SMALS: a novel database for two-dimensional object location

A database called SMALS, for storage minimizing automatic level sifting, that minimizes memory requirements and query response time for object location in 2-space is discussed. Memory usage is minimized through object insertion and deletion algorithms that allocate and deallocate the database memory dynamically, obviating the need for garbage collection. A fixed-level multidensity mesh representation and its associated sifting algorithm efficiently map the number of database nodes (grid boundary intersections) in which an object is stored. In contrast to the case for conventional tree structures, the object query time is well bounded and independent of object density in a region, making SMALS ideal for use in large interactive CAD (computer-aided-design) systems.<
>