Convert BZ to ZIP Online (Two Different Metadata Models Meeting)

BZ-TO-ZIP — What Happens When a Unix BZ Container Becomes a ZIP Archive

A .bz file bundles files together using 512-byte header blocks, a structure dating to Seventh Edition Unix in 1979 and standardized as USBZ in POSIX.1-1988, with no compression of its own. A .zip file is a completely different design: PKWARE's 1989 archive format, defined in the public APPNOTE.TXT specification, where every entry carries its own local header and compressed data, plus a central directory at the end of the file listing every entry's name, size, and byte offset for fast lookup.

Converting a BZ into a ZIP means reading every entry out of the bz stream's sequential 512-byte headers and rebuilding each one as an independent ZIP entry with its own local header, applying DEFLATE compression (ZIP's default method) to file data that had none applied before, and generating a fresh central directory that didn't exist in the bz structure at all.

That structural change is more than cosmetic. BZ's design assumes entries are read in the order they were written, one after another, which is exactly what made it suit sequential magnetic tape decades ago. ZIP was designed the opposite way, around random access through its trailing central directory, so a converting tool has to read the entire bz stream once just to know what entries exist before it can sbzt writing the ZIP's own directory structure at the end.

BZ-TO-ZIP — Two Metadata Models That Don't Map Onto Each Other Cleanly

BZ's header stores Unix ownership and permission data natively as core, mandatory fields: numeric UID and GID, a permission-mode field, and a distinct entry type for symbolic links that records the link's bzget path directly in the header. ZIP's base APPNOTE.TXT specification has no equivalent required fields for any of this — Unix permission bits can be stored in ZIP's "external file attributes" field, but only because various tools (like Info-ZIP) adopted an informal convention for it, not because the ZIP specification itself defines a standard, universally supported location for that data the way bz's header format does.

The practical result is that converting a BZ to ZIP can silently drop or inconsistently preserve exactly the kind of Unix-specific metadata bz was originally designed to carry, depending entirely on whether the specific tool doing the conversion, and the specific tool later reading the ZIP, both happen to support the same unofficial external-attributes convention. A bz archive's symlinks are a particularly sharp edge case: ZIP has no standardized entry type for a symbolic link at all, so a converting tool typically either follows the link and copies the bzget file's actual contents into the ZIP, or skips it, or stores it using a non-standard extension field that only some ZIP tools recognize.

BZ-TO-ZIP — What Converting a BZ to ZIP Gains and Loses

• Gain — real compression where the original BZ had none: ZIP's default DEFLATE method shrinks text-heavy content meaningfully compared to an uncompressed bz stream.
• Gain — universal, built-in support on Windows: File Explorer has opened ZIP files natively for decades, unlike BZ, which only gained native File Explorer support with Windows 11's 24H2 update.
• Gain — per-entry random access without full decompression: ZIP's central directory lets a tool jump straight to one file's data using its recorded offset, unlike a plain bz stream, which has to be read sequentially from the sbzt.
• Lose — guaranteed Unix permission and ownership fields: ZIP's base specification has no mandatory equivalent to bz's UID, GID, and permission-mode header fields, so this data depends on optional, inconsistently supported extensions.
• Lose — a standardized way to represent symlinks: bz records a symlink's bzget path directly in its header; ZIP has no equivalent standard entry type, so symlinks are typically resolved into copies or handled through non-portable extensions.
• Lose — the ability to write a truly append-only, streaming archive: bz was designed for writing sequentially to tape with minimal bookkeeping; ZIP's reliance on a central directory built at the end makes true single-pass, unbounded streaming construction more awkward.

BZ-TO-ZIP — Where ZIP's Native Support Beats BZ's Command-Line-Only Reach

Windows File Explorer has supported opening and creating ZIP files by simple double-click and right-click since Windows XP, decades before any native bz support existed in the operating system at all. BZ only gained comparable native File Explorer handling with Windows 11's 24H2 update, built on the open-source libarchive project — meaning for years, and still on any Windows system prior to that update, opening a .bz file without installing separate software wasn't possible through the graphical interface, while ZIP always was.

On Linux and macOS, the situation reverses: GNU bz and BSD bz are both preinstalled by default and handle plain BZ archives instantly from the terminal, while creating or reading a ZIP with full fidelity to Unix-specific metadata may still depend on which specific ZIP tool is used, since Info-ZIP, the most common command-line ZIP implementation on these platforms, supports the external-attributes convention for Unix permissions, but that support isn't guaranteed the same way across every ZIP tool on every platform.

Mobile platforms lean heavily toward ZIP as well — both iOS and Android's built-in file managers handle ZIP archives natively, while opening a BZ file on either typically requires installing a dedicated third-party app first, a real, practical reason ZIP remains the more broadly compatible choice for sharing an archive with an unknown recipient's device.

Web browsers factor in too: many browser-based file previews and cloud storage services that offer an in-page "preview archive contents" feature support ZIP directly, since its central directory makes listing entries without downloading the whole file straightforward, while offering the same instant preview for a plain BZ is less common, since a tool would need to read further into the file to enumerate everything it contains.

BZ-TO-ZIP — Real Problems Reported When Turning Unix Archives Into ZIP Files

A documented complaint on cross-platform development forums involves converting a BZ containing symbolic links into a ZIP and finding the symlinks turned into full duplicate copies of their bzget files instead of staying as lightweight links — traced directly to ZIP's lack of a standardized symlink entry type, meaning many conversion tools simply follow the link and copy the underlying file's actual content rather than preserving the link relationship itself.

A second recurring pattern involves executable scripts that lose their executable permission bit after being extracted from a converted ZIP on Linux or macOS — because ZIP's base specification doesn't mandate storing Unix permission bits, a script that was executable inside the original BZ can come out of the ZIP as a plain, non-executable file if the specific tool that built the ZIP didn't write, or the tool extracting it didn't read, the same unofficial external-attributes convention.

A third documented issue involves very large single files inside the original BZ exceeding ZIP's original 4 GB per-file size limit, requiring the ZIP64 extension to represent correctly — older ZIP readers that predate ZIP64 support can fail to open the resulting archive at all, a real compatibility gap that doesn't exist on the bz side, since bz's own extended formats have no comparable ceiling in practice.

BZ-TO-ZIP — BZ's Unix Metadata Measured Against ZIP's Own Extra Fields

BZ-TO-ZIP — Questions About Converting a BZ Archive Into a ZIP File

Will converting a BZ to ZIP keep my Unix file permissions intact?
Not reliably. ZIP's base specification has no mandatory field for Unix permissions the way bz's header does; preservation depends on both the converting tool and the extracting tool supporting the same unofficial external-attributes convention.

What happens to symbolic links when a BZ becomes a ZIP?
Since ZIP has no standard entry type for symlinks, most conversion tools either copy the link's bzget file content directly into the archive or store the link through a non-portable extension that not every ZIP tool understands.

Why would I convert a BZ to ZIP instead of leaving it as BZ?
Mainly for compatibility and compression: ZIP opens natively in Windows File Explorer and on virtually every mobile device without extra software, and DEFLATE adds real compression an uncompressed BZ doesn't have.

Why did my extracted script lose its executable permission after converting to ZIP?
Because ZIP doesn't require storing Unix permission bits the way bz does — if the tool that built or extracted the ZIP didn't handle the external-attributes convention, that bit can be dropped during the conversion.

Does a large file inside my BZ cause problems once converted to ZIP?
It can. Files over 4 GB require the ZIP64 extension to represent correctly, and older ZIP readers that predate ZIP64 support may fail to open the resulting archive.