I'm not sure. Currently the compression variables are somewhat inefficient. I've compressed similar files down to about 17mb using hardcoded variables but I've yet to stumble across one that will reliably compress a variety of data. I was hoping to test it enough to come across patterns that would enable me to develop heuristics, changing the variables according to different input file characteristics. This is simply extremely time consuming. I have a non-open version of a listener for server applications that I've sort of moved my testing over to since I created the client compressor/extractor above.