Zola: a language to operationalise conceptual models of reasoning

This paper presents Zola, a language that allows the operationalisation of conceptual models of expertise. Zola is a general knowledge representation language, more specifically defined to support the definition of conceptual models by Knowledge Level prototyping. It does not include constructions specific to a particular methodology (as tasks or roles), but provides general primitives that allow the definition of such specific concepts. The objective is, by providing a “neutral” language, to allow the modelling team to elaborate a conceptual model with the least possible bias introduced by the language concepts. Zola provides three levels of representation: domain knowledge, operations (that can manipulate the domain knowledge), and strategies (algorithms). Zola is a reflective language: operations can analyse other operations. This allows the construction of reflective operations that can give the modelling team some feed-back on the implemented model. After a short presentation of the objectives of the project underlying Zola construction, we describe how Zola has been used to operationalise a conceptual model of reasoning for a fault diagnosis system. Finally, we analyse Zola according to criteria proposed in the literature to compare such languages.