Hello Philippe,

I thought that I would add some thoughts that are probably not concrete
enough to be useful, but who knows?  The thread that has been running on
ontologies is been very interesting, but has seemingly been polarized in two
camps.  A school of thought that believes in evolutionary growth and the
school that contents that central growth is preferable.  I did a search on
http://www.google.com  with the search term of ontology and I received back
154,000 hits.  A link that I followed that was pretty good with sorting out
some of the differences in the use of the term was
http://www-ksl.stanford.edu/kst/what-is-an-ontology.html  It seems that a
great deal of work has been done in this field, is there no room for two
methods of usage?

My opinion is that the "correct" answer if there can be a "correct" answer
lies in the middle of the domain.  Dr. Ho's evolutionary growth mimics the
entropy growth found in the open-source movement which mimics other patterns
that others very much more observant then I have seen.  However, I think
that it has an initial-state problem, in the fact that I don't think it can
bootstrap from nothing.  One of Eric Raymond's (ESR) preconditions for an
successful open-source project is that there has to be something that works
partially before anyone starts.  Please, I know that ESR papers on
open-source have nothing to do with ontology development, but I think that
some similar principles are at work.  He, (ESR) is much more eloquent then I
and if I can explore other ideas from some of his frameworks, then I believe
I communicate more effectively.

On the other hand, Philippe's centrally managed ontology which will start
out strong is vulnerable to being too rigid and vulnerable to organizational
and political problems.  It has been my experience that an organization will
modify or bend rules to accommodate certain clinicians.  If a planning
community is building a ontology then it is possible that they might be
contaminated by political bodies or industry groups.  Dr. Ho has pointed out
that it may be vulnerable to drift effects as well.

My thoughts are to try to build a ontology that has a strong core that can
be grown evolutionary from that core.  A model that would have a built-in
"blindness" to other concepts to allow them to be grown on demand.  A search
query engine that "knows" what is  core and what is not and runs different
patterns or methods based on that "knowledge"  At some point in the future
if the core is obscenest then update the core with the latest snapshot that
incorporates key features of the grown "outer" material

FM-200 system on standby
I await the flames.

Todd Smith <[EMAIL PROTECTED]

Reply via email to