Open pieter-edelman-nictiz opened 2 years ago
I think option 1 will be a good choice. We can develop the profile as far as possible. In addition, we can indeed create a separate package and then test it with the implementers. The implementers can then provide feedback on what they encounter and we can then adjust this again in the profile.
My view: don't make an FHIR implementation unless someone or some informationstandard is using the zib and asks for it. If so then make a FHIR implementation as far as reasonable.
Make a clear and findable notion about this... so people can reach out if they need an FHIR implementation for one of these zibs.
I left some remarks in the GlasgowComaScale mapping Excel on August 1 ~12:30 on some issues that at least need to be clarified before that could potentially mean the zib would be 'un-usable' as is. But what happend between my remarks and the creation of this ticket? I cannot find any zib-ticket(s) on the subject.
A few zibs turn out, when profiling, to have some fundamental issues which make them in practice unusable (on a functional level). As far as we can see, this pertains to:
We should consider what to do regarding profiles for these zibs. If these zibs are not usable on a functional level, a FHIR implementation is moot but would require major effort. Options: