Stream: implementers
Topic: Devdays breakout: packages and versioning
Alexander Henket (Nov 14 2018 at 13:33):
This is a call for participants for a possible breakout session on packages (and versioning). Problem space: FHIR offers a way package profiles and other conformance resources and have packages depend on other packages. This could solve a number of versioning issues.
Packages are not(?) described as part of R4 and tooling support is just starting up. Implementations will need to work with a mix of packaged and non-packaged 'stuff' for some time to come. What's the recommendation for modelers and/or implementers to deal with that?
Are there any fallback scenarios? Support for the pipe syntax for pointing from profile version to profile version is not supported very broadly either.
A day/time has not been set yet, but leaning towards Thursday afternoon.
@Simone Heckmann, @Grahame Grieve , @Lloyd McKenzie, @Michel Rutten
Patrick Werner (Nov 14 2018 at 13:35):
will join
Grahame Grieve (Nov 14 2018 at 14:02):
I'm in, of course. I have nothing particular yet tomorrow afternoon
Lloyd McKenzie (Nov 14 2018 at 14:03):
I've got a session at 16:05. Clear otherwise
Patrick Werner (Nov 14 2018 at 14:11):
@Simone Heckmann prefers friday afternoon, also ok for you?
Lloyd McKenzie (Nov 14 2018 at 14:30):
Ok for me
Alexander Henket (Nov 14 2018 at 17:34):
Hi all. A bit of a back'n'forth but I've just made reservations for Thu 11:05 - 11:45 at Chalet 3
Patrick Werner (Sep 12 2019 at 19:46):
Just found this thread and asked myself: did we ever come up for a resolution to track the origin package of a StructureDef on non-package aware servers?
Grahame Grieve (Sep 12 2019 at 19:54):
meta.source?
Grahame Grieve (Sep 12 2019 at 19:55):
or did I say this last time?
Patrick Werner (Sep 12 2019 at 20:08):
can't remember - that's why i'm asking. meta.source sounds great, so for the CG IG i would set meta.source to: hl7.fhir.uv.genomics-reporting|0.3to keep the version of the IG?
Grahame Grieve (Sep 12 2019 at 20:12):
it's a uri... so either that canonical or we invent an npm pacakge scheme
Patrick Werner (Sep 13 2019 at 17:42):
just saw it on your slide. Your format for a package was: hl7.fhir.uv.genomics-reporting#0.3
Which leads to three possible formats:
- hl7.fhir.uv.genomics-reporting|0.3
- hl7.fhir.uv.genomics-reporting#0.3
- the npm format would be hl7.fhir.uv.genomics-reporting@0.3
Last updated: Apr 12 2022 at 19:14 UTC
