Dynamic generalisation from single detailed database to support web based interaction

It is believed that Web-based visualisation of spatial information could be greatly enhanced through the use of dynamic or 'onthe-fly' map generalisation. This involves dynamically deriving scaleand theme-dependent displays from a single detailed dataset, thus eliminating the need to maintain duplicate datasets at different resolutions. Dynamic generalisation also prevents the generation of cartographically-poor maps that result from the display of spatial information at a considerably different resolution from that at which it was captured. The implementation involves combining automated generalisation and symbolisation techniques together in a simple rule-base to produce a virtual map of central Edinburgh in which both the level of detail and the symbology is dynamically tailored to a user-specified scale and map theme (tourist or topographic maps). The paper explores the inherent differences between the generalisation of a virtual map and that of a paper map, and also considers issues such as those arising from the relationship between map scale, map theme and utility. The research demonstrates that it is possible to achieve effective results using a limited number of simple mechanisms, and argues that the true potential of dynamic generalisation lies in two main areas firstly, in an ability to narrowly-define the map theme, thus enhancing map clarity and more closely matching user requirements; and secondly, in using control over the map scale as a gateway to other types of map use and other themes.