The author describes the compressor in the documentation:
Some words about base lexicographic ordering: at this stage of development of 'ecp' I use bzip algorithm without quadrants, but later 'ecp' will include one more new algorithm(s), as different keys.About sorting:
I use preprocessor, based on LZP54. It works very fine on 'water' data, as RLE does not. And also lzp let improve speed of lexicographic ordering.
The algorithm in this module is almost identical to used in bzip by Julian Seward. The next version of ecp will include some original algorithm and probably reprocessed lexicographic ordering, found in bzip.About LZP:
In this module you can find preprocessor algo, based on LZP by Charles Bloom. This preprocessor especially useful on so called 'water' data, and also it can help improve BWT speed.the program source code
It is not clear what the
water data refers to.
The compressor is featured in the BWT comparison.
(Have info that should be added here? E-mail.)
Using custom MinGW32 3.4 compiled binary, see the linked package.