My History with UNIX and Linux

My history with UNIX

This article describes how I came to know about UNIX, and later, Linux.

Part of this is personal, documenting to myself when I started using each of the aforementioned systems.

Part of it is quasi academic research using digital archeology techniques to dig up old mails, search old files and directories, and try to piece together the sequence of events on how the above happened.

First Contact with UNIX

UNIX first came to my attention in the late spring / early summer of 1985. In that year, I got my first formal computer training, initially by NCR Egypt Ltd., and then by Price Waterhouse in St. Louis, Missouri. This was all part of developing a Management Information System (MIS) for my employer at the time.

The hands-on part of the training was in NCR's education center in El Agouza, Cairo, right on the banks of the Nile river. There, I saw two NCR employees working in the System Services Division (SSD), Nagla F. and Hazem C., using UNIX on an NCR TOWER. I remember seeing the lscommand in action, and was fascinated how creating a file was so easy using the cat > filename command! Then displaying it using the same cat command. This was a far cry from using OLPD on an NCR Criterion mainframe running the VRX operating system. I also saw a few games, including ASCII character “worms” running across the screen.

Later that summer, in St. Louis, MO, I bought my first book on UNIX, which was A User Guide to the Unix System, 2nd Edition, by Rebecca Thomas and Jean Yates, published by McGraw-Hill Osborne Media. This was a great book for a beginner, covering everything from shell commands, shell scripting, vi editing, text processing using nroff, and much more.

I did not get to use a UNIX system until a bit later, specifically in 1987, when I did part time work for a software house. The customer was a building contracting company in Alexandria, Egypt, and they had bought an NCR TOWER, which ran a UNIX V.2 variant on Motorola 680x0 CPUs. After working on a proprietary mainframe (NCR Criterion V-8555) and a proprietary batch oriented operating system (NCR VRX), this was a really refreshing and delightful experience to work on an interactive system, with the power of shell scripting. I just loved it!

In 1989, I started working for NCR Corporation itself, and did a lot of software support, consulting and development on NCR's UNIX systems, despite never getting formally trained on it. These were first the NCR TOWERs, which by then were running UNIX V.3. In 1990/1991, NCR's System 3000 made its debut, and they were Intel x86 based, and were using the NCR UNX System V Release 4 (SVR4). This later evolved into MP-RAS (a more robustized version with very good symmetric multi-processor SMP support).

SVR4 was interesting because AT&T and Sun worked together toattempt to end the fragmentation of UNIX, and merge SunOS (a Berkeley variant), V.3, and Xenix into one UNIX.

Check our collection of web pages on UNIX history.

My Quest for a Home UNIX

XINU

Perhaps in 1990, I bought Douglas Comer's book, "Operating Systems Design, the XINU approach". XINU was an interesting concept: it had all the source book published in the book. But to run a program, you have to do a build of the operating system, and link your program into it.

UNIX System V Release 4 (SVR4) for home

It was not feasible for me to get a mini-computer at home,although I did entertain the idea of having an old NCR TOWER at one point. The NCR TOWERs were really nice minis based on Motorola's 680x0 processors, and running UNIX V.2/V.3.

Eric S. Raymond (ESR) used to maintain the "386 UNIX on clone hardware Buyers FAQ", which is now named UNIX Hardware Buyer's HOWTO. ESR also authored the JargonFile (hacker terminology), and became famous later as an Open Source theorist and advocate, when he wrote his essay on The Cathedral and The Bazaar, and its followups, and then theHalloween Documents.

I started looking for such vendors, and comparing their products, inan attempt to find what would be suitable for me.

  • UHC
  • Microport (uPort)
  • Dell
  • Esix (by Everex)
  • Consensys
  • MST (Micro Station Technologies)

I contacted several of those vendors and got literature from them. However, their systems were expensive, several hundreds of US dollars. Moreover, there was no guarantee that after buying a system, that it would work on my specific configuration. The high price, and the possibility of completely wasting that money stopped me form purchasing any of them.

First Contact with Linux

Everyone who heard of UNIX and Linux must know that Linus Benedict Torvalds is the father of Linux. One of Linus firstposts on Usenet was on 28 May 1991, asking for help on an XT PC that his sister gave him!

A few months later, on 28 August 1991, Linus announced the birth of Linux in the comp.os.minix Usenet group, and again on 21 October 1991, and on 31 October 1991.

I heard about Linux around that time, or a little bit after that, from Usenet. I had earlier subscribed to the Minix Usenet group, part of my quest for a PC UNIX system. Linux was completely free, unlike Minix. In February or March 1992, I started emailing several people on how to get Linux.

We had no TCP/IP (and hence no FTP) at the time (only UUCP over X.25 -actually X.28 since it was dialup), and it was not feasible at all for me to download 3+ MB on a 2400 bps modem at the time.

I was finally referred to Linus, and emailed him asking if it is possible to send me a copy on floppies via postal mail.

His reponse, which I did not keep a copy of, went something along the lines of: "I am just a poor student and do not have the means to do that, but Ted T'so may be able to help you".

So, I contacted Ted Ts'o about getting Linux on floppies.

I do not recall exactly what Theo replied with, but I could not get Linux from him at the time

Here, you can read an early history of Linux was written from notes from Linus. There is also Ragib Hasan's History of Linux.

Creation of comp.os.linux

On 20 February 1992, Ivoted “yes” for the creation of the comp.os.linux Usenet group. Shortly afterwards, on 26 March 1992, the results ofthe vote were announced, and it turns out to be that I was one of 858 people who voted “yes” for the creation of adedicated comp.os.linux Usenet group, dedicated to Linux discussions. Only 5 people voted against it (I wonder why?)

Help with FAQ

I also helped with some FAQs in comp.unix.sys386.

Further attempts

So, in February 1992, this is What I knew about Linux

Unix User - another attempt to get diskettes

In the early 1990s, a guy by the name Steve Hanson started to publish a short lived newletter publication called Unix User. He dedicated one issueto Linux. I subscribed to this newsletter, and asked about his service of providing Linux on diskettes.

It seems I got no reply to this, because I resent it on December 8.

Again, I tried to get Linux on diskettes, but to no avail!

Google Groups (Usenet) search on Steve Hanson's newsletter. Here is the UnixUser article on Linux

Even in Canada, people were asking for Linux on a disk.

Hiatus from the Quest for a Home UNIX

I then continued my quest for a UNIX that would run on my home PC. I was not able to finda resonably priced UNIX System V Release 4 that was guaranteed to runon my home machine. The video drivers, the SCSI tape requirements,and the cost all were deterrents for me.

Linux, again! Five years later ...

I then had a five year hiatus due to life distractions, like having young kids and keeping busy with work.

Then in April 1997, I bought the Infomagic CD set, which included several Linux distributions. I got Slackware 3.0, with Kernel 1.2.xx) installed on my 486 NCR Globalyst 200S laptop (made by NEC), and I was really amazed with its quality.

Shortly afterwards, on 18 July 1997, I switched to Red Hat 4.1 (Vanderbilt), and on Oct 14, switched to a Toshiba 460 CDT, which had a Pentium 133MHz with MMX, 2.1GB disk, 12.1" active TFT LCD display, 32MB of RAM, and a CD-ROM. Then I upgraded to Red Hat 5.1 in November 1998. In May 1999, I upgraded to 2.0.36 kernel.

Thanks to Linux developers

On October 16, 1997 I emailed several Linux contributors and companies, thanking them for what they have done for everyone out there. These included Linus (of course!), InfoMagic (selling CDs with Linux distributions), DavidHinds (PCMCIA), Red Hat (RPM and distro), Stewart Allen (Visual TCL), Christian Bolik (Tk Desk).

I mentioned in it that I was using Linux for all my UNIX SVR4 development at the time on Linux (C, TCP/IP). I also mentioned that my brother is converting his business to Linux as well.

Later events, upgrades, ...etc.

Digging in old directories and files, I found that I kept a log in those early dayson what I have done. It was interesting to read through all this.Here is the log for anyone interested in this aspect of digital archeology.

Later I moved Linux to my home deskside machine. I do not recall if I upgraded to a later Red Hat release or not.

In December 2000, my PC was stolen in a break in. Luckily, I had backed it up on Travan Tapes (HP Colorado drive, 2.5 GB). For a while I did not have a home computer, and used the company issued laptop.

On april 14, 2002, I installed Mandrake 8.2 on a second hand Pentium II machine that I bought to be a server at home. The kernel was 2.4.18. I bought a used Seagate STT20000A IDE Travan tape drive for backup.

Later, in Late spring 2003, that Pentium II was replaced by a Dell Optiplex Gx1p with PIII 550 MHz. I upgraded that to 640MB of RAM, a Promise 66MHz IDE card, a Seagate 13GB disk for root, and an IBM Deskstar 20.5GB disk for /home. It also has a Seagate STT20000A IDE Travan Tape drive with 10GB tapes for backup, and a Misumi 2X CD-R.

Also, around May 2003, I upgraded to Mandrake 9.1, and got more computers at home for the kids, with each having her own in her room. I chose not to upgrade to 9.2 because its stability is less than adequate.

In June 2004, I upgraded all my home machines to Mandrake 10.0 Official with 2.6.3 kernel.

I had a problem with the CD drive causing a kernal halt. Check the details in that other page.

Contents: