Résumé: Many large- scale endeavors, especially the ones that need to be performed on an occasional basis, are performed on a de novo basis. Whether one is designing a spacecraft or a dam, or launching a sales campaign, each endeavor tends to have a life of its own. Very little knowledge, if any, is transferred from one business or engineering design endeavor to a successive one. This paper describes an approach that involves the acquisition of raw information from the keystrokes entered by the user while performing the primary activity. Such information is distilled to produce knowledge for future use, without imposing additional burden on human users involved in the .rst endeavor or the succeeding endeavors. The paper also describes a complementary approach that signi .cantly reduces the time and e .ort spent by human beings in articu- lating their requirements, and speci .cally, in terms of delineating the indi .erence points among a set of feasible alternatives. Both these approaches have been validated in a complex multi- organizational situation related to spacecrafts. The approaches can be applied, either singly or jointly, to a broad class of business and engineering scenarios. They incorporate the notion of automated knowledge acquisition, discovery, management, and dissemination. The proposed approach is discussed in the context of supporting workteams that are distributed both on geographic and temporal basis, thereby creating a ‘‘24- Hour Knowledge Factory ’’environment.

Langue: Anglais