Microsoft Opens the Vault: PC-DOS 1.0 Source Code Goes Public

Microsoft has made the source code for its inaugural operating system, PC-DOS 1.0, publicly available, offering historians and developers a rare window into the foundational era of personal computing. The release traces back to 1980, when IBM approached Bill Gates seeking a software partner to power its upcoming personal computer. At the time, Gates had previously focused on BASIC interpreters, and his company had already shipped Xenix, a Unix distribution. Because AT&T licensing restrictions prevented Microsoft from porting Xenix to the x86 architecture, the company pivoted to an alternative solution. Microsoft purchased 86-DOS, also known as QDOS, from Seattle Computer Products and its creator, Tim Patterson, for approximately $100,000. The firm adapted the software’s CP/M-style application programming interfaces and delivered it to IBM as PC-DOS 1.0 in August 1981. Crucially, the agreement permitted Microsoft to market the software independently as MS-DOS to other hardware manufacturers, a strategic move that cemented its long-term industry dominance.

Before this latest publication, the earliest accessible DOS code consisted of MS-DOS 1.25 and 2.0. Microsoft originally shared those files through the Computer History Museum in 2014, followed by a GitHub publication in 2018. The initial 2014 distribution carried strict limitations, permitting only non-commercial academic and research use while prohibiting derivative projects. The subsequent GitHub upload shifted to the MIT license, a permissive framework compatible with the GNU General Public License that encourages modification and redistribution. This latest publication extends that open framework back to the very beginning of the PC timeline. The code is now hosted as a navigable Git repository, enabling developers, educators, and enthusiasts to compile and test the software using modern development tools.

According to Microsoft, the publication extends beyond standard operating system files. The company explained that the archive contains point-in-time working states and handwritten annotations maintained by Patterson, which the firm described as resembling a “printed commit history of a Git repository.” The collection features the 86-DOS 1.00 kernel, multiple development snapshots of the PC-DOS 1.00 kernel, and early utilities like CHKDSK. Notably, it also includes the assembler listings themselves, providing an unprecedented look at the actual development environment of the era. Running on first-generation 8086 hardware, the compact codebase remains highly accessible for study, offering a stark contrast to the complexity of modern operating systems while preserving a critical chapter in technology history.

The release also addresses persistent historical confusion regarding early DOS iterations. No commercial product ever carried the exact name MS-DOS 1.0, forcing historians to reconcile IBM’s PC-DOS 1.0, Microsoft’s internal tracking numbers, and OEM variants such as MS-DOS 1.25. By publishing a clearly labeled DOS 1.0 codebase tied directly to the original IBM PC launch, researchers now have a definitive reference point. The original software operated on 160KB floppy disks, lacked hard drive support, and contained no subdirectories, yet it established the architecture that would govern personal computing throughout the 1980s and into the early 1990s. The open release ensures that this foundational code remains a living educational resource rather than a static archive.