This seminar aims to bring together researchers from the areas of computational geometry, automated cartography, and geographic information systems (GIS), spatial data acquisition and image analysis and practitioners from institutions and industries that deal with spatial information and GIS. It will be the second Dagstuhl seminar with this title, further forming the rather young interdisciplinary community.

Two of the most important capabilities of GIS are the visualization of geographical information and the quantitative analysis of spatial data. This seminar will focus on the former issue of (cartographic) visualization and the relevant computations. Since visualization of geoinformation generally concerns large quantities of data, efficient algorithms to handle the data are essential. Furthermore, a significant number of manual techniques of traditional cartography are still awaiting adequate modeling and formalization in order to be translated into algorithms and automated procedures useful for GIS and automated map production.

Main topics are:

• Map generalization:
reducing the detail of a map to fit a target scale and map purpose
• Terrain modeling and visualization:
representing, analyzing, and visualizing continuous data defined over the plane
• The fusion of spatial data and image data
• to support image interpretation and to enhance terrain visualization
• Dynamic and animated maps:
showing changing maps and maps of change on screens, solving the necessary representations and computations
• Map labeling:
automatically placing text at the appropriate place of maps to label map elements.

Other important issues directly or indirectly related to computational cartography are:

• Feature and object extraction from images
• Image and data fusion
• Image sequencing for process monitoring
• Spatial data analysis:
includes combination of themes through overlay or spatial statistics, network analysis, and spatial interpolation methods
• Automated rectification of digitized data:
checking for and correcting error in acquired data sets.
• Error representation, error propagation, and quality control
• Data structures for representation of spatial data
• as well as the relations and hierarchies contained in them.

The technological advances of the recent past -- steadily increasing graphics capabilities, Image analysis, multimedia technology, distributed computing and the Internet, to name but a few -- have lead to an increasing importance of visualization and cartographic products not only in conjunction with GIS, but also in everyday life applications, such as access to indicator and locator maps via the Internet (so-called Internet map servers). Despite the growing importance of such cartographic products, however, the further spread of maps is impeded by the fact that digital products in most cases still cannot be produced automatically with high cartographic quality (largely for lack of suitable algorithms). This implies that digital map products today either have to be tediously 'polished' interactively (i.e. by hand) or sub-optimal cartographic quality is reached by limited automated procedures. Since maps are intended to be communication tools, low quality maps can be rather useless or even misleading if the 'message' they contain cannot be properly read.

The above main topics for the seminar have been selected because they are representative of cartographic techniques and are of central importance to providing cartographic functionality to users of GIS as well as other systems working with spatial data. Common to all of these problems is that they are complex, resisting straight-forward formalization. Furthermore, solving these problems requires that principles established in traditional cartography are observed in order to maintain the desired level of quality of the graphical output. This situation, therefore, clearly calls for a collaborative effort bringing together algorithm developers and cartographic experts, with the aim to develop efficient algorithms which also achieve high cartographic quality.