FHIR Chat · Canonical URL of search params in FHIR Core · conformance

Stream: conformance

Topic: Canonical URL of search params in FHIR Core


view this post on Zulip Vassil Peytchev (Dec 03 2021 at 16:50):

Are the canonical URLs of the search parameter defined in FHIR core supposed to resolve? E.g. http://hl7.org/fhir/SearchParameter/Patient-identifier

view this post on Zulip Eric Haas (Dec 06 2021 at 16:40):

no they do not and never have.

view this post on Zulip Lloyd McKenzie (Dec 06 2021 at 16:50):

That said, it'd be nice if they did. Would require some publisher magic, particularly for the 'shared' ones as they have no dedicated page in the publication.

view this post on Zulip Eric Haas (Dec 06 2021 at 17:14):

(deleted)

view this post on Zulip Grahame Grieve (Dec 07 2021 at 05:14):

it sure would be nice, but I'm not actually sure where they'd redirect to?

view this post on Zulip Lloyd McKenzie (Dec 07 2021 at 05:16):

For the shared ones, we'd need to have a page. (That might be a good thing to have anyhow.) For the resource-specific ones, probably sufficient to just jump to the "search parameters" section of the resource.

view this post on Zulip Eric Haas (Dec 07 2021 at 16:07):

Lloyd McKenzie said:

For the shared ones, we'd need to have a page. (That might be a good thing to have anyhow.) For the resource-specific ones, probably sufficient to just jump to the "search parameters" section of the resource.

I think that if we do this ( and I am not convinced it is worth the effort) then it should reference a full blown page like we did for US Core. e.g., : http://build.fhir.org/ig/HL7/US-Core/SearchParameter-us-core-careteam-role.html Not saying the narrative here could not be improved upont - should get all the views and add a table header like we do for valuesets, etc and an optional blurb and example syntax for completeness.

view this post on Zulip Eric Haas (Dec 07 2021 at 16:15):

So far the only folks I know who dove into this that deeply are IG authors and they can figure it out by looking at the package,. but if there is a wider audience then I can mock up a solid proposal

view this post on Zulip Lloyd McKenzie (Dec 07 2021 at 17:55):

Another possibility is to just have the canonical resolve to one of the technical representations (probably JSON) if no content type is specified.

view this post on Zulip Grahame Grieve (Dec 07 2021 at 18:00):

I can't retrospectively add pages like this to R4

view this post on Zulip Lloyd McKenzie (Dec 07 2021 at 20:15):

I'm not super-fussed about R4. I was thinking more about what might be feasible for R5+

view this post on Zulip Eric Haas (Dec 08 2021 at 23:56):

Lloyd McKenzie said:

Another possibility is to just have the canonical resolve to one of the technical representations (probably JSON) if no content type is specified.

then I would argue just inspect the package

view this post on Zulip Lloyd McKenzie (Dec 09 2021 at 03:34):

When a developer plops the link into a browser, they want something useful. Saying "go inspect the package" isn't useful.

view this post on Zulip Eric Haas (Dec 09 2021 at 08:19):

I don't think your proposal adds enough value to be worth the effort. Where else do we link to the json format?


Last updated: Apr 12 2022 at 19:14 UTC