4/22/2004 6:53
AM
Key questions on Common Upper Ontology
Communication from Dean Allemang
John
F. Sowa wrote:
The greatest misconception is the notion that "controlling vocabulary"
is (1) possible, (2) desirable, and (3) capable of making some still undefined
and unimplemented programs interoperable.
That hope is doomed to failure.
This is exactly what I would have said (and indeed, have
said) before I heard Deb's talk. The
power of the semantic web, I believed, could not possibly come from any
top-down imposition of a controlled vocabulary (or "upper"
ontology). Not only is this a lot of
work, it is doomed to failure (and has failed, many times before).
Or so I thought.
But Deb pointed out some actual
cases from her consulting where considerable value was brought to the
organization by providing some very simple vocabulary control. So while I have always been a staunch critic
of top-down imposed ontologies, her case studies convinced me that in certain
situations, not only is this possible, but even valuable. I am always strongly influenced by real
success stories, even when they disagree with what I previously thought. And as I have no reason to believe that Deb
was exaggerating or fabricating her stories, I take them as facts to be reckoned
with. Having said this, nevertheless,
Deb herself went on to say that she saw more value in the "uncontrolled"
approach of wikipedias and blogs; but here I am putting words into Deb's mouth,
so I'll stop that.
As for the rest of your comments about controlling
vocabularies for wikipedias, I am a bit confused, as we seem to be in violent
agreement; I was using blogs and wikipedias as a /contrast/ to controlled vocabulary
approaches, since I take it (as you have pointed out in great detail) as given
that no controlled vocabulary can or should be imposed on such a process.
Dean