FHIR Chat · Content-Type for Turtle · implementers

Stream: implementers

Topic: Content-Type for Turtle


view this post on Zulip Grahame Grieve (Feb 02 2018 at 09:37):

we are using text/turtle for the content type for the turtle format. Several users have asked me if we can use application/fhir+turtle instead.

view this post on Zulip Grahame Grieve (Feb 02 2018 at 09:37):

that would require that we register +turtle here: https://www.iana.org/assignments/media-type-structured-suffix/media-type-structured-suffix.xhtml. @Eric Prud'hommeaux is that a sensible thing to do?

view this post on Zulip Grahame Grieve (Feb 02 2018 at 09:40):

@Patrik Sundberg that raises the same question for protobuf... should that be application/fhir+protobuf, or would it be better to be application/protobuf; message=org.hl7.fhir.Patient; fhir-version=3.0.... looks like +protobuf would be better, but see https://stackoverflow.com/questions/30505408/what-is-the-correct-protobuf-content-type

view this post on Zulip Grahame Grieve (Feb 02 2018 at 09:46):

anyway, do people have a preference about text/turtle vs application/fhir+turtle; fhir-version=3.0 ?

view this post on Zulip Kevin Olbrich (Feb 02 2018 at 17:18):

I like the text/fhir+turtle; fhirVersion=3.0approach for consistency with json and xml. It might be worthwhile to go ahead and preemptively request a few other mime types, like application/fhir+msgpack (or whatever the right one for that format is).

view this post on Zulip Grahame Grieve (Feb 02 2018 at 17:52):

we can't add fhir-version to text/turtle - at least not officially

view this post on Zulip Patrik Sundberg (Feb 03 2018 at 01:08):

Typically when people send protobuf over the wire, they use gRPC, mainly to enforce that the client and server agree on the particular protobuf message schema. That said, especially for bulk data export, I agree that it would be nice to be able to return binary content using the standard REST api. I would prefer to be consistent with the content types for json and xml, and we will need the message type to be specified in the header. Something like "application/fhir+protobuf; message=org.hl7.fhir.stu3.Patient; fhir-version=3.0" would work. We could possibly drop the fhir-version since it'll be implied by the message type, or keep it for clarity and consistency.

view this post on Zulip Grahame Grieve (Feb 03 2018 at 02:33):

So is google interested in registering “+protobuf”?

view this post on Zulip Grahame Grieve (Feb 03 2018 at 02:33):

Because I agree completely with what you wrote

view this post on Zulip Grahame Grieve (Mar 04 2018 at 01:16):

Bringing this back to life. @David Booth @Harold Solbrig @Patrik Sundberg @Lloyd McKenzie @Ewout Kramer @Josh Mandel @Paul Knapp @Dale Nelson @James Agnew @Eric Prud'hommeaux do we want to action this for R3?

That is, I propose that we document:
- application/fhir+turtle for turtle instead of text/turtle
- application/fhir+protobuf for protobuf

Note that both imply that someone (!) will register +turtle and +protobuf with iana... and I bags it not be me... I'm still going through the procedural process on application/fhir+json... hoops of no substance to pass and now just stuck going nowhere

view this post on Zulip Patrik Sundberg (Mar 04 2018 at 03:43):

I will inquire about what happened to the +protobuf registration, and see if we can revive it.

view this post on Zulip Grahame Grieve (Mar 04 2018 at 04:49):

thx

view this post on Zulip Grahame Grieve (Mar 04 2018 at 04:49):

It doesn't have to registered for us to use , but we won't be able to register it unless it is


Last updated: Apr 12 2022 at 19:14 UTC