Object-oriented framework for knowledge-based systems

Transforming elicited knowledge acquired fi om a human expert in some domains into a machine readable format has always been a crucial step in developing knowledge-based systems (KBS). This issue of knowledge representation has long been considered as a bottleneck in the development of a KBS. Schemes used to represent such knowledge include production rules, sementic nets, scripts, frames, and logic, as well as more conventional representations such as arrays. These representations suffer several drawbacks which mainly result in (1) poor reusibility of knowledge across multiple KBSs, and (2) poor transportibility, which makes it difficult to move KBS across multiple operating platforms. These two main shortcomings have contributed to the characterization of KBSs as being single-domain systems which lack the ability to have their knowledge shared across multiple domains. This paper presents a discussion on utilizing object-oriented features to deal with the drawbacks addressed above.