At the beginning it was just for fun, but after some research and some "googling" over the Internet I came up with many new ideas. Soon after, my program became comparable to bzip2 and WinAce. The execution time can be improved very much since the algorithm I used to make the transform is not one of the best known.
If you are interested here you can download the sources (as a gzip archive) using one of links below. The program runs under Linux (it was not heavily tested). To compile it, a simple "make" is enough. Be aware that it implements some algorithms (eg. Arithmetic Coding) that are patented.
Chile is benchmarked on different websites. Here are some of them:
- http://www.cs.fit.edu/~mmahoney/compression/text.html
- http://www.maximumcompression.com
- http://uclc.info/
Use one of the links below to download the program. Version 0.5 is the latest. While latest version is faster or has better compression on most cases, I recommend 0.3d because suffix sorting is done using an algorithm with worst case O(n*log(n)) and it was tested much better.
