Enhancing KOS Productivity
While we have not estimated the aggregate productivity impact of KOS,
our discussion does allow us to identify strategies for enhancing that
productivity impact.We mentioned one of these above. The premise of the
SemanticWeb is that computers can draw inferences across databases if
these are each coded in terms of a shared (or at least interoperable)
controlled vocabulary and set of syntactic rules. But after decades of work
the SemanticWeb community is far from achieving consensus on a KOS.
Indeed research has lately turned away from the formal ontologies that it
was hoped would provide the necessary shared vocabulary and rules [3].
Databases are thus coded in terms of a range of different KOS grounded in
different assumptions. The potential advantages of the SemanticWeb will
only be realized if consensus on a particular KOS is achieved. I have argued
elsewhere that the solution may lie in a bottom-up strategy where we first
classify the phenomena (things), relationships and properties required by
the SemanticWeb and then add syntactic rules as necessary [4]. It is
noteworthy that coding for the SemanticWeb necessarily takes a synthetic
approach: the RDF triplets employed take the form (object)(predicate or