Speculations

33.3 Speculations

Speculation about longer-term scientific data access infrastructure.

I/O will become even more of a bottleneck for high-performance models
Observational data will be generated too fast to store, so filtering it will become as important as storing it
Access to compressed lower-resolution data from multi-resolution represenations such as discrete wavelets will speed up I/O of huge datasets
Format obsolescence will not be an issue, with new software supporting old formats
Semantic web technologies will add more meaning to online data
Higher-level data access, such as use of coordinates instead of index ranges, will dominate
More valuable data will be lost for lack of organizational support for data stewardship and preservation
Format-independent data conventions will evolve incrementally, collaboratively, and too slowly
...