hi David I think my comment may have created more concern than is warranted. RDF > does have named graphs, which we could use to delineate a certain set of > triples. But we haven't been doing that for FHIR RDF and I don't think it > is needed either. In practice, the FHIR resource that you get contains > whatever data it contains, whether it's in XML, JSON or RDF. And if it > somehow requires anything that isn't there, then it is non-conformant, > whether it is XML, JSON or RDF. If something is not there but not > required, then that just means that the sender didn't include that > information, whether it is in XML, JSON or RDF. So I think the question of > missing data boils down to a question of conformance, which would be the > same whether it is in XML, JSON or RDF. >
ok, fine. I can go with that. and mine about how you know > >> where to start are probably more serious >> > > I think we addressed this on yesterday's 5pm call. We decided to add a > fhir:resourceType property (analogous to the JSON resourceType property), > which would identify the starting instance data element (as its subject) > and its FHIR resource type (as its value): > https://www.w3.org/2016/02/16-hcls-minutes.html#resolution03 so you don't add this to contained resources, or resources in an entry in a bundle? ok. I'll update my code. Grahame