Abstract
The leap from an atemporal to a temporal geographical information system involves issues that can be solved only by understanding precisely what information a particular organization wishes to represent. Uncertainty, incomplete information and budgetary realities combine to present a spatiotemporal system's implementors with difficult choices. No right or wrong approach exists in these cases; rather, the solutions are based on what information is required and how an organization chooses to maintain it. The major issues involve representation, updates and longevity. Concerning representation, a system can use timestamps to describe the last time a feature was checked for correctness, or the last time the feature changed; feature histories can be traced in aggregate or individually; retrieval of past states may require some generalization over time; and many features need not be tracked temporally. Concerning incremental updates, set–up time and contention must be minimized; continual updates of geometry produce artifacts and edge effects; and incremental dissemination of updates requires careful coordination. Finally, because temporal systems are intended for posterity, ensuring that data will be useful to future generations is paramount.

This publication has 3 references indexed in Scilit: