Rong Chen wrote: > Hi Sebastian! > > It is a very nice piece of work you and Stefan have done! I fully > agree that we need to make archetypes compatibility checking explicit > and this is a very good start. > > It will be useful to publish the algorithm that you used for > comparison so others could comment. And the same algorithm could be > implemented in different tools to ensure interoperability of > archetypes. > I have just been thinking about the specification for this area myself, and how we should approach it. In our software, we have just successfully done the 'flat-form' -> 'differential form' transformation of specialised archetypes, and we are now working on the differential -> flat-form transform (i.e. compression through inheritance). These two transforms allow hierarchies of specialisations to be properly processed. The excellent work Sebastian mentions above gives us some generic differencing, even between non-parent/child pairs. There are probably some more variants we need as well.
My proposal would be that once we have perfected the transformations/comparisons in software, that we document the algorithms in the Archetype Semantics document (currently an early draft lives here - http://svn.openehr.org/specification/TAGS/Release-1.0.1/publishing/architecture/am/archetype_semantics.pdf). This will provide guidance for other tool builders, and a solid reference for what we believe the semantics are. - thomas
