November 28, 2025
Merge wars meet history buffs
A Repository with 44 Years of Unix Evolution
44 Years of Unix in One Repo—Fans Cheer, Merge Math Sparks Drama
TLDR: A single GitHub repository reconstructs 44 years of Unix history with 659k commits. Commenters are wowed but bicker over low merge counts and which versions were included, while the author drops free links—making this a treasure for history geeks and fresh fuel for merge-count debates.
History nerds, start your forks: a 1GB “time machine” of Unix from 1972 to 2015 just dropped on GitHub, stitching 24 preserved snapshots into one mega-repository with 659k commits and 850 named contributors. The crowd’s first reaction? Awe, followed by a fierce audit of what made the cut. As werdl put it, “the decisions… regarding which flavours… to include” became the main sport, and aap_ chimed in with the fan chant: “Where’s UNIX v4?” To many, it’s software archaeology you can clone, a rare chance to see how a five-thousand-line kernel grew into a 26-million-line giant.
Then came the merge math drama. phplovesong did a double-take at “670,000 commits” vs “only 2K merges,” joking that maybe everyone just “pushed straight to master.” Cue hot takes: trunk-based before it was cool, or the inevitable quirks of stitching old histories together. In a classy plot twist, the project’s author DSpinellis parachuted in with a paywall-free paper and the actual repo links: read the study here and poke the code here. Verdict from the comments: epic exhibit, spicy curation debates, and merge-count side-eye—the perfect HN cocktail.
Key Points
- •A 1GB synthetic Git repository on GitHub reconstructs Unix’s evolution from 1972 to 2015.
- •The dataset aggregates 24 snapshots and multiple repositories from Bell Labs, UC Berkeley/CSRG, the 386BSD team, two legacy repositories, and the modern FreeBSD repository.
- •It contains about 659,000 commits, 2,306 merges, and ~850 identified contributors (23 Bell Labs, 158 CSRG, 660 FreeBSD).
- •Caldera International’s liberal licensing of early Unix components enabled inclusion of key historical material.
- •Repository structure includes an Epoch tag and Research-VX tags for six Bell Labs research editions, beginning with Research-V1 (4,768 lines of PDP-11 assembly).