Tim Churches wrote: > Thomas Beale wrote: > >> W > OK, I know this is an openEHR list, but nevertheless I don't think that > openEHR can claim exclusive use of the word "archetype" to refer only to > artefacts which are expressed in openEHR ADL. From such a claim it is a > slippery slope to having to refer to them as archetypes?, archetypes(TM) > or archetypes?. > we're not interested in that obviously. But to my knowledge, "archetypes" are used in the language of IT in only two places: openEHR, where they are formally defined, and by the plone people, where they area kind of templating mechanism for use within plone/zope systems. Only the first has a formal definition as far as I know. One of the reasons we chose the term (in 1997) was that it wasn't used in IT (the plone stuff didn't exist then). > Strictly speaking, an archetype is not a set of specifications or > constraints at all but rather (according to WordNet) "an original model > on which something is patterned" - that is, the master **instance** of a > thing, a prototype, from which specifications can be derived. openEHR > seems to be using the term archetype in the later, Jungian sense of "an > inherited pattern of thought or symbolic image that is derived from the > past collective experience of humanity and is present in the unconscious > of the individual". > strictly speaking, in openEHR, an "archetype" is exactly what it is: a specification defined (or definable) in the Archetype Definition Language. Archetypes in openEHR are indeed "original models on which something is patterned", which is why we thought this definition appropriate. In a mathematical sense, it is possible to have 1,000 instances of BP measurement data, all conforming to the BP measurement archetype. I think this is a very reasonable use of the language. > My practice has been to refer to "openEHR Archetypes" to clearly > distinguish them from other uses of the English language word "archetype". > Since we are in IT, personally I don't see that it matters too much - as I say, there are only openEHR archetypes and plone archetypes (so far). > Perhaps on this and other openEHR lists, the term archetype could be > taken to mean "OpenEHR Archetype", and other types of archetype could be > distinguished as necessary by suitable qualification, including the > generic "non-openEHR archetype". But alternative uses of the term > archetype should not be denied. > I disagree: since we have taken the trouble to define the use of the word in a formal informatics sense in a certain way, and to define a language and model for archetypes, then using the word "archetype" in the IT space either means the openEHR usage of archetype, or else the plone usage. But calling other things that are not archetypes by that name just creates confusion, since people will expect to find real archetypes that can be expressed in ADL, processed by the tools and so on. I don't see this as unreasonable, since we pioneered the use of this word in informatics to mean what it now means - it is not as if anyone would have previously pointed to HL7 RMIMs or any other kind of model and called them "archetypes".
In the hope of a less confused world... - thomas beale
