The Subversion project’s own book dates the start of the project to early 2000, when the company CollabNet began “seeking developers to write a replacement for CVS.” This entry uses 2000 as the year of the project’s beginning; the book records design work commencing in May 2000, while the first 1.0 release came later, in 2004.
The motivation was direct. According to the book, CVS’s “limitations were obvious from the beginning,” and CollabNet set out to build a system that retained “the basic ideas of CVS, but without the bugs and misfeatures.” Jim Blandy had already conceived the Subversion name and the basic design of its data store before the company’s effort formally began.
The book records an early milestone in the project’s own development: on 31 August 2001 Subversion became self-hosting, meaning its developers stopped using CVS to manage Subversion’s source code and switched to Subversion itself. That moment marked the tool crossing from prototype to usable system.
The official Apache Subversion site confirms the founding by CollabNet in 2000 and the project’s later move to the Apache Software Foundation. The release of Subversion was a milestone in version control because it gave the open-source world a centralized tool that fixed the long-standing frustrations of CVS, and it dominated the centralized era of the 2000s.