AN OBJECT - BASED DYNAMIC SPATIAL MODEL , AND ITS APPLICATION IN THE DEVELOPMENT OF A USER - FRIENDLY DlGlTIZING SYSTEM

A possible set of requirements for a future GIS spatial data structure is that it should have no topological gaps in it at any time, that it should be locally updateable and hence permit real-time updating, editing and querying, that full object insertion and deletion should be supported (i.e. dynamism), that a primitive object on the map should equal one object in the data structure, that objects should be capable of being moved within space (and, by implication, time) while preserving topology, and that the underlying methods need not be restricted to only two dimensions. Such a system would thus be capable of being used interactively, either for data entry and editing, or for “what-if” simulation. As a preliminary platform for the concepts, a user-friendly digitizing system is being implemented that permits a high level of user interaction with the growing map, based on a dynamic spatial data structure the Voronoi diagram. To most users of geographic information systems the underlying spatial data structure linking the parts of the map together, along with the way the system designers imagined “space” to be constructed, appear to be of little direct relevance. Nevertheless, many of the apparently incomprehensible bugs/features of a GIS are directly related to the underlying spatial model used. Despite the fact that there may be possible “work-arounds” of the underlying limitations, this does not excuse a lack of understanding of the properties of a chosen spatial data structure and the underlying perception of space that was used in it (the spatial model). In this paper the importance of the spatial data structure being dynamic is emphasized, using the two-dimensional digitizing process as an example. Further discussion examines the benefits of the dynamic Voronoi spatial model for single-coverage problems, and then touches on multi-layer overlay processes and the extension to higher dimensions.