The EuroKnowledge standardisation initiative grew from awareness that knowledge technology has an increasingly important role to play in systems development, and that full exploitation has not been made of established knowledge formalisms.
An underlying issue is that without the ability to reuse knowledge from one system to another, the cost of development will continue to be excessive. In addition, without the ability to communicate knowledge, computer systems will remain data processors. It is the consortium's view that the way to remove these obstacles is to develop standards facilitating knowledge reuse and exchange:
- reuse means a more efficient and less costly development process leading to more predictable and reliable result; furthermore, it is possible to capitalise on corporate knowledge.
- exchange permits both intra and inter organisation knowledge interchange and sharing, with standard interfaces enabling interaction between conventional software systems and knowledge-based systems.
To expedite knowledge reuse and exchange, EuroKnowledge is establishing recommendations for knowledge level standards; currently, at this level, there is evident deficiency. It is here that Europe has greatest experience and renown. Americans are addressing the "lower" levels of knowledge representation (file-interchange and implementation/source levels): these computationally oriented expressions of knowledge are forcibly the "lowest common denominator" and are simply inadequate for implementation independent representation of knowledge interchange with non-computer professionals.
Knowledge level elements addressed by EuroKnowledge are:
_-knowledge representation formalisms: recommendations for high-level modelling of knowledge, covering both static description of concepts and their inter- relations, and description of the dynamics (i.e., know- how: the way in which reasoning is carried out on static knowledge),
- domain ontologies: the structure, criteria for construction and acceptance of domain-specific ontologies, building on the knowledge representation formalisms just noted
- problem-solving models: the structure, criteria for construction and acceptance of generic problem-solving knowledge models, again building on the knowledge representation formalisms noted.
EuroKnowledge results take the form of recommendations on:
- Knowledge level representation formalisms.
- Domain ontology expression and criteria.
- Problem-solving model expression and criteria.
EuroKnowledge works by way of issuing Requests-for-Commentary on consortium proposals, and Requests-for-Action for original contributions, to the knowledge engineering community. The project maintains a Web page, participates in relevant conferences and workshops, and issues bulletins ("EuroKnowledge News").
EH1 1HN Edinburgh