ArchetypeTerm.code is mandatory in specification and BMM, but often null

[J3173]

The ArchetypeTerm.code is mandatory in specification and BMM, but often null when parsing an archetype that doesn't explicitly set them. Maybe this should be automatically filled from the map key?

[borutjures]

If it helps – I fill it from the map key in my implementation when parsing ADL.

[wolandscat]

It is always filled from the 'map key' - that key is the code. We never included the code inside the Archetype Term structure to avoid noise and duplication. Although we are not consistent about that - in BMM itself, we do have the duplication all over the place. But at least archetypes are easy to read ;)

[JeroenBrinkman]

The same goes for ArchetypeTerm.description, null is accepted when the spec says its mandatory. There is a workaround in defaulting to empty string, but the mismatch with the spec is annoying.

For example something like this is accepted even though it is illegal according to the spec:

terminology
    term_definitions = <
        ["en"] = <
            ["ac1"] = <
                text = <"Illegal Term">
            >
            ["ac27"] = <
                text = <"Legal term">
                description = <"this one is legal">
            >
      >

[pieterbos]

It would probably be good to add a validation for these. For example in https://github.com/openEHR/archie/blob/master/tools/src/main/java/com/nedap/archie/archetypevalidator/validations/BasicTerminologyValidation.java , with a clear error message. And then fix the test so the error never triggers :)