Map Generalization for On-demand Web Mapping

Maps, owing to their graphic nature, are used increasingly in many applications to visualize complex spatial concepts. This is particularly true for the WWW which emphasizes graphic communication. Though maps are in wide-spread use in web mapping, however, capabilities for adapting the visualized data and maps to the display scale (e.g., as a result of zooming operations) are still rather poorly developed. Most systems that provide scale-dependent web mapping use a series of raster maps (e.g., GIF or JPEG images) at multiple, fixed scale levels that are served depending on the zoom level. For future on-demand web mapping applications, however, more flexible and progressive ways of transmitting and displaying vector (not raster) data across the network are needed. To date, only few papers have addressed the issue of transmitting vector data at progressive levels of detail. One of the notable exceptions is the paper by Bertolotto and Egenhofer (1999) which discusses options for structuring multiple representations of vector data in light of progressive transmission.