Joao: Goal of shex?
david: Validation.
Joao: Difference from SHACL?
david: Mostly stylistic. SHACL is SPARQL oriented; ShEx is much more concise, which is nice for humans.
Joao: Use of BFO in FHIR?
tony: None. FHIR has its own
structure definitions.
... Instances need to be round-trippable, so they need to be
isomorphic with JSON and XML representations of FHIR.
... BFO is sort of above the FHR ont, more like the HL7
RIM.
... Should be possible to map FHIR into BFO.
Joao: Looking at United
Foundational Ontology (UFO)
... Been working with UFO and Dolce. Nice thing to notice in
the W3C SSN ont is that it uses Dolce to align the main ont w
high-level ont. REcommend looking at this approach. Maybe could
do this with FHIR also.
tony: Would require re-thinking
how FHIR is designed. I was bringing them into Protege, and you
cannot do anything unless you have the TBox part in
Protege.
... So I built a TBox ont based on the FHIR structure..
... Once i got the ont into Protege it came alive.
... Found a better tool, with structured data capture, but not
so pure. Takes some of the OWL stuff and makes it
scalable.
... No reasoner, but SPARQL queries.
... User wants drop-down list of SNOMED terms, and I've done
that.
david: That gap between FHIR's records-oriented view versus the fact-oriented ontologies like BFO will require mapping, and we'd like to do that.
tony: We exchanged object graphs a while ago, but FHIR is locked into its current structure.
(Eric joins, Tony drops off)
david: Eric, can you say more about how you and Harold are using shex?
eric: different validation
definitions available, and shex is one of them. Look below
"Resource content" table on a given FHIR resource page, such as
http://build.fhir.org/observation.html
... One of the links will be to the shex def.
... FHIR pub process involves validating all ofthe
examples.
david: A de facto test
suite.
... Here is shex for fhir:Observation: http://build.fhir.org/observation.shex.html
(Eric does demo of shex Constellations doc: https://hangouts.google.com/_/elUi/chat-redirect?dest=https%3A%2F%2Fwww.w3.org%2F2016%2FFHIR-tutorial%2FConstellations)
(Rob Hausam joins)
(Eric does demo of ShExMap playground: https://rawgit.com/shexSpec/shex.js/push-down/extensions/shex-map/doc/shexmap-simple.html)
eric: looking at Blood Pressure
example
http://build.fhir.org/observation-example-bloodpressure.ttl.html
... When the instance data is validated, the variables in the
shex %Map statements become bound to the instance data, and you
can output it in the form of a different shex schema.
... This allows you to translate from one shex schema to
another.
Joao: Similar to xslt?
eric: partly. with xslt it is hard to see the target schema. With shexmap you have two, clear schemas: source and target.
david: and they're bidirectional! either shex can be the source or target.
eric: And this test suite shows both directions.
Joao: Similar to RML?
eric: RML is mostly from RML to RDF. One of the RML architects is looking into this.
david: We experimented with ShExMap for mapping from one RDF data model to another. Essentially used it like a rules language.
Joao: Looked at complex event use
case?
... I can show it on March 27
... Also looking at lightweight FHIR for IoT.
... Intend to bring this to the project, and they want to put
itinto their core ont.
... Can talk about it later.
david: IoT was also discussed at
the US Semantic Technology Symposium last week http://www.us2ts.org/
... I will see if I can find their slides and point you to
them, in case they are of interest.
Joao: I expect to be able to join next with, with more questions :)
ADJOURNED
This is scribe.perl Revision: 1.152 of Date: 2017/02/06 11:04:15 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00) Succeeded: s/BFO/BFO in FHIR/ Succeeded: s/playground/Constellations doc/ Succeeded: s/shex play/ShExMap play/ Present: David_Booth EricP Rob_Hausam Tony_Mallia Joao_Moreira No ScribeNick specified. Guessing ScribeNick: dbooth Inferring Scribes: dbooth Found Date: 06 Mar 2018 People with action items: WARNING: IRC log location not specified! (You can ignore this warning if you do not want the generated minutes to contain a link to the original IRC log.)[End of scribe.perl diagnostic output]