Confusion about the role of JSON-LD

https://chat.fhir.org/#narrow/stream/179166-implementers/topic/RDF.20Support.20in.20hapi-fhir

eric: SOLID uses Turtle, so we'll be optimizing for that, and want to streamline that, with errors on line numbers. That's motivating the HAPI server work.

harold: Line number issue is important. Should call attention to it.
... In another project we decorated the string datatype with line numbers, to carry them along.

extensions

https://github.com/fhircat/FHIRCat/issues/31

eric: For modifying extension, user should not accidentally access the data.

AGREED: For modifying extension, user should not accidentally access the data.

eric: THe RDF approach would be to change the predicate. The FHIR way would be to change the property in a predictable way, such as leading underscore.
... Could use `observation._status`
... I kinda like that, because it feels RDFy and FHIRy.

david: FHIR already uses the underbar convention for non-mod extensions, so I'm concerned that it may create confusion with non-mod ext.

eric: We won't ever see the _status in RDF.

harold: But someone who knows FHIR may see the RDF and assume that the semantics are the same.
... Better to alter the token to prevent that confusion.
... We punted on this problem in R4, but now in RDF you can access things in pieces, so you might not notice the modifier tag. I think the FHIR folks might run into this problem soon, because they're starting to use graph query graphQL.

david: Note that FHIR/JSON already uses a "modifierExtension" property https://www.hl7.org/fhir/extensibility.html
... but they also use Is-Modifier = true .
... Let's table this discusison until we better understand how mod ext work in FHIR.

<scribe> ACTION: All read how mod ext work in FHIR

List ordering

https://github.com/fhircat/FHIRCat/issues/30

<scribe> ACTION: David to add example of option 2

<scribe> ACTION: Harold to add Turtle to list ordering option 1 example in https://github.com/fhircat/fhir_rdf_validator/blob/master/tutorial/FHIRR5.md

harold: How does schema.org represent ordered lists?

<scribe> ACTION: David to figure out what schema.org uses to represent ordered lists

david: Anything in the data cube work for ordered lists?

eric: DDI is close to data cubes. IDK.

<scribe> ACTION: David to look at data cube work to see if they use ordered lists and how

eric: OWL can say that something should not have more than one index the same.

harold: downside of option 4 is that primitive values must be inside objects
... Does FHIR allow the same member to appear twice in an ordered list?
... The default JSON-LD processing assumes that [ a, b, c ] are unique -- not [ a, b, a ].

<scribe> ACTION: David to add example to option 4 in https://github.com/fhircat/FHIRCat/issues/30

harold: example here: https://tinyurl.com/tlahmfc

david: All of the options we have so far suffer from the parallel downside: "List status [ "good", "fair", "good" ] will by default be turned into two RDF values for status, not three."

harold: The worst is to have two ways to access the same thing.

david: Could ask others for input on this question after we narrow down the options.

harold: Need to show: 1. what JSON transforms are needed; and 2. how the resulting RDF looks and will be used.

<scribe> ACTION: Harold and David to edit list and extension options to show: 1. what JSON transforms are needed; and 2. how the resulting RDF looks and will be used.

ADJOURNED

This is scribe.perl Revision: 1.154  of Date: 2018/09/25 16:35:56  
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/david: should add at least one more option, since existing options are all parallel style.//
Present: Daniel_Stone Dazhi EricP Gerhard Harold_Solbrig David_Booth hsolbrig Gopi
No ScribeNick specified.  Guessing ScribeNick: dbooth
Inferring Scribes: dbooth

WARNING: No date found!  Assuming today.  (Hint: Specify
the W3C IRC log URL, and the date will be determined from that.)
Or specify the date like this:
<dbooth> Date: 12 Sep 2002

People with action items: all david harold

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.)