My Vision
Support for Knowledge Modelling, Management and Maintenance


Today's economy may be called a Knowledge Economy indicating that stake-holders of the right kind of knowledge may gain competitive advantages and strive in this new economy. This knowledge may be roughly divided into two types of knowledge: internal and external knowledge. The internal knowledge is the Corporate Knowledge within an organisation and the external one is knowledge about the economic environment the organisation operates in.

A modern enterprise today is often a virtual entity which consists of many sub-organisations which are distributed across different geographical areas, each possessing different expertise and specialising in certain functions. This complicates the task of treating Corporate Knowledge as a whole and making effective use of it. Furthermore, the economy in which an organisation operates is very dynamic which requires an organisation to react appropriately and promptly --- in adapting their goals and processes.

The proposed work investigates potential methods and (automatic or semi-automatic) support that can help the process of capturing, managing and maintaining the Corporate Knowledge. A focused approach is the Multi-Perspective Modelling (MPM) approach.

The Multi-Perspective Modelling (MPM) approach makes use of multiple models to describe different aspects of an organisation to present insights for the concerned business areas and to allow specific analysis to be performed on them. These models together present a consistent and coherent view of the Corporate Knowledge.

This approach is already used in research and practice: UML has offered a suite of inter-supportive modelling notations[1], Ulrich Frank makes uses of multiple notations and builds a multi-perspective knowledge management system (MEMO)[2], in Zachman's Framework for Enterprise Architecture, it suggests using a variety of modelling languages to capture and describe the different aspects of a domain[3], Common KADS methodology embodies several modelling languages to help understand and capture domain knowledge and to help the design of knowledge based systems[4].


Although the MPM approach is a valuable approach for knowledge management and has been accepted and practised by industries and researchers, there are still several challenging issues to be resolved.

Some of these challenges are given below:

1. Quality of Knowledge:

Typically, several models are used in a MPM approach, although they are used to describe the same (knowledge) domain, the content of each model may not be consistent. The interesting issue is how these models may be checked, criticised and corrected. How AI techniques, e.g. formal methods, may be applied to help such a process. Can generic methods be derived from it? To which extent can such support be provided?

2. Analysis of Knowledge:

As a part of modelling activities, knowledge that is represented in the (partially) developed models can be extracted for analysis of specific purposes. This analysis can be on the domain-specific knowledge, organisational context, focused issues, or purely on the demonstration of knowledge that the models have described.

3. Use of Knowledge:

After models have been formed, it is important that insights of the models are delivered to the right people at the right time and in the right format. This links directly to the objectives of building the models and motivations for using them. It will be interesting to see how the process of making use of knowledge can be helped by providing (generic) automatic support (under-pinned by formal methods).

4. Forming a Consensus:

All models that have been built in a MPM initiative together reflect a consensus of vision of key personnel in an organisation. The process of forming such a consensus is typically informal and not structured. It will therefore be interesting to see to which extent (automatic and semi-automatic) support can be provided. This is to investigate technologies and methods which provide help to collect and document modelling rationale and gives a platform for discussing and debating issues, and eventually help to reach a desirable consensus.

5. Testbed Application:

If some of the above proposed work actually happens, when it is appropriate, it may provide a basis for one of the testbed applications.


[1] Grady Booch, James Rumbaugh, and Ivar Jacobson, The Unified Modelling Language, Object Technology, Addison-Wesley, February 1999.

[2] Ulrich Frank, Multi-Perspective Enterprise Models as a Conceptual Foundation for Knowledge Management, Proceedings of Hawaii International Conference on System Science, Honolulu, 2000.

[3] John A. Zachman, The Framework for Enterprise Architecture. and Zachman Institute for Framework Advancement.

[4] G. Schreiber, B. Wielinga, and J. Breuker, KADS: A Principled Approach to Knowledge-Based System Development. Academic Press, November 1997.