From a young age, I have been in love with UNIX, having been exposed to it at my mother’s office. She used to take me with her to the office, when she went into the office, on the weekends. Being a geek, this was nirvana, as my mother was a publisher who worked for a major satellite communications corporation. So, they had almost every type of publishing and graphics related hardware and software that was available, in the 1980s. On one of my earliest visits, my mother left me in front of a Wang-made UNIX System V system, and started Colossal Cave Adventure (ADVENT) for me. I was immediately entranced, and played for hours. Later that afternoon I got my first taste of the shell when I managed to exit from the game.
After that first experience, I pined for my own UNIX workstation. However, being a kid born in the mid 1970s, the cost of such an endeavor was far too large for my own means, and would have to wait for the availability of a free UNIX system that would run on the limited IBM PC clone hardware that I had at my disposal. The Softlanding Linux System, or SLS, for short, was one of the first complete Linux distributions, and was first available in August in 1992. That availability gave me my chance, as I was informed of its release by an older friend who lived in Washington, DC. He even made for me a partial copy of the 3.5” floppy-based installation media, but unfortunately either the copy was damaged, incomplete, or I was too lacking in understanding to get it working. However, that attempt did strongly re-motivate me to get a complete system working.
SLS was available for download from the Internet, or on CD or floppy from the publisher, or most commonly, copied from someone else who had downloaded it. For my second installation attempt, I bought a new box of cheap white-label 3.5” floppy disks from Egghead Software, which was one of the local computer store chains in Maryland, where I grew up. I then took those floppies with me to my high school, where I volunteered in the library and computer programming labs, to image with the new copy of SLS that I downloaded using their (for the time) fast 1.5 Mbps Internet connection. Once I had all of the disk images downloaded, I started creating floppy disks from them. Of course, being inexperienced, I had no idea what disks I would actually need, and which I could skip, so I made them all. This amounted to 22 floppy disks, and this is where I ran into my first significant problem: cheap floppy disks were unreliable! I had no computer, at school, to test the installation on, so every time I wanted to try to replace bad floppy disks, I had to go to Egghead, exchange the failed disks, and return to my high school, on the other side of town. I didn’t have access to a car, so this was all on my bike. Thus, I could make (at most) one attempt per day. It took me several weeks to finally get a complete set of installation media!
Once I had a complete installation set, the real fun began. I can still remember my nervousness as I erased DOS from my 105MB Seagate (yeah, I’m a geek, and I remember the exact hardware that I used!) hard drive, to prepare for the Linux build. Hard drives were very expensive, and I could not afford a second drive, or to get one large enough to also install DOS on. As you might expect, my first installation did not succeed. If my memory is correct, I got all of the software loaded on the system, but the bootloader (the software that starts Linux, when the system is powered on) failed to function properly. I repeated the installation process over and over, until I got a very basic running system.
Folks who didn’t use early Linux probably wonder why I didn’t ask for help, or just look on the Internet for a detailed explanation of the process. In the early 90s, you couldn’t just sign up with an Internet Service Provider, and use your fancy new telephone modem (the predecessor to today’s cable modems) to make a phone connection to the ‘net. In order to have an Internet connection, at the time, one had to be a scientific or educational institution. Commercial access to the Internet didn’t come about until 1995, in any widespread manner. So, I didn’t have anyone to ask, or any easy source of information, since my high school didn’t subscribe to USENET, which is the predecessor to today’s online forums. Because of that lack, I had to complete the process using only the documentation present on the installation media, and trial and error.
With my newly completed base system, I next wanted to be able to use a graphical environment, and not just a text-based command line interface. The system that provided such an interface, at the time, is called the X Window System, and was created by MIT in 1984. There was a version created for PC-based UNIXes, like Linux, and 386/Free/NetBSD, called XFree86, in 1991. That system was included on the installation media for SLS, but it was very much a manual process to configure it. Not only did you need to know exactly what video hardware was in your system, but you had to manually specify the exact parameters of the video signal that you wanted your system to produce. To get the necessary information, you needed to have the instruction manual for your monitor and your video card. If you made an error in the video configuration, not only would you not get a functional graphical environment, you could actually damage your monitor. Given all of this, and the fact that I, like many other computer users of the time, did not buy my monitor or video card new, it took me several weeks to get the X Window System properly configured.
Once the system was up and running, with the graphical user interface functional, I proceeded towards the last obstacle: customizing the Linux kernel. With modern operating systems, including Linux, you no longer have to build your own kernel, but with all UNIX variants until the mid 1990s, kernel compiling was required to add support for your specific hardware configuration. For example, if you wanted to add support for a sound card, like a Sound Blaster, Adlib, or Pro Audio Spectrum, you needed to build a custom kernel. Building the kernel was a mostly straightforward procedure, with one exception: you had to know exactly what hardware was in your system, and how it was configured. There was a large list of yes/no questions that had to be answered correctly, and once complete you could install the new kernel. If you made any errors, the new kernel would fail to boot the system, and you would have to fall back on the previous one (assuming that you hadn’t replaced it!).
With all of these caveats, and the clear complexity, you might well ask yourself why anyone would attempt to install and run Linux, in 1992. The answer varied by the individual, but for me, it gave access to the advanced tools and efficiency of a UNIX system, on PC hardware that were within my budget. All of the difficulties were just part of the journey.
