Mediation in Information Systems GK3 WIEDERHOLD Department of Computer Science, Stanford Uruversltv. Stanford. Cal Lforn~a ( gLo@db.stanford.edu). Information systems can be defined as computing systems that provide information for their customers, often in a decision-support role. They are a large and increasing fraction of modern computing, slowly eclipsing the effort expended on massive data processing, cycle-intensive scientific computing, and critical real-time systems. They are also quite visible, since they have a large number of direct users. However, they do not stand alone. Planning and other decision-making tasks require information from diverse sources such as databases, reference systems, data obtained in real time from sensors, and analyses projecting trends over the totality. The databases are mainly part of operational systems, built to satisfy local objectives. Useful data are available from public services, both free and for-profit. Processing of sensor inputs and simulations requires the technologies of scientific computing and produces results in widely different formats. Most of these resources are autonomous with respect to decision-making tasks. Autonomy, in turn, leads to heterogeneity, including representation, scope, level of abstraction, and context. Rapid construction of information systems is being aided today by the many tools provided by vendors of client-server systems.
/lp/association-for-computing-machinery/mediation-in-information-systems-SOwPNvR3XK