A METADATABASE APPROACH

Repository systems have emerged recently as a cornerstone to information management. Their envisaged purposes include (1) applications development in a CASE (Computer-Aided Software Engineering) environment, (2) enterprise information resources management (as a passive kernel for information integration), and (3) global information processing and management (as an active facilitator/integrator). Accordingly, their contents comprise metadata that model functional (sub-)systems and metadata that formulate systems interactions. In other words, the scope of metadata encompasses both data resources and knowledge, and hence requires new methods and techniques to manage these information resources in a unified way. Most previous metadata systems are limited to containing only data resources models. A metadatabase system is being developed at Rensselaer for information integration in heterogeneous and distributed environments. This paper presents its core structure, the GIRD (Global Information Resources Dictionary) model fo