLUWl6SWp3LUJLN2c#gid=7

<ericP> brb

<ericP> b

Bob: It's sort of an open vs. closed world problem. If you define a wild type based on not having a particular set of SNPs but, later, a functionally relevant SNP is discovered, the wild type definition might not hold anymore.

Michel - you just dropped out for me before "the variant cases"

<michel> i was just saying that assigning a role for the wild type basically falls to a default reasoning case - non-monotonic reasoning

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.136 (CVS log)
$Date: 2012/04/26 16:11:12 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.136  of Date: 2011/05/12 12:01:43  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

No ScribeNick specified.  Guessing ScribeNick: mscottm2
Inferring Scribes: mscottm2

WARNING: No "Topic:" lines found.

WARNING: No "Present: ... " found!
Possibly Present: Bob BobF Eric G1 Helena Joanne JoanneLuciano MS Matthias P49 Scott aacc aadd boycer egonw_ ericP harry_ https iker matthias_samwald michel mscottm2
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy

WARNING: No meeting title found!
You should specify the meeting title like this:
<dbooth> Meeting: Weekly Baking Club Meeting

WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Got date from IRC log name: 26 Apr 2012
Guessing minutes URL: http://www.w3.org/2012/04/26-HCLS-minutes.html
People with action items: 

WARNING: No "Topic: ..." lines found!  
Resulting HTML may have an empty (invalid) <ol>...</ol>.

Explanation: "Topic: ..." lines are used to indicate the start of 
new discussion topics or agenda items, such as:
<dbooth> Topic: Review of Amy's report

[End of scribe.perl diagnostic output]