本文由 AI 分析生成
建立時間： 2026-03-28 來源： https://blog.brachiosoft.com/posts/git/
Summary
A detailed history of Git’s origins from the “Software Archaeologist” blog. The story: Linus Torvalds had become a scaling bottleneck for Linux kernel development (“Linus doesn’t scale”), which led to adopting BitKeeper — then a controversial crisis (Andrew Tridgell reverse-engineered the protocol, triggering BitKeeper’s withdrawal) forced Linus to write Git in two weeks.
「軟體考古學家」部落格對 Git 起源的詳細歷史。故事是：Linus Torvalds 已成為 Linux 核心開發的規模化瓶頸（「Linus 無法擴展」），這導致採用了 BitKeeper——然後一場爭議危機（Andrew Tridgell 逆向工程了協議，觸發 BitKeeper 撤銷授權）迫使 Linus 在兩週內寫出了 Git。
Key Points
- 1998: Linus was the sole gatekeeper for Linux changes — everyone emailed him patches; he became the bottleneck
- Larry McVoy: Sun engineer who built NSElite (Perl) → saw it replaced by TeamWare → developed BitKeeper (first modern distributed VCS)
- BitKeeper was free for open-source use, which allowed Linux kernel to adopt it
- The crisis: Andrew Tridgell (rsync, Samba) tried to reverse-engineer the BitKeeper protocol → Larry McVoy withdrew the free license in anger
- Linus wrote Git in ~2 weeks (April 2005) to replace BitKeeper, informed by what he’d learned using it
- BitKeeper/TeamWare lineage: distributed VCS concepts (local commits, clone/pull/push) originated at Sun in the early 1990s
Insights
The most striking detail: the distributed VCS concept (local commits, pull/push) originated at Sun in the early 1990s via TeamWare — over a decade before Git. The Tridgell incident is a perfect example of how technical communities’ values (open protocols, reverse engineering as legitimate practice) can clash catastrophically with commercial license models, and how that tension ultimately produced Git — the most important open-source tool of the era.
Connections
Raw Excerpt
問題癥結在於 Linus 擴展不了（“Linus doesn’t scale”）。我們不能期待 Linus 一個人能跟上 Kernel 變化的速度，我們也不想看到 Linus 失去對 Kernel 的掌控。