FHIR Chat · R4B publication location

Stream: fmg

Topic: R4B publication location

Josh Mandel (Jan 26 2022 at 22:48):

I'm assuming R4B will replace the content at http://hl7.org/fhir as "current version of FHIR"?

Lloyd McKenzie (Jan 27 2022 at 00:15):

That is also my presumption

John Moehrke (Jan 27 2022 at 13:48):

so what is the plan for all of the IGs that are today based on R4. Will they be compelled to revise to R4B? If not, what does that mean?

Lloyd McKenzie (Jan 27 2022 at 18:19):

If/when you update your IG, you can choose whether you move or not. I don't think we'd force it the way we did with 4.0.1, but I'd expect the FMG to ask questions if you didn't move to 4.3.0

John Moehrke (Jan 27 2022 at 18:24):

so, no one is going to explain "what are the factors that I need to know about?"

Lloyd McKenzie (Jan 27 2022 at 19:00):

I'm not sure what factors you're looking for. No breaking changes (and generally no changes at all) to anything other than EBM, Med governance and subscription stuff. (And subscription stuff and EBM are net new, not changes per se.) The "breaking" stuff is things like "enumerations of possible resources" will now allow additional resources, Bundles now allow new resources, etc.

John Moehrke (Jan 27 2022 at 19:01):

I thought subscription stuff was in an IG, not being pulled into R4B. Why is the IG not sufficient?

John Moehrke (Jan 27 2022 at 19:02):

but if these are the three things.. then I think it should be clear when and why an IG would need to care about R4 vs R4B... and thus they should not be based on R4B unless they need one of these three things. right? Thus, if anything the IG population should be discouraged from building off of R4B.

Gino Canessa (Jan 27 2022 at 20:02):

Re: subscriptions, there are two new resources: SubscriptionTopic and SubscriptionStatus. Modeling them in R4 was a lot of Parameters/Basic/Extension. Given that anyone implementing topic-based subscriptions is at least doing a software update (to actually implement them), adding the new resources seemed like a lower bar. Especially given the context of a not-too-distant R5 publication.

Gino Canessa (Jan 27 2022 at 20:07):

On John's other note: R4B also contains technical corrections on top of R4:4.0.1, so I assume we want to direct people to 4.3.0 over 4.0.1 as a general rule.

I spoke about it offline with someone, but assume there is a need for further technical corrections in R4 about a year out. Applying them to 4.0.1 means also trying to backfill changes from 4.3.0... I think that makes publication of 4.3.1 significantly more likely than 4.0.2. I also assume trying to publish both a 4.0.2 and 4.3.1 release would cause a lot of people to resign their positions ;-)

Lloyd McKenzie (Jan 27 2022 at 20:23):

They should absolutely be based on R4B whenever possible because R4B includes a lot of technical corrections. They should only use 4.0.1 when absolutely necessary (and I can't think of a situation where that would occur...)