Sustainability of Digital Formats: Planning for Library of Congress Collections

The bzip2 file format is a freely available, patent-free data compression program created by Julian Seward. It is both the name of the format and the program used to create it. The program is designed for compressing single files only. It was created as a successor to its predecessor, bzip, to avoid potential patent issues. See: History for more information.

Different versions of bzip2 maintain file format compatibility. Newer versions can work with files created by older versions, ensuring a level of stability. However, the format creator acknowledges limitations in the compressed file format and provides source code for decompressing older files created by bzip-0.21.

A bz2 stream consists of a 4-byte header, followed by zero or more compressed blocks. An end-of-stream marker contains a 32-bit CRC for the plaintext whole stream processed. The compressed blocks are bit-aligned, and no padding occurs.

Two unofficial documentation resources are commonly cited.

• " Not an Official Specification" by Joe Tsai.
• Repository for the bzip2 program on GitLab maintained by Micah Snyder.

Widely adopted.

The bzip2 file format “ships standard on many Unix/Linux systems.”

Often compared to gzip and ZIP File Format (PKWARE).

The bzip2 homepage states that the license is a GNU’s Not Unix (GNU) General Public License (GPL). It is unclear which version of GNU GPL would apply. Other sources state conflicting information about bzip2’s license, stating it is a Berkeley Software Distribution (BSD) style license. Comments welcome.

Identifies self as a bzip2-compressed file with magic numbers (see magic numbers section). There is no specific language for the inclusion of other metadata. However, documentation is sparse. Comments welcome.

Accessibility Features

No specific features in the file format. Features to support accessibility would be found in the bundled and compressed files (such as embedded captions and subtitles in audiovisual content, tagged and structured text in textual documents, and alt text for images). Aggregate files can also contain separate files for transcripts, timed text or captions as part of the bundled package. See Relationships to other formats for details.

For more details see:

• " Not an Official Specification," by Joe Tsai.
• “ BZip 2,” a poster by Ange Albertini.
• Gary Kessler's File Signature Table.

Note this header, when converted from Hexadecimal to ASCII, is "BZh". “BZ” stands for “bzip”, and the "h" is for "Huffman coding," the compression algorithm used with bzip2. Some sources, such as Wikipedia, will cite the magic numbers as “BZh” instead of the hexadecimal.

The bzip2 program, and by extension the bzip2 file format, is based on its predecessor bzip. Despite similarities in appearance and name to bzip, bzip2 is rewritten and re-engineered. It was developed to address potential patent issues with bzip. The format created by bzip2 is not compatible with bzip, and efforts to make them compatible were avoided to maintain the purpose of patent avoidance. Seward expressed commitment to backwards compatibility for future changes. The predecessor program bzip is no longer available.

Julian Seward released bzip2, version 0.15, in July 1996. The compressor’s popularity grew over the next several years due to its stability.

Julian Seward released version 1.0 in late 2000.

In June 2019 Federico Mena became the new maintainer of bzip2.

In 2019, Mark Wielaard began maintaining a bzip2 stable repository at Sourceware. In June 2021 Micah Snyder became the new maintainer of the Sourceware repository.