| Popis: |
A basic feature of human nature is the propensity to construct boundaries which define territorial possession. Such assumed possessions are often jealously guarded by their owners, and a consequence of this primitive instinct is the emergence of subject specialists who exercise in-depth ‘lordship’ over their domain of expertise. For many years computer science has made attempts to relate to this phenomenon of expert intellectual property by developing mechanisms in software to emulate reasoning capability. Correspondingly this has resulted in the development of intelligent knowledge-based (or expert) systems, along with their attendant processes of knowledge elicitation, representation and exploitation. This paper attempts to define a context for knowledge engineering, the term being used to define spanning the void between domain expertise and the intelligent knowledge based system. It goes on to describe the systemic development of a particular solution to the knowledge engineering problem which is underpinned by a software environment called VEGAN (a Visual Editor for the Generation of Associative Networks). Many attempts have been made at bridging this gap, and VEGAN represents a significant aid to the knowledge engineering task, in the context of frame-based systems. Rather than attempt to create a unidirectional information path from expert to computer system (or knowledge engineer), VEGAN presents a common forum for discussion about, and exploration of, the expertise of the domain specialist. By doing so it helps the flow of information between the two parties. VEGAN represents an approach to a human-natured design of a software system which: 1. empathises with the ‘culpture’ of the expert, 2. provides a bridge between the expert and the computer system by shielfing the expert from the underlying complexity of the system, 3. aids the study of the organization of expertise, and thus induces further information. |
