Strengthen LINK class semantics


[Due to discussions at the Jan 2003 CEN meeting in Rome]
There may be a need to add a time_validity attribute to LINKs to
model the fact that some causal links will only be valid for
certain times. Solid use cases need to be found for this. It
appears that all the Act.effective_time justifications are already
taken care of in particular classes in openEHR. It is not clear if
there are other reasons to recorded intended/expected
time-validity of a link - one example may be if a provisional test
result is supplied from a lab - the "result" link from a causing
order might have a time-limit set on it.

TB: A more general approach might be to add other_details:
STRUCTURE<T> to LINK, enabling it to include any level of details
needed in the future. The most likely need is further meta-data
supporting workflow data representation, going by experience in
Denmark, for example.

DK: I think we need to be clearer about use cases for this. I am
concerned about putting time validity on links without being clear
about the implications for applications retrieving the source and
or target components. Suggest defer to v1.0 giving us time to
explore further.





Raised By