Tim Benson wrote:
> > -----Original Message-----
> > From: Thomas Beale [mailto:[EMAIL PROTECTED]]
> > One of the mistakes termset designers are making, I believe, is that they
> are trying to model
> > everything; it would be better if they used something like archetypes or
> templates for
> > informational compositions like test results etc. A biochem test panel
> today
> > might have only 3 items in it, tomorrow, there may be four. These kinds of
> > things are too volatile for modelling in termsets. Volatility has to be a
> > primary criterion for termset inclusion (after relevance).
>
> I agree. However, you need a way of naming commonly used terms such as Full
> Blood Count or Urea and Electrolytes. Then there is the problem of
These terms do need to be in the termset, since they are commonly used. The real
question (I htink) is what they are semantically connected to in the termset
(versus what they might be connected to or aggregated with in an archetype).
> distinguishing the term FBC in an order and in a result. Are they one and
> the same concepts or are they different?
I think they should be the same. Termsets should be contextless.
> There has long been a debate between those who think that terminologies
> should be comprised of:
> 1. Atoms and hence taking the view that most terms are compound
> 2. Molecules (useful stable things)
> 3. Combinations of molecules (such as an FBC)
>
> Personally, I find the molecular approach the most straight forward. Where
> do you stand? If we use archetypes or templates (and I am afraid I still
I'm not the terminology expert in GEHR, but I feel comfortable with 2. as well.
3. implies context which termsets (IMO) should not know about (not on principle
so much as it's just impractical, since you get an exponential explosion of
similar terms in different combinations due to differeing contexts).
Naming of elements in archetypes is dependent entirely on underlying termsets -
we are not trying to come up with new names or codes, just modelling how they
are put together in particular usages.
- thomas beale
