In the late 1980s and into the 1990s, I considered myself a power user of the PC operating system MS-DOS. DOS was a modest system, running only one task at a time, and interacting via the command line to launch applications or simple utilities.
As an undergraduate physics student, I relied on DOS to do much of my work. And like many other DOS power users at the time, I wrote my own tools and utilities to expand the features of the DOS command line.
Aside from the tools I created myself, every DOS application I relied on was proprietary or "closed source software." Although at the time, we didn't have the terms "free software" or "open source software," everything was just "software." The normal way to obtain and run software was to buy it in a store. Yes, that could get expensive, but we didn't have other options. I ran the typical gamut of software for the era, including WordPerfect to write papers for English classes, Lotus 1-2-3 to analyze lab data for physics classes, and ProComm to dial into the university's network to use the campus Unix systems for certain coursework.
Then I discovered a new software paradigm: shareware.
Shareware would let you try a program before you bought it. Usually, you could use the program for about a month, after which you "registered" the program by sending a check or money order to the author. Even better, shareware encouraged you to share the programs with your friends so they could try it too.
Most shareware programs were pretty inexpensive, and there was a variety of high-quality programs available under the shareware model. I discovered a shareware word processor called Galaxy Write that displaced WordPerfect, a shareware spreadsheet program called As Easy As that replaced Lotus 1-2-3, and a shareware modem dialer called Telix that did the job of ProComm. These programs gave me all the functionality for a fraction of the cost of typical off-the-shelf software.
By Spring 1993, I was ready to try something new. I loved the power of DOS, but it ran only one task at a time, while the big Unix systems in our campus computer lab could support multiple programs running at the same time. Using those big Unix systems had spoiled me. Also, I admired the flexibility of the operating system and the maturity of its tools, such as awk and ksh.
MS-DOS 6.0 had just come out in March 1993, and I was unimpressed with its new features. Version 6.0 wasn't much different from 5.0., and I wanted more. I began to look around. Most internet discussions at the time were via a distributed group system called Usenet, and it was in Usenet that someone mentioned this new thing called "Linux." It was a Unix-like operating system, but it ran on PCs. I could run it on my '386 computer.
Best of all, Linux was "free software," like several of the tools installed in the campus computer lab, such as GNU Emacs. In fact, Linux had all those tools (and more) that I had used on the big Unix systems. I was immediately attracted to a free Unix-like system that I could run at home, without having to dial into the campus computer network. To me, "free software" was like shareware, but better because I had access to the source code, so I could make my own changes.
I paid someone $99 to send me the floppies needed to install SoftLanding Systems Linux 1.03. It worked great. The tagline "Gentle Touchdowns for DOS Bailouts" proved true, as the installer was very DOS-like. After booting Linux for the first time, I was pleasantly surprised to find familiar tools: awk and sed to manipulate files, less and cat to examine files, and GNU Emacs for editing. For more advanced work, I had gnuplot to display data, gcc to write my own utilities in C, and f2c to write custom data analysis tools in FORTRAN.
While Linux sported a familiar Unix-like user space, I missed my favorite DOS shareware programs. For that, I kept a DOS partition on my hard drive and would reboot my computer with a DOS floppy to use the shareware word processor or spreadsheet. But it didn't take long to find DOSEMU, allowing me to run MS-DOS inside Linux, saving me the need to reboot the entire system for the sake of one DOS application.
A career in open source
Linux was a boon to me. I could do all my work from home, on my own system, without dialing into the campus computer lab—or trekking to campus in person if the modem lines were busy. My introduction to Linux was also my first taste of a new career option.
I graduated in 1994 with a bachelor's degree in physics but started my career as a Unix systems administrator. I never looked back.
Since then, I've "grown" with Linux. I've changed Linux distributions over the years, from SLS to Slackware to Red Hat to Fedora. I still run Linux on my desktop, but I've also introduced Linux to every organization that employed me.
My first job was at a geographics company. We printed custom maps. I moved our server support infrastructure (DNS, YP, LPD) off our aging Apollo/Domain and HP-UX servers to a few Linux systems. With Linux, we found the total cost of ownership to be lower than the big Unix systems. The day-to-day support effort was the same on Linux and Apollo, but the hardware cost was almost nothing compared to the purchase and support costs for the big Unix systems.
My second job was supporting a Unix environment for a document management company. Again, I transitioned our core "back office services" (DNS, file, web) from our expensive AIX and HP-UX systems to less-expensive Linux systems.
At my third job, a Big Ten University, I transitioned our AIX system to Linux. This was more than just installing a few Linux systems to run backend services. Instead, it was our first experiment with Linux in the enterprise, supporting a key web system. Linux not only performed well, it outperformed the AIX systems. On that merit, we began rolling out Linux across the enterprise, displacing AIX and Solaris servers at a fraction of the cost. By the time I left the university some 12 years later, two-thirds of our server infrastructure ran on Linux.
I'm a CIO in local government now, and while we have yet to install Linux in the year since I've arrived, I have no doubt that we will someday. The benefits of Linux are too great to ignore. Linux servers are cheap to install and easy to maintain. The total cost of ownership of Linux has proven to be much lower than any of our "big Unix" systems, especially for high-bandwidth applications such as web or file services.
And to think, it all started with a hobby Linux install in 1993. Linux has certainly grown in that time, and I've grown with it.
Read my interview with Jason Baker about founding the FreeDOS Project.