Acta Electrotechnica et Informatica No. 4, Vol. 2, 2002 11 AN ACTIVE, OBJECT-ORIENTED MEDIATOR-BASED FRAMEWORK FOR THE DEVELOPMENT OF GIS APPLICATIONS Leonid STOIMENOV, Dragan STOJANOVIĆ, Department of Computer Science, Faculty of Electronic Engineering, University of Niš, Beogradska 14, 18000 Niš, tel. +381(0)18 529-331, E-mail: leni@elfak.ni.ac.yu SUMMARY Most current geographic information systems assume and present the static world. But, new generations of GIS applications have much more demands in comparison to possibilities, which could provide traditional database management systems. Information that exists in the spatial database may be updated over time. Also, the very important demand is time capability to answer to state changing in database. The goals of our research activities are defining an active architecture, identifying the need for mediators in GIS, adding the active rules to the GinisNT spatial database and realize the active mediator level which perform active behaviour of the GinisNT. Paper describes how complete active database semantics can be supported on an existing framework by adding a mediator between the GinisNT spatial database and the client applications. ECA rules are fully supported through the ActiveMediator component without changing applications or the spatial database schema. The Active Mediator prototype provides integration of active and object-oriented features in one system. To provide support for temporal GIS application design and implementation, a suite of spatio-temporal object modelling and management technologies and tools has been developed constituting spatio-temporal object modelling and management system. The openness and scalability of developed frameworks enable their further improvement and refinement to satisfy the needs of specific, temporal GIS applications. Active, temporal GIS would improve understanding of the dynamic geographic processes caused by man, the nature or both, provide methods for detecting and analyzing trends and cycles in geographic phenomena, an would enable prediction of the future geographic states. Keywords: GIS, mediator, active databases, temporal databases. 1. INDRODUCTION Geographic information systems are computerized systems for managing data about spatially referenced objects. GIS differ from other types of information systems in that they manage huge quantities of data, require complex concepts to describe the geometry of objects and specify complex topological relationships between them [11]. In addition, GIS data are typically used by various groups of users with different views and needs. Most current geographic information systems assume and present the static world. But, new generations of GIS applications have much more demands in comparison to possibilities, which could provide traditional database management systems. Information that exists in the spatial database may be updated over time. The very important demand is time capability to answer to state changing in database [1]. To support applications such as GIS, database management systems (DBMS) need to have some support for time. This limitation of current GIS capabilities has recently become the focus of growing research interests within GIS community [2]. Solution for these demands is the active [1,6] and temporal database systems [14,3]. In an active DBMS it is desirable have active rules that can access the time when event occurred. A temporal DBMS usually has extensions to the basic relational operations to support queries that use the time dimension. The growing demand for better understanding and analysis of everlasting geographic changes has emphasized the need for GISs with temporal and active capabilities developed around spatio-temporal database system [9]. Applications in domains so diverse as earth sciences, economical and socio- economical studies, urban planning, traffic control, land management, environment protection, medical imaging, or fleet tracking, can benefit from integration of spatial and temporal information. Temporal GIS could be used to more closely depict the knowledge of the real world, to better explain dynamic geographic processes and phenomena, reveal patterns in their occurrence and explore cause and effect relationships between them. This caused the integration of research work and technology of spatial and temporal databases and resulted in design and development of spatio-temporal data models, database management systems and information systems. Active, temporal GIS would improve understanding of the dynamic geographic processes caused by man, the nature or both, provide methods for detecting and analyzing trends and cycles in geographic phenomena, an would enable prediction of the future geographic states. The research group at the Computer Graphics and GIS Lab at the University of Nis, Yugoslavia, has been developing GIS software for ten years now. One of the research directions pursued has been the