Space-Time is the Key: For Expressing Spatial-Temporal Computing
2019-09-18T03:37:22Z (GMT) by
This paper outlines a new spatial-temporal programming model. The model builds on key-value programming models such as MapReduce by enabling bounded space and time to be keys and a collection of spatial-temporal data to be values. This reconceptualization of key-value programming exploits spatial-temporal characteristics in data to facilitate the parallelization of spatial methods and models. This paper outlines the new space-time key-collection programming model and a proof-of-concept implementation For Expressing Spatial-Temporal computation in parallel,
called ForEST. Three use cases for the programming model and ForEST language are outlined. First, as a platform to advance research in geospatial computing and algorithm development for spatial problems. Second, as a teaching tool to help learners understand the complexities
around handling data, expressing computation, and executing code in spatial-temporal applications. Third, as a language to help process, mine, and analyze spatial-temporal data in a number of fields including GeoComputation, Geographic Information Science, and Spatial Data Science.