Practical compressor: Simple code, efficient both on speed and compression ratio!

Latest development version on https://github.com/richox/comprox

Also check out another lossless compressor of mine: slimox

Comprox is a generic data compression program, with command-line based UI, implementing static dictionary coding, LZ77, and arithmetic coding. It's time performence is similar to Bzip2, but with a better compression ratio (especially for large text files).

Comprolz is a fork to comprox, which implements ROLZ instead of LZ77 matching algorithm. It provides better compression ratio but costs more time on compressing and decompressing.

Comprop is another fork to comprox, using LZP instead of LZ77. It works best on data with high redundancy (e.g. fp.log).

All of the three compressors are experimental still under developing. Think before using them for your important data.

Here is a Qt4 GUI front-end for comprox.

Some benchmark for English text files: (auto generated by comprox/benchmark/Makefile) ``` Benchmark with world95.txt (original size = 3005020): encode decode compressed size gzip: 0.458s 0.056s 878236 bzip2: 1.373s 0.519s 578797 xz: 4.856s 0.163s 577232 comprox: 1.115s 0.553s 546580 comprolz: 2.224s 1.158s 524943 comprop: 1.164s 1.397s 546639

Benchmark with bible.txt (original size = 4047392): encode decode compressed size gzip: 0.901s 0.091s 1191061 bzip2: 1.672s 0.680s 845635 xz: 8.160s 0.228s 885184 comprox: 1.297s 0.753s 768724 comprolz: 3.209s 1.348s 757862 comprop: 1.412s 1.578s 759474

Benchmark with enwik7.txt (original size = 10000000): encode decode compressed size gzip: 1.904s 0.217s 3693800 bzip2: 4.011s 1.748s 2916026 xz: 20.874s 0.678s 2723036 comprox: 4.015s 2.388s 2514558 comprolz: 7.942s 4.132s 2471818 comprop: 3.891s 4.687s 2505889 ```

Large Text Compression Benchmark (Thanks to Matt Mahoney)