This paper describes a language that is intended to support software engineers in the development of information systems throughout the software lifecycle. This language is not a programming language. Following the example of a number of other software engineering projects, our work is based on the premise that information system development is knowledge-intensive and that the primary responsibility of any language intended to support this task is to be able to formally represent the relevant knowledge.

Information systems technology, computer-supported cooperative work practice, and organizational modeling and planning theories have evolved with only accidental contact to each other. Cooperative information systems is a relatively young research area which tries to systematically investigate the synergies between these research fields, driven by the observation that change management is the central issue facing all three areas today and that all three fields have indeed developed rather similar strategies to cope with change. In this paper, we therefore propose a framework which views cooperative information systems as composed from three interrelated facets, viz. the system facet, the group collaboration facet, and the organizational facet. We present an overview of these facets, emphasizing strategies they have developed over the past few years to accommodate change. We also discuss the propagation of change across the facets, and sketch a basic software architecture intended to support the rapid construction and evolution of cooperative information systems on top of existing organizational and technical legacy. 1.

Data Type which consists of an object and an execution submodel. Both are influenced by the context and 4 organization in which their instances are used -- large communities of software developers, users, and managers. Change, complexity, and heterogeneity of representation are ubiquitous in specification management. They must be addressed in the object submodel, e.g., by representations for versioning, configuration, and mapping between different object views. The execution model has to cope not only with the problems of multi-version concurrency control but also with explicit communication and coordination in large development teams, possibly distributed in space and time. The meta data model that supports such a setting could in principle be very loose (e.g., hypertext), or it could require full formalization of the available language concepts, methods, and tools, even of the roles of human developers. Probably, neither extreme should be followed. Simple hypertext structures have ...