Ontological versus knowledge engineering

The author first discusses the difference between a knowledge base (KB) and a database (DB), which seems to hinge on the `gray box´ verses `black box´ nature of the entries. He then discusses the need for a huge KB to break today´s bottleneck in intelligent systems, i.e. their brittleness when confronted by unforeseen problems. That same brittleness-the representation trap-is what prevents multiple expert systems from cooperating or even sharing rules. The author then considers the central question of the present work: How is the task of building a huge KB different from that of building n small KBs? It is shown that this leads into the realm of ontological engineering, and it is found that there is no single, elegant `use-neutral´ solution to the problem, at least not at present, but that a kind of variegated `tool-box´ approach might succeed