This is the simplest method to do threading, which splits
the uncompressed data into blocks and compresses them
independently from each other. There's room for improvement
especially to reduce the memory usage, but nevertheless,
this is a good start.
Lasse Collin