Hi!
I agree with Knut's points. It's probably often a good idea least have
a look in Snomed CT for inspiration, sometimes you will find that
things are well modelled and in other cases that another modelling
would be desirable. The archetypes should be usable no matter if a
country is an IHTSDO-member or not. The same probably goes for other
ontologies like FMA (http://fma.biostr.washington.edu/) or in some
areas other OBO ontologies (http://www.obofoundry.org/). We don't want
to re-invent the wheel _every_ time when archetyping, just replace the
non-functional and invent non-existing wheels...
When looking in ontologies (including Snomed CT) not only the medical
content but also the built in metamodel of the ontology should be
considered so that one can understand the kinds of automated reasoning
(e.g. equivalence checking and production of normal forms) that is
supposed to be possible. In the case of Snomed CT it's referred to as
the Snomed CT Concept Model. For an archetype-to-Snomed-CT-binding
example take a look at the "constructor" mechanism outlined in section
7.4 of the document "Terminology Binding Requirements and Principles"
by Markwell et.al. available at
http://www.ehr.chime.ucl.ac.uk/download/attachments/3375121/TerminologyBindingRequirementsAndPrinciples_v1.0.pdf
While you're at it read the whole document, it partly shows another
perspective (more termimology centric) than the ones often seen on
this list (more archetype centric). The 'truth' (as in scalable
practically useful semantics) probably lies somewhere in between.
Regarding the separation of 'ontology of information' and 'ontology of
reality' I thought it was pretty clear that archetypes deal with
_both_, they refer to 'reality' either formally through
terminology-system-bindings or informally by binding to medical words
and expressions.
Best regards,
Erik Sundvall
erisu@imt.liu.se http://www.imt.liu.se/~erisu/ Tel: +46-13-227579