Data fission is the data management equivalent of splitting the atom. Unfortunately, whereas atom splitting can generate significant amounts of energy, data fission often loses significant amounts of synergy by making the sum of the parts less than the whole.
For example, we can take a hypothetical (but not VERY hypothetical) data stream from the NOAA Polar Orbiting Environmental Satellites. Ancillary data, such as buoy-based sea near-surface temperatures, may also be brought into the processing stream. The satellite imgry may receive geometric and radiometric corrections. Corrections may be made to the buoy data to make them approximate skin temperatures measured by the satellites. Additional computation produces Sea Surface Temperatures, atmospheric aerosols, global-scale Normalized Difference Vegetation Index, and perhaps other products. Each of these products is produced primarily giving consideration to its own user community, perhaps downplaying optimization for integrated use. The data may be produced on different projections, or at different resolutions, or with different output file formats. The data are "separated at birth" for storage in different locations.
In the past, when specialists attempted to pull the compromises in satellite design to their benefit (if necessary at the cost of another application), attempting to optimize the processing of, say, the Sea Surface Temperature, made sense. Now, however, when we attempt to combine Sea Surface Temperature, atmospheric aerosols, and Normalized Difference Vegetation Index for a particular study of El Niño, we are challenged by sometimes incompatible time periods, projections, resolutions, grid referencing, etc.
Thus, the potential attraction of data fusion.
Submit your favorite definition or example(s) of data fission! Send us your favorite link to discussion on data fission! Comments are welcome:
URL: http://www.ngdc.noaa.gov/seg/globsys/gis/fission.shtml
Revised: 21 December 1997