In vector databases as well as raster databases, the quantity of data becomes overwhelming. Practical considerations dictate that the data must be partitioned and considered in smaller chunks. People who converted to GIS and computer cartography systems from paper maps were accustomed to dealing with map sheets, quadrants, and tiles. Other users, however, demand a seamless structure-maybe even one that allows a wide range of levels of detail. Debates rage about what data structure is best for the partitioning: uniform grid, quadtrees, R-trees, binary space partitions, BANG files.