[dehrmann]

Huge caveat to this: systems like Hadoop accomplish the parallelization by adding keys, and those FP constructs don't broadly have this. It's better to think of it as parallel divide-and-conquer.