This paper describes OPAL, a knowledge editor intended for use by medical professionals for encoding cancer-therapy management expertise. Techniques for building high-level knowledge acquisition tools are in short supply, which makes OPAL’s contributions particularly interesting to read. The authors employ the semantics of the application domain to structure and control the means by which a domain expert can view, modify, and create knowledge bases. The emphasis in OPAL’s approach is on minimizing the reliance upon a knowledge engineer for constructing the application knowledge bases. This necessitates the requirement to abstract the internal representations to structures that the experts are comfortable in directly dealing with. The authors rely upon advanced graphics techniques in providing an interface to OPAL. The interface allows experts to enter various types of domain entities, relations, actions, predicates, and procedural knowledge. Through an intermediate representation, this information can be transformed back and forth into the internal hierarchical object representation, with its associated production rules and finite state table.

The best part of this paper, though, is the clear and succinct style in which the OPAL techniques are categorized and are compared with other knowledge acquisition techniques that have been developed to date. Given the highly empirical nature in which most of this research is conducted, this is a worthy accomplishment. The paper therefore also provides an excellent review of the general class of knowledge acquisition techniques to which OPAL belongs.