Standardise the following change of behaviour for VERSIONs that have lifecycle_state set to incomplete :
When committing incomplete data, the validation should be relaxed, i.e. validation of [occurrences and] cardinality lower bounds is skipped, in other words, allow missing data.
All other validation (cardinality upper bounds and constraints on values) would be performed normally.
Additionally, when querying, data from incomplete versions would be completely ignored, unless accessed in a specific way. This requires a rule to be added to AQL.
see also https://discourse.openehr.org/t/special-treatment-of-incomplete-versions/420
I’m not sure, but you might be right to just include that into the AQL spec.
Is it a requirement to have versioning also on incomplete or should we support overwrite of incomplete versions?
I would allow versioning - that is one of the advantages of persisting the data within the EHR rather than externally like the Better /draft compositions.
I agree with - I would not change the versioning behaviour just because the authoring mode is draft. I would say that an external ‘crash-save-cache’ can still be offered by any product (even this editor here in Jira now remembers your uncommitted comments), but this is another level of functionality.
I agree.