This was my senior year as an ECE student at purdue. I gathered via archie and FTP the GNU utilities and the Linux kernel and got things working on my AMD 386 with a 387 math coprocessor. I remember around 4 Meg of RAM, and 20 Meg of hard drive. I mostly played around with it, used it a bit to work on programs that I then uploaded to the engineering network to test out. At the end of the year, I graduated, re-installed Windows, and gave the PC to my sister.
To work on my Master's Thesis at the University of Illinois, I installed Yggdrasil's version (the very first CDROM-based distribution of GNU/Linux) on my girlfriend's PC. She didn't really trust it yet, so I had to put it onto a separate disk drive and use LOADLIN.EXE to boot into it. I did virtually all of the C/C++ programming and much of the writing (in LaTeX) at home. The real number-crunching was done on SunOS, on a four-processor SPARC-10 in our research lab at The Computer Vision and Robotics Lab.
About this same time, I fell into being the system administrator for my research group at The Computer Vision and Robotics Lab at the University of Illinois. This system could, without much exaggeration, be called GNU/Solaris. We had everything from the shells and shell utilities on up running as GNU. This made my migration to GNU/Linux much easier.
I was getting ready to start real work on my dissertation, and another member of the lab group was graduating, so I bought his Toshiba Protoge 486SX-25 laptop with a 9" greyscale screen, 12M RAM, and a 200M hard drive. I pushed and shoved, and managed to get Redhat-4.1 to fit on it. I now was free to handle email (via POP), read netnews (via SLRNPULL), and write my dissertation (again, LaTeX) in the comfort of my favorite coffee shop in Urbana.
I was happy enough with the Redhat distribution that I installed Linux/Sparc on one of the Sun IPCs in the robotics laboratory. I had it up and running for a few months, but when the lab upgraded its version of Solaris, I dumped it. I was running a network of Solaris machines and one Linux box: not worth it.
I got a job as an Assistant Professor of Engineering Science at Trinity University in San Antonio, TX. I installed Redhat 5.1 on my wife's 486 to use while waiting for my university-owned computers to arrive. When my two machines (a Pentium-2-400 with 128M RAM for my desktop, a dual-Pentium-2-400 with 512M RAM for the lab) arrived, I put Redhat 5.2 on them. I also got a Dell laptop (Pentium-2-233 with 96M RAM) to write papers, and put Redhat 5.2 on it.
I upgraded all my machines to Redhat 6.1. Notice that most of my system maintenence gets done in the summer?
I was at NASA/JSC in houston for a fellowship. Both the robot that we were working with (an RWI ARTV Jr, with a dual pentium-2 AND a single pentium-2) and all the workstations around (with the exception of the vision guys' machine ... NT) were Redhat 6.2.
Trinity's computing center has begun "officially" supporting Redhat Linux, so I'm yeilding control of some of my machines to them. The standard replacement cycle for faculty PCs has kicked in, yeilding a bunch of new machines. About half the department (4 people) have decided to have Linux put on their new machines by the computing center, so I'm no longer the only Linux user around here.
I had to include a picture of my four-year-old with an abandoned P166 that I gave a home. I have to confess that I removed Linux and put W95 on it so that he could play his games.
I worked on a summer fellowship at JPL this summer. I worked on an old SGI workstation, and essentially created an GNU/IRIX system. I put all my favorite programs (Gnome, mutt, rxvt, mozilla, etc) on it and felt right at home! Most of the robot work for this project was done on an embedded DC computer running RedHat 7.0.
I've now installed Linux (RedHat 7.3) on my first PC/104 computer. I'm still playing with the interfaces, but I have enough confidence from talking to the guys at JPL that I worked with last summer to think that it'll work. So far, it seems to "just work"!
I've now completely given up my desktop, in favor of a Dell M50 Precision workstation (very cool laptop) with a docking station. I've got RedHat 7.3 on it. I'm getting support from the computer folks on normal type issues, but still have to do the hard/strange stuff myself, such as setting up automagic switching between the external Intellimouse and internal GlidePadPS/2, or the four (!) different networking configurations.
The Dell M50 now has Fedora Core 1 on it, and automatically switches between the (wired) port in the docking station, the wireless card for campus, and the wireless card for home.
The Dell M50 now has Fedora Core 2 on it. Many more things are automatic, I'm quite pleased with how solid the more modern desktop stuff is becoming (gnome). I still had to hack scripts to get the three network configs (all of which show up as eth1!) working, but more and more stuff is working "out of the box".
Update (04/02/2005) - someone asked for details about my network setup - here is the answer. Probably of interest (at most) only to people who want to run FC2 on laptops and have multiple networks that they use. If any such people have better solutions, let me know!
The PC/104 computer now has Gentoo Linux on it (installed by a student), and we have a linux device driver for the 48-pin digital IO card on it. We can read/write 48 TTL-compatable inputs/outputs from the console and from code now!
I finally gave up my Dell m50. I got tired of having to futz with things to make them work - when I got back from my 05-06 academic leave, wireless didn't work anymore, my backups did't work, ... I ended up getting a MacBook Pro with OSX on it. What convinced me was a demonstration of Apple's X11 (XDarwin) with Fink/DarwinPorts - ports of most all of the GNU Software to OSX, which is a reasonably full-featured UNIX. So if Summer 02 was a GNU/IRIX system, this one is a GNU/OSX system, and without me having to do it myself! I've still got most of my favorite tools (xfig, latex, vim, xdvi, espresso, matlab, etc.), but things mostly just work. Suspend/resume in particular just never made it to maturity on the M50. Wireless/wired networking connects automatically, external monitors hook up without a hitch, etc. My normal workflow hasn't changed a whole lot - I still write in LaTeX using vim, etc. By and large, I'm very happy with this computer/OS. I still use Linux on my research computers and on the Engineering network, which is now dual-boot WinXP/FC5.