bzip2 (BZ2) is a free and open source lossless data compression algorithm and program developed by Julian Seward. Seward made the first public release of bzip2, version 0.15, in July 1996. The compressor's stability and popularity grew over the next several years, and Seward released version 1.0 in late 2000
bzip2 compresses most files more effectively than more traditional gzip or ZIP but is slower. In this manner it is fairly similar to other recent-generation compression algorithms. Unlike other formats such as RAR or ZIP (and similar to gzip), bzip2 is only a data compressor, not an archiver. The program itself has no facilities for multiple files, encryption or archive-splitting, in the UNIX tradition instead relying on separate external utilities such as tar and GnuPG for these tasks.
As stated in the front page of the bzip2 Web site, in most cases bzip2 is surpassed by LZMA and PPM algorithms in terms of absolute compression efficiency. According to the author, bzip2 gets within ten to fifteen percent of PPM, while being roughly twice as fast at compression and six times faster at decompression.
bzip2 uses the Burrows-Wheeler transform to convert frequently recurring character sequences into strings of identical letters, and then applies a move-to-front transform and finally Huffman coding. In bzip2 the blocks are generally all the same size in plaintext, which can be selected by a command-line argument between 100 kB–900 kB. Compression blocks are delimited by a 48-bit sequence (magic number) derived from the binary-coded decimal representation of π, 0x314159265359, with the end-of-stream similarly delimited by a value representing sqrt(π), 0x177245385090.
Originally, bzip2's ancestor bzip used arithmetic coding after the blocksort; this was discontinued because of the patent restriction to be replaced by the Huffman coding currently used in bzip2.
BZip2 is known to be extremely slow at compressing, making people opt for alternatives such as gzip or rar when time is an issue. This problem is asymmetric, as decompression is relatively fast. Motivated by the large CPU time required for compression, a modified version was created in 2003 that supported multi-threading, giving significant speed improvements on multi-cpu and multi-core computers. Unfortunately as of 2007 this functionality has still not been incorporated into the main project. No reason has been given by the maintainer.
