kads

KADS is a structured methodology for the development of knowledge-based systems, emphasizing knowledge elicitation, modeling, and design. It originated in the 1980s as a framework to improve the analysis and construction of expert systems, promoting a systematic approach to capturing and organizing domain knowledge and the reasoning processes that use it.

The core idea of KADS is to model a problem domain from multiple viewpoints to ensure a comprehensive and coherent design. The methodology employs five interrelated views often summarized as Organization, Task, Agent, Knowledge, and Communication. These views help separate organizational context, task requirements, the capabilities of agents or actors, the domain and inference knowledge, and how knowledge is exchanged or communicated. Together, they support a General Model of knowledge-based systems and guide the selection or construction of appropriate problem-solving templates.

A typical KADS development process includes domain analysis to understand purpose and constraints, knowledge modeling to capture domain rules and reasoning patterns, and system design to translate models into a workable architecture. The approach favors iterative refinement and validation, using explicit representations such as diagrams and templates to keep stakeholders aligned.

Impact and use: KADS influenced knowledge engineering practices beyond early expert systems, contributing to structured methods for requirements analysis, knowledge specification, and design. While later methodologies have evolved and adopted different notations, KADS remains noted for its emphasis on multi-view modeling and its structured approach to separating organizational context, tasks, and knowledge from implementation details.

See also: knowledge engineering, expert systems, knowledge modeling.