Many of the well-known impediments to automating map generalization derive at least in part from a poverty of information regarding roles of specific locations in delineating features. Although the current paradigm features�=�coordinates�+�topology�+�attributes�+�metadata often works well, it is still possible to improve on it. In particular, geometric descriptors of features (coordinates) have distinct limitations when changing map scale and filtering cartographic details; it is often necessary to refer to attributes and metadata to support such operations. However, this leads to difficulties when attempting to transfer cartographic databases between dissimilar systems, and to complicated computational strategies that could be avoided if better notations for geometric data were employed. This paper describes a notation for location that offers scale-specific positional encoding, as well as describing multiple characteristics of each vertex along every vector in a cartographic database. The method encodes locations as addresses in eight spherical quadtrees describing coordinates, attributes and metadata. Addresses are stored as 64-bit bitfields, capable of capturing point locations at a hierarchy of map scales down to less than 1:20, usually with room left over to encode other spatial properties. An encoding convention for this notation is described, and suggestions are given for how it might be used to convey qualitative and quantitative aspects of map data that can be applied across a range of cartographic scales. An approach to detecting overlapping features using hierarchical nodal regions is also described.