Over the past few years for engineering groups there has been a steady migration from simple collaboration and database software to data mining and knowledge management systems. This trend stems from the need to organize data and knowledge in ways that facilitate users getting relevant information spontaneously pushed to them whenever the system detects that it is something they should be informed of. No longer does a user have to make a proactive search. Over time the system becomes context aware and develops a peripheral vision™ capability from which the users benefit.

The basis of knowledge management is a linked network of data artifacts or knowledge network. By creating various types of links under version control it is possible to develop very sophisticated algorithms which detect facts, trends, and anomalies that would not be available in traditional tools.

Consider for example the case where a software company is producing a new product release every six months. With knowledge management it is possible for the system to track the arrival rate and closure rate of design defects and compare it with the profile of the same parameters averaged out over all previous releases at this stage in the release cycle.

Another often cited example for linked artifacts is where a rigorous requiirements management design process is being followed. Product managers can establish product visions and map feature releases to customers and business cases. Systems engineers can link the features to detailed system and sub-system requirements. Test engineers write test cases which are linked to groups of requirements so that an accurate test coverage can be calculated and at the same time test cases can be linked to defect reports so that the fault level can be monitored.

As the network oflinked artifacts gets more heavily populated it becomes possible to make sensitivity and dependency analyses. If a customer declares that he no longer has an intrerest in a product, then this information can flow in real time right through the value chain. The product manager can identify which features were being developed specifically for that customer and are no longer needed. The project manager can identify in real-time who is working on these requirements in the entire team of systems architects, design engineers, testers and start to optimise the project plan accordingly.