Small inconsistency in EHR 1.0.2 PDF vs. HTML versions

In the PDF, EHR.system_id is a String, in the HTML version EHR.system_id is a HIER_OBJECT_ID.

It seems at some point in the HTML generation that change was introduced and was also migrated to newer versions of the EHR spec, though there is no change log associated with the change in the type.

Looking at the JSON schemas, it seems that modification was taken there too, so the type is HIER_OBJECT_ID.

If this was a mistake in the HTML generated, we are spreading this error all around. If this change was planned, certainly the v1.0.2 HTML is incorrect, but if v1.0.3 does include this change on purpose, we are missing the change log for this.

@thomas.beale will know better, but looking in my archives, I think the EHR.system_id was introduced in v0.96 and was always a HIER_OBJECT_ID, or at least I found it in v1.0, v1.0.1 and v1.0.2 pdfs (see for instance EHR IM pdf in openEHR - Release 1.0).
The problem is that the diagram was incorrectly defining it as string, while the class description was always as HIER_OBJECT_ID.

You are right @sebastian.iancu , I was only checking the UMLs.

1 Like

Checking other parts of the model, AUDIT_DETAILS.system_id in RM 1.0.2 is String in both UML and text definition. Shouldn’t that be HIER_OBJECT_ID too?

@sebastian.iancu @thomas.beale what do you think?