FHIR Chat · Backporting elements to DSTU2 · implementers

Stream: implementers

Topic: Backporting elements to DSTU2


view this post on Zulip Vadim Peretokin (Jan 26 2017 at 15:14):

Has anyone had to backport elements introduced in STU3 for use in DSTU2? I'm thinking of just copying the element and modelling it as an extension. Interested to hear if anyone has done the same or differently

view this post on Zulip nicola (RIO/SS) (Jan 26 2017 at 15:17):

Wouldn't be nice if core elements and extensions were uniform? :)

view this post on Zulip Vadim Peretokin (Jan 26 2017 at 15:20):

How would it work with data processing then? Whole idea of extensions is that you don't encounter unknown elements. Interested to hear about this, but in another stream perhaps

view this post on Zulip Lloyd McKenzie (Jan 27 2017 at 23:05):

In theory we could auto-generate extensions for all new elements in STU 3. We already have transforms that convert data between STU 2 and 3, so this would allow such transforms to be more complete. Thoughts @Grahame Grieve ?

view this post on Zulip Brian Postlethwaite (Feb 02 2017 at 23:57):

I like the idea also.

view this post on Zulip Brian Postlethwaite (Feb 02 2017 at 23:58):

would make this much easier for the HAPI server and others exposing the same content over both interfaces, not just a translation

view this post on Zulip Vadim Peretokin (Feb 03 2017 at 07:32):

+1

view this post on Zulip Grahame Grieve (Feb 06 2017 at 04:22):

well, we wouldn't want to generate extensions for R3 things that actually exist in R2. in probably 95% of cases, it's binary: the element does or doesn't exist in R2. but in 5%, it's not at all obvious. People can read the R2/r3 conversion scripts if you want to be educated on this.

view this post on Zulip Lloyd McKenzie (Feb 06 2017 at 04:43):

Have we created extensions for the things that are obviously new in R3?

view this post on Zulip Brian Postlethwaite (Feb 06 2017 at 04:53):

It could be based off the r2map tab...

view this post on Zulip Grahame Grieve (Feb 06 2017 at 06:26):

it's only obvious to a human, not to a computer

view this post on Zulip Vadim Peretokin (Feb 06 2017 at 08:29):

Just having the obvious stuff would be a boon, and I don't see a problem with generating extensions for the non-obvious cases anyway: I'd definitely look in the resource first to see if I can map an element before going on a trek to find the right extension


Last updated: Apr 12 2022 at 19:14 UTC