Philippe AMELINE wrote:
> Thomas Beale answered :
>
>> I have to admit that I don't yet have a special term for this -
>> usually I jsut say something like "data conforming to archetype xxx"
>> or so. Actually, I don't think "archetype instance" is such a bad
>> term. Maybe "archetype exemplar". I'll have to think on it.
>> Suggestions welcome...
>
> From my point of view, an Archetype is a model ; it may also be seen
> as a "data mold".
in this sense, "template" is the usual word, but we consciously avoided
this because archetypes are constraint models, not just "cookie cutters"
- that is, two pieces of structured data which look quite different can
in fact conform to the same archetype.
> Hence - back to the debate about ontologies and Archetypes - one can
> only speak, as Thomas said, of a description structure conforming to
> an Archetype.
>
> I think it is important to distinguish 2 functions for Archetypes :
> description validator and description mold.
right. In my original paper and in some other places I have talked about
the validator function, which is what the GEHR kernel does. The
implementation of our new openEHR models will do it in a much improved
way...
> The function of description validator is easy to understand ; I can
> give an actual example of the function of data mold :
> The Odyssee project has been granted some funds from the french
> Ministry of research in order to investigate the domain of continuity
> of care. The key concept in that matter is to be able to manage the
> concept of "points of view" : how can we display a common set of datas
> in a way that is accurate for a GP, but also in a different way,
> accurate for a cardiologist, and in a third way accurate for a
> gastroenterologist, and so on.
>
> We plan to use Archetypes as "transformation rules" from a global
> description to specific descriptions.
aha, very intersting Philippe - we must talk about this!
> I hope Thomas didn't plan too much hollydays for years to come ;-)
uniquement en France...
- thomas beale
