Tacit Knowledge Extraction
Explicit knowledge: knowledge that can be easily articulated, stored and transferred. Within IT, explicit knowledge is the information that can be transferred from domain experts to business analysts and developers - the information which is found within requirements documents.
Tacit knowledge: knowledge that cannot be easily expressed or understood, or is not readily visible. Even if it can be transmitted, this knowledge often requires complex shared activities to be fully expressed.
While it is straightforward for explicit knowledge to be written into system requirements, it is another matter entirely to encode the subtle inferences and exceptions that exist in real world human processes. This difficulty is at the heart of IT systems� inability to meet project requirements satisfactorily and can lead directly to project failure.
When turning complex processes into IT systems there is a need to capture the requirements of those processes exactly. With traditional IT development techniques, the accuracy of the requirements document will directly affect the success of the project. This is because project IT architecture is chosen based on starting requirements. Once it has been decided, then any changes to requirements may well affect the architecture, and architectural changes are very difficult to orchestrate after the architecture has been built.
As the project progresses and the developers and business experts share experiences tacit knowledge will become apparent within a project as extra requirements. Unlocking tacit knowledge and requirements changes will occur within any sufficiently complex project, and the task of the IT department is to make these changes as easy to implement as possible.
Erudine�s Behaviour Engine operates around the concept of Case Based Reasoning � learning behaviour through cases presented by a domain expert. Erudine learns in a similar fashion to the way humans recall and learn knowledge, meaning the Behaviour Engine can extract the tacit knowledge held by the expert, constructing system behaviour to reflect the results of increasingly complex situations.
The expert uses the Behaviour Engine to build system behaviour through Conclusion & Justification. The software learns as it is told by the domain expert �what� to do with each case and �why�.