| A reasonable goal for geographic information systems is that they be capable of tracing changes in an area by storing historic and anticipated geographic data. This work provides a conceptual, logical, and physical basis for developing such a capability. It introduces a conceptual model of geographic change that treats only changed data over time, then extends the conceptual model to common geographic data types. The literature of time in information processing is reviewed at length and ways to apply this research to geographic needs are suggested. The discussion extends to such implementation issues as clustering, quality control, and volume control.;The later chapters focus on the problem of responding to ad hoc queries to a spatiotemporal database. The discussion introduces a taxonomy of multidimensional access methods. Four distinct classes are selected from the taxonomy and a representative of each class is chosen to implement on small spatiotemporal datasets. Two of the four methods are found to perform acceptably. |
