I agree it is a balancing act in how far the semantics should be in the
RM or in the archetypes.
Both ways have their pro and contra.
Thanks for explaining it
Bert
On 28-08-15 19:17, Thomas Beale wrote:
Hi Bert,
On 28/08/2015 16:32, Bert Verhees wrote:
On 27-08-15 19:54, Thomas Beale wrote:
I would suggest that CIMI has been simiplified to the point of not
being directly usable as an RM by openEHR or 13606 - most of the
needed context information is gone in CIMI, and it doesn't
distinguish any kind of 'Entry' or clinical statement.
Are you saying, that the context information from the reference model
is not used?
the CIMI RM
<https://github.com/opencimi/rm/blob/master/model/Release-3.0.4/BMM/CIMI-RM-3.0.4-generated-from-UML.bmm#>has
no context information in it.
This was a conscious choice in the CIMI community, designed to get
buy-in from a much wider range of stakeholders than openEHR or 13606
deals with. Technically, the CIMI approach is to soft-model nearly
everything in 'reference archetypes'.
and the archetypes fill in the missing reference model context parts?
that's the idea.
If so, then this makes the two level modeling approach, of course,
much more flexible, a kind of new database approach/technique, usable
for virtual anything.
it makes it more flexible in one sense, but also harder for
implementers - now they cannot know where even basic context like
subject, times, locations etc are - all that has to be obtained from
archetypes. The 'flexibility' comes with a price...
What goes in any particular RM for some particular domain or industry
needs to be the result of careful analysis of
* the need for being able to build reliable software components that
can assume some things
* the need for a base model with enough useful primitives that it
doesn't force endless repeated modelling of the same basic
concepts in archetypes
* but sufficient flexibility so that all the variability of the
domain, and also localization can be accommodated.
It's a balancing act.
So far in openEHR, the context and most other structures etc have
proven to be good. We'll probably get rid of / simplify the ITEM_TREE
stuff in Release 1.1, but I can't imagine getting rid of most of the
other semantics.
- thomas
_______________________________________________
openEHR-technical mailing list
[email protected]
http://lists.openehr.org/mailman/listinfo/openehr-technical_lists.openehr.org
_______________________________________________
openEHR-technical mailing list
[email protected]
http://lists.openehr.org/mailman/listinfo/openehr-technical_lists.openehr.org
