A legend from the compression arms race

Every field has a name that keeps coming back in old forum threads, and in file compression that name is FreeArc. Built by Bulat Ziganshin, it took one idea further than anyone else at the time: do not pick a single algorithm for the whole archive, pick one that suits what is actually inside. Plain text went to PPMd, executables were preprocessed with a BCJ filter so that call addresses stopped fighting the compressor, and audio and image data got filters written for their particular statistics. Stack those decisions and the resulting ratio landed in the top group of the comparison tests published back then.

The engineering around the compressor was just as deliberate. Solid blocks let similar files be compressed as one stream, so a folder of near-duplicate documents stops paying the same overhead dozens of times. Smart updating meant a new archive could reuse the compressed data of unchanged files instead of redoing the work. Recovery records borrowed from the Reed-Solomon family could repair a damaged archive instead of forcing a full re-download. For an era when archives travelled on scratched optical discs and flaky drives, that last feature was not decoration.

It is worth being precise about what this class of tool is for. Archivers are not only about saving space; they are about a file surviving the trip. Ratio, speed, and recoverability form a triangle, and FreeArc leaned hard toward ratio and recoverability while accepting that the clock would suffer on large inputs.

Encryption and open source, the old-fashioned thorough way

Security configuration was unusually generous. AES, Twofish and Serpent could be chained, one after another, so a password was not protecting a single layer but a stack of them. Very few archiving tools of that generation offered anything comparable, and the habit of letting the user decide how paranoid to be is a distinctly old-school virtue. The project shipped under the GPL, ran on both Windows and Linux, and offered a command line alongside a graphical front end for people who preferred menus to switches.

Then the story took its quiet turn. After releasing version 0.67 in 2014, the author moved his attention to other compression work, and the ideas from that research later fed into Zstandard. FreeArc was never declared dead; it simply stopped moving. The practical consequence is honest and unambiguous: it is no longer the tool at the front of the update queue, but the archives it produced during its peak still open perfectly today, which is the only promise an archiver really has to keep.

FreeArc is a footnote from the golden age of compression algorithms: take one goal to its limit, then bow out with grace. Its reasoning lives on inside every modern compressor you use.

Where it still earns a place

For new archives and everyday work, the sensible choice remains a tool that is actively maintained, 7-Zip being the usual default. FreeArc earns its space in two narrower situations. The first is extraction: old archives created during its peak years may use algorithm combinations that newer tools handle poorly, and the original is the safest key. The second is study. If you want to understand why per-file-type selection beats a single global setting, its options screen is a compact lesson.

It is free and open source, the project page is still standing, and historical builds remain reachable there. Treat it as a reference volume on the shelf rather than the daily driver, and it will keep paying you back the way good old software does.