See also: IRC log
<trackbot> Date: 05 May 2015
<dbooth> ACTION: [DONE] dbooth to suggest to the FIHR group to make it explicit in the FHIR spec tha a profile URI *should* be dereferenceable to an authoritative definition of that profile [recorded in http://www.w3.org/2015/04/07-hcls-minutes.html#action01]
<dbooth> Apr 28 minutes: http://wiki.hl7.org/index.php?title=ITS_RDF_Concall_Minutes_20150428
<dbooth> Dale is co-chair today
<scribe> scribenick: ericP
<dbooth> David moved to approve minutes; eric seconded.
<dbooth> RESOLVED: Apr 28 minutes approved!
<dbooth> FHIR RDF and Validation/Translation Task Force
<dbooth> http://www.w3.org/2015/04/29-hcls-minutes.html
dbooth: we went through the
issues list and close issues that we'd already handled (and one
duplicate)
... we mostly talked about bnodes [in RDF representations of
FHIR/XML]
<dbooth> Options: 1. use blank nodes and let applications deal with the problem by minting skolem URIs or something else; 2. define a standard way to form a URI; 3. use a skolem URI using the well-known convention defined by RDF.
dbooth: the easiest way to represent objects nested in FHIR/XML Resoruces is to create a fresh bnode
<dbooth> eric: Concerned that creating permanent URIs for bnodes may lead to an erroneous conclusion that if the same URI is generated in two different documents, it will be treated as the same entityu.
<dbooth> ... For example, if the are named numerically based on position first item, next item, etc., if the items are not the same in those two Observations then they'll still have the same URI, though they shouldn't.
<Rafael_Richards> Is the original cause or need for blank node poor definition? I.e. use of "NOS"?
dbooth: let's see what we want to
do at Paris F2F
... we can go through issues
... we can go through Tony's side-by-side
... i.e. continue doing this sort of stuff
claude: we could do some
outreach
... engaging folks. invite folks to our meetings
Rafael_Richards: we have european
counterparts at e.g. University of Geneva
... (interested in Linked Data)
... do we have a list of attendees?
Lloyd: not yet. occasionally distributed in advance. could ask lillian but she's busy and traveling tomorrow
claude: maybe we can do this between sessions
<dbooth> ACTION: Claude to get list of HL7 meeting attendees to see who may want to join our meeting [recorded in http://www.w3.org/2015/05/05-hcls-minutes.html#action02]
<trackbot> Error finding 'Claude'. You can review and register nicknames at <http://www.w3.org/2014/HCLS/track/users>.
claude: during the session we should tell people how to reach us
<dbooth> eric: Only within HL7 people? Or others?
claude: let's start with HL7
folks but they can reach out to others
... give them some contact info
<dbooth> claude: Start w HL7 people, and people they know.
<dbooth> lloyd: Non-HL7 attendees would have to register and pay non-member rates.
<dbooth> eric: Would be great to get Alejandra
dbooth, it would be good to let people know, even if they probably can't come
dbooth: it would be good to let
people know, even if they probably can't come
... so contact anyone you think of
<Rafael_Richards> Program for Paris HL7 meeting: http://www.hl7.org/documentcenter/public_temp_6E8EF6E0-1C23-BA17-0CED013A099E3264/calendarofevents/Paris-onsite_v9.pdf
dbooth: can you own the agenda
for our session?
... we had overview materials last time
<dbooth> ACTION: Claude to plan Paris agenda [recorded in http://www.w3.org/2015/05/05-hcls-minutes.html#action03]
<trackbot> Error finding 'Claude'. You can review and register nicknames at <http://www.w3.org/2014/HCLS/track/users>.
dbooth: any more for Paris right now?
Rafael_Richards: i pasted the meeting schedule. ops for claude to network e.g. FHIR session at 9am
claude: and i'll be at the connectathon
Rafael_Richards: there's a 9-5 connectathon on fri but there are FHIR talks every day
dbooth: what are appropriate ways to announce to others?
Lloyd: propose ITS list (which is
where folks who care about syntax hang out) and FHIR list
... can also mention it on the implementors skype
chat
<dbooth> ACTION: Claude to send invite to FHIR list about the RDF subgroup meeting in Paris [recorded in http://www.w3.org/2015/05/05-hcls-minutes.html#action04]
<trackbot> Error finding 'Claude'. You can review and register nicknames at <http://www.w3.org/2014/HCLS/track/users>.
<dbooth> eric: I won't be in Paris, but I will spend some time w Josh Mandel.
<dbooth> ... So I might be able to get him to answer tough questions about how to get stuff from the FHIR spec to our ont.
<dbooth> Older version of Tony's comparison: http://wiki.hl7.org/images/2/25/FHIR_RDF_Sample_side_by_side_comparisons.pdf
<dbooth> :resource a fhir:Observation;
<dbooth> fhir:Observation.code [
<dbooth> a loinc:54411-4;
<dbooth> fhir:CodeableConcept.coding [
<dbooth> fhir:Coding.system [ fhir:uri.value <http://loinc.org>] ;
<dbooth> fhir:Coding.code [ fhir:code.value "54411-4" ];
<dbooth> fhir:Coding.display [ fhir:string.value "Rh immune globulin given Qualitative"];
<dbooth> ];
<dbooth> fhir:CodeableConcept.text [ fhir:string.value "Rh immune globulin" ];
<dbooth> ].
<dbooth> eric: Need to encode fhir:CodeableConcept.coding
<dbooth> instance as a URI instead of blank node, for OWL reasoners.
<dbooth> ... I verified the problem with an OWL reasoner, though it is not a problem in SPARQL.
<bpech> I apologize, I have to leave early.
<dbooth> ACTION: David to check with an OWL reasoner expert to see if the blank node issue is inherent or a tool limitation [recorded in http://www.w3.org/2015/05/05-hcls-minutes.html#action05]
<trackbot> 'David' is an ambiguous username. Please try a different identifier, such as family name or username (e.g., dbooth, dderour, dhansen2, dnewman, dshotton, dwild).
-> http://piratepad.net/hcls hcls pirate pad
<dbooth> [[
<dbooth> @prefix : <http://BlankNodeTest/> .
<dbooth> @prefix owl: <http://www.w3.org/2002/07/owl#> .
<dbooth> @prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
<dbooth> @prefix xml: <http://www.w3.org/XML/1998/namespace> .
<dbooth> @prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
<dbooth> @prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
<dbooth> @base <http://BlankNodeTest> .
<dbooth> <http://BlankNodeTest> a owl:Ontology .
<dbooth> :hasChild a owl:AsymmetricProperty , owl:ObjectProperty .
<dbooth> :hasHairColor a owl:ObjectProperty ;
<dbooth> rdfs:range :HairColor .
<dbooth> # :hasSSN a owl:ObjectProperty .
<dbooth> :hasSSN a owl:DatatypeProperty .
<dbooth> :value a owl:DatatypeProperty .
<dbooth> :code a owl:Class .
<dbooth> :HairColor a owl:Class ; rdfs:subClassOf :code .
<dbooth> :BrownHaired a owl:Class ;
<dbooth> owl:equivalentClass [
<dbooth> a owl:Class ;
<dbooth> owl:intersectionOf (
<dbooth> :HairColor
<dbooth> [ owl:onProperty :value ; owl:hasValue "brown" ]
<dbooth> )
<dbooth> ] ;
<dbooth> rdfs:subClassOf :HairColor .
<dbooth> :brown1 a :HairColor , owl:NamedIndividual ;
<dbooth> :value "brown" .
<dbooth> :Person a owl:Class .
<dbooth> :frank a :Person , owl:NamedIndividual ;
<dbooth> :hasChild
<dbooth> [
<dbooth> a :Person ;
<dbooth> :hasSSN "123-45-6789" ;
<dbooth> :hasHairColor [
<dbooth> a :HairColor ;
<dbooth> :value "brown"
<dbooth> ]
<dbooth> ], :frankson .
<dbooth> :frankson
<dbooth> a :Person ;
<dbooth> :hasSSN "123-45-6790" ;
<dbooth> :hasHairColor [
<dbooth> a :HairColor ;
<dbooth> :value "brown"
<dbooth> ]
<dbooth> .
<dbooth> ]]
<dbooth> Postpone this issue until talking to OWL experts
dbooth: four possibilities:
<dbooth> Some possibilities:
<dbooth> rdf:List
<dbooth> Ordered List Ontology
<dbooth> Collections Ontology
<dbooth> Simple List Conventions
<dbooth> http://www.w3.org/TR/rdf-schema/#ch_list
dbooth: RDF default -- use the
existing RDF list expression
... this causes trouble for SPARQL queries
<dbooth> (a b c d)
dbooth: it's hard to get those elements in order in a SPARQL query
<dbooth> http://smiy.sourceforge.net/olo/spec/orderedlistontology.html
<dbooth> http://www.essepuntato.it/lode/owlapi/http://purl.org/co/
dbooth: so lots of people use other techniques, e.g.
<dbooth> http://goo.gl/8PNuAG
dbooth: Simple List Convention is similar to OLO but more concise
<dbooth> lloyd: is inference the same for these options?
ericP: SLC is more terse in SPARQL and Turtle because it leverages rdf:List syntaxes
<dbooth> ... Sometimes we'll care that an item is first; other times we'll care about relative ordering.
ericP: do we care about order in inferecning use cases?
Lloyd: we care about first or one
before another
... let's enumerate UCs:
... ... essential UC: regenerate XML and JSON in the same
order
... .. in some circumstances, for some profiles, you care about
the 1st member
... .. relative position: is diagnosis 1 more important than
diagnosis 2?
... .. is this result newer or older than some other
result?
<dbooth> dbooth: UC 1 is covered by all of these optoins.
dbooth: do you mean testing by SPARQL or a reasoner?
Lloyd: if we can do it in a
reasoner, that's ideal
... if we can't but we can do it in SPARQL, that will probably
suffice
... if two can do it in SPARQL and one can do it in the
reasoner, that's a point for the one that the reasoner
handles
... in general, reasoners suck at order or arithmetic
dbooth: UC to involves looking
for a certain number
... UC2 involves looking for a certain number
... UC3 involves comparing values
<dbooth> eric: reasoners do not compare numeric values. SWRL does though.
ericP: there will be one extra predicate to traverse in SLC
<dbooth> eric: SLC would be more applicable when using exclusively Turtle or SPARQL
<dbooth> My pref: OLO or SLC
dbooth: i propose we skip
Collections Ontology as it is complex
... and not rdf:List [because it's hard to work with in
SPARQL]
... my pref it to put it on the list to solicit
preferences
<dbooth> ACTION: Ask on the list for preferences about which list convention to use [recorded in http://www.w3.org/2015/05/05-hcls-minutes.html#action06]
<trackbot> Error finding 'Ask'. You can review and register nicknames at <http://www.w3.org/2014/HCLS/track/users>.
<scribe> ACTION: dbooth to ssk on the list for preferences about which list convention to use [recorded in http://www.w3.org/2015/05/05-hcls-minutes.html#action07]
<trackbot> Created ACTION-23 - Ssk on the list for preferences about which list convention to use [on David Booth - due 2015-05-12].
<scribe> ACTION: ericP to arrange an HCLS bridge for 09:30-11:00 EDT [recorded in http://www.w3.org/2015/05/05-hcls-minutes.html#action08]
<trackbot> Created ACTION-24 - Arrange an hcls bridge for 09:30-11:00 edt [on Eric Prud'hommeaux - due 2015-05-12].
estimate 10 people
<Rafael_Richards> eric: have you looked at the OMG XML model standard? Round trip this to RDF?
<Rafael_Richards> http://www.omg.org/mof/
<dbooth> dbooth: we'll see on the list who will be available for our regularly scheduled calls next week, which conflict w HL7 paris mtg.
<dbooth> AJOURNED
-> https://www.w3.org/wiki/File:Snomedct_extract.ttl SNOMED CT Extract for C-reactive protein
-> https://www.w3.org/wiki/File:Snomedct_extract_descriptive.ttl SNOMED CT Extract for C-reactive protein (with descriptive URLs)
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/based on/based on position/ Succeeded: s/cloudcell_/claude/ Succeeded: s/thought/though/ Found ScribeNick: ericP Inferring Scribes: ericP Default Present: dbooth, ericP, claude, Kerstin_Forsberg, lloyd, +1.916.783.aaaa, DNelson, Rafael_Richards, bpech, +1.512.680.aabb, Darrell Present: dbooth ericP claude Kerstin_Forsberg lloyd DNelson Rafael_Richards bpech Darrell Found Date: 05 May 2015 Guessing minutes URL: http://www.w3.org/2015/05/05-hcls-minutes.html People with action items: ask claude david dbooth ericp[End of scribe.perl diagnostic output]