Providing Knowledge Management Support to Communities of Practice through Agent-oriented Analysis

This paper concerns the need for methodological support in providing Knowledge Management (KM) IT solutions. Due to the distributed nature of knowledge, the support of KM often requires complex, distributed IT systems, which are inherently difficult to design. We propose an agent-oriented methodology for the analysis and design of KM systems that offers appropriate abstractions for modelling and designing the characteristics of the organizational setting of the system. The method is illustrated using a fictitious scenario where a newcomer in a knowledge organization decides to join an existing Community of Practice (CoP) in order to share knowledge and adjust to his new working environment. Our proposal uses the Tropos methodology to analyze an organizational setting and to elicit the requirements for a KM system, and proposes the Agent-Object-Relationship Modelling Language (AORML) for the architectural and detailed design. Aside of this primary goal, we also intend to point out needs of extending the expressiveness of the current agent-oriented modelling language we are using and to check where the methodology could be improved in order to make it more usable.