Remove statement requiring node-id redefinition in case of occurrences redefinition.




Ian McNicoll
May 21, 2020, 12:20 PM

Sure but you may not want to rename the template,at least as part of the root name - this probably applies more to Entry or SECTION level templates. And I guess that is probably the difference between a specialised archetype and an embedded template. The former is a genuinely new ‘thing', whereas the latter is really just a set of pre-package constraints e.g a template around a blood pressure entry, that may or not have a slot filled is still just a blood pressure semantically from the patient perspective, even if the template (id) needs to be quite localised.

Thomas Beale
May 21, 2020, 12:45 PM

I do get your point, but nevertheless, one can write 10 (or a hundred) variant templates, including embedded Entry-level ones based on the same original archetype. With no specific naming, how would you distinguish each of those derivatives from each other? Now you might say: but we only want one! To which I respond… True, until someone wants the second one…

Pieter Bos
May 21, 2020, 4:00 PM

Maybe tooling can help here? It could notify which nodes in which child archetypes need to be reviewed after a parent archetype change.

Also I think the deprecated validation rule VSONIR may need to be removed or updated in the AOM 2 specification. It now reads:

Deprecated: VSONIR specialised archetype redefined object node identifier condition: the node identifier of an object node in a specialised archetype that is a redefinition of a node in the flat parent must be redefined if any of reference model type, node identifier definition in the terminology, or occurrences of the immediate object constraint is redefined, with the exception of occurrences being redefined to {0}, i.e. exclusion.

Which is in direct contradiction with the ADL 2 specification.

Ian McNicoll
May 21, 2020, 8:59 PM

yup - we will definitely need tooling to resolve some if these issues - just anxious to minimise that need unless there is clear technical requirement. I had a chance ot see an actual example and it is probably more readily solved by tooling - option to update the translation on a particular node, as the real problem is just with partially translated archetypes with ‘dummy’ translations. They are then pretty obvious and can be fixed on a node by node basis

Ian McNicoll
May 22, 2020, 9:33 AM

“With no specific naming, how would you distinguish each of those derivatives from each other?“ I agree but this is best done in the templateID or proposed ‘template name' , rather than the root nodeId name. However - needs a bit of discussion in the modelling community.


Thomas Beale

Raised By

Thomas Beale


Affects versions