Ewout Kramer

Interestingly, microsoft has left out the Async methods on the newer System.Text.Json serializers - probably because the overhead is too large. In fact, our experience with the validator shows that...

The new `FhirPathCompilerCache` solves this problem now!

We're actually concentrating our effort on System.Text.Json nowadays, so I don't want to do this.

Yes, that seems like a good idea. There is already a universal Cache class but that's not used across the board. I will keep this in mind when I refactor...

> Up until SDK 5.7.0 this happened to work because TypedElementOnSourceNode retained its location verbatim from the underlying ISourceNode. Indeed, "happened" to work ;-) We fixed that bug (#2642) in...

BTW: since in many cases this is a wrapped source node on a wrapped itypedelement that wraps an isourcenode, we can save cycles here by polling the `Annotations` for `ISourceNode`...

Pretty sure @brianpos was working on this?

Yes! It has all landed on the right spot I think!

Though this looks nice, I think this is only applicable when you do the "built-in" object -> json serialization. FHIR Json serialization is complex enough that we wrote our own...

So, to finalize this discussion: # On the serializer It is a feature that the serializer is able to serialize incorrect data, for three reasons: * The POCO model can...