You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We have a number of places in the ontology where we either have two versions of the same structure, one of which is a document, and one of which is not (e.g. software.software_component and software.implementation). The idea I think was to ensure that the latter are composed, and the former are linked. We also have places where the clear outcome is that some documents should be composed in and some should be linked.
In the case where documents are composed in, the intention is that the "internal" documents are subject to the parent's life cycle, that is, they don't need their own doc_meta_info.
One assumes in those cases there is no reason for external documents to link to them either.
At the moment, implementations can choose whether to replace attribute values with linked_to stereotypes with a doc_reference instance or an actual instance of the target document type. In the latter case, many documents are serialised into one json document.
Is there a case for allowing document instances not to hold meta info in the instance where they are composed? Allowing this would encourage specialisations to choose composition in some cases, and it would mean we could remove the first case of repeated attributes with document and non-doc versions.
This could be a metamodel issue.
The text was updated successfully, but these errors were encountered:
bnlawrence
added
cimv3
For consideration in a major release
for paper
Need to be addressed before core paper publication
metamodel
Things which are about how the python definition language works and/or what it requires.
labels
Oct 10, 2019
We have a number of places in the ontology where we either have two versions of the same structure, one of which is a document, and one of which is not (e.g.
software.software_component
andsoftware.implementation
). The idea I think was to ensure that the latter are composed, and the former are linked. We also have places where the clear outcome is that some documents should be composed in and some should be linked.In the case where documents are composed in, the intention is that the "internal" documents are subject to the parent's life cycle, that is, they don't need their own
doc_meta_info
.One assumes in those cases there is no reason for external documents to link to them either.
At the moment, implementations can choose whether to replace attribute values with linked_to stereotypes with a
doc_reference
instance or an actual instance of the target document type. In the latter case, many documents are serialised into one json document.Is there a case for allowing document instances not to hold meta info in the instance where they are composed? Allowing this would encourage specialisations to choose composition in some cases, and it would mean we could remove the first case of repeated attributes with document and non-doc versions.
This could be a metamodel issue.
The text was updated successfully, but these errors were encountered: