Earlier this month, I spent a day working in the throwback world of DOS. More specifically, it was FreeDOS version 1.1, the open source version of the long-defunct Microsoft MS-DOS operating system. It’s a platform that in the minds of many should’ve died a long time ago. But after 20 years, a few dozen core developers and a broader, much larger contributor community continue furthering the FreeDOS project by gradually adding utilities, accessories, compilers, and open-source applications.
All this labor of love begs one question: why? What is it about a single-tasking command-line driven operating system—one that is barely up to the most basic of network-driven tasks—that has kept people’s talents engaged for two decades? Haven’t most developers abandoned it for Windows (or, tragically, for IBM OS/2)? Who still uses DOS, and for what?
Old school: I work in DOS for an entire day
Open source MS-DOS alternative lives—but using it nearly killed me.
To find out, Ars reached out to two members of the FreeDOS core development team to learn more about who was behind this seemingly quixotic quest. These devs choose to keep an open-source DOS alive rather than working on something similar but more modern—like Linux. So, needless to say, the answers we got weren’t necessarily expected.
Doing very little, very well
Jim Hall kicked off the FreeDOS project 20 years ago while he was an undergraduate studying Physics. Hall is now the IT Director at the University of Minnesota-Morris, and he’s still working to keep the DOS prompt blinking. Hall just returned to the project after completing a second master’s degree program, and he says that group has between 30 and 50 active developers involved. (That’s down from the hundreds who were active while pushing FreeDOS to its 1.0 release in 2006.)
While it’s a small group, this is no rag-tag assemblage of DOS holdouts. Many of them do develop for Linux and other operating systems, work for commercial software vendors, or hold other technical jobs. They largely contribute to FreeDOS for the intellectual challenge. Pasquale “Pat” Villani, the man who contributed the kernel to FreeDOS, was for many years a lead software engineer at Digital Equipment Corp., Compaq, and then HP, working on various Unix operating systems.
Jim Hall kicked off the FreeDOS project 20 years ago, when he was an undergrad. Now he’s a university IT director.
University of Minnesota-Morris
FreeDOS has a very loyal core constituency. Eric Auer, a long-time core contributor to FreeDOS, explained in an e-mail why he thinks DOS is still relevant. “It is small enough to get an idea of the inner workings,” he said. Auer said FreeDOS now has support for developers who want to do full-blown 32-bit applications thanks to the DJGPP toolkit, a port of the GNU Compiler Collection to FreeDOS. This allows developers to build monolithic applications that use all of the computing power of the machine they run on without the overhead of more complex operating systems. Also, Auer said, “It runs games that I knew 20 years ago.”
Because FreeDOS is, as some have called it, “barely an operating system,” it allows developers to get very, very close to the hardware. Most modern operating systems have been built specifically to avoid this for security and stability reasons. But FreeDOS has become much more friendly to being virtualized or running in hardware emulation, such as the Linux DOSEMU emulator
The direction the project has taken hasn’t exactly followed the road map Hall envisioned after version 1.0. He once had ambitious plans for a next-generation of DOS, originally envisioning a modern FreeDOS along the lines of an alternative history of computing. “For a while, I was thinking, ‘If MS DOS survived, where would DOS have gone in the last 10 to 15 years?’” Hall said. “I was advocating some sort of multitasking—we could have task switching like what was supported in the 286, where you can put one process to sleep while you do another process. I wanted to have TCP/IP added to kernel.”