Traditionally, building knowledge-based systems has been a process involving professional knowledge engineers. While this process works for well-defined tasks in limited domains, it is too expensive for domains where the representation must change frequently and is not feasible for problems where experts' understanding of the domain and task changes during the course of problem solving. The alternative approach investigated in this project is one of "incremental formalization." This approach allows users to initially enter their understanding of their domain, task, and solutions in less formal representations and provides computer support for the gradual formalization of this knowledge. The project includes: (1) designing knowledge representations and interfaces for capturing emergent understandings and enabling incremental formalization, (2) developing mechanisms for recognizing implicit structure within informally represented information, (3) investigating interfaces for suggesting formalizations, uses of automated formalization, and trade-offs between suggesting and automating formalization, (4) creating tools which aid the development of shared languages and support collaborative knowledge-building tasks, and (5) improving understanding of knowledge-building processes and the relative merits of prototypes through observations of individual and collaborative tasks requiring domain modeling. www.csdl.tamu.edu/CKBE/
