A Personal History of Linux  

By Joe Klemmer <joe@webtrek.com>

 

              In light of all the visibility that Linux is getting these days I thought I'd write about some of it's history, from a personal point-of-view.

 

              Today, the world knows about Linux.  It's in the news and all over the 'Net.  From enterprise triumphs to law suits.  Propaganda, both positive and negative, abounds.  You can't go into a bookstore or read a trade magazine without seeing a plethora of subjects about it.  But it wasn't always this way.

 

              I remember the first time I ever heard of Linux.  It was in November 1991.  Back in those days I was just starting with PC's.  I'd been working with mainframes for 9 years but had only been using PC's for a year.  The BBS scene was just reaching it's golden era, the Internet was something only for researchers & universities and UNIX was perceived as an old and dying thing.  Man, was that going to change.

 

              One day, in that long ago November, I got a call from a friend of mine telling me about this really cool new thing.  It was a UNIX-like operating system that was available on two 5.25" floppies.  The friend was Erik "The" Ratcliffe.  He brought over the disks and we played with it a bit over an evening.  You'd boot from one of the disks then switch it out for the other.  We ended up with a shell prompt.  The only experience I'd ever had with any kind of UNIX amounted to being able to spell it.  Now I was staring at a prompt with no clue what to do at all.  But the excitement I felt was incredible.  Back then, the closest thing to a personal UNIX OS was Minix and Coherent.

 

              In those days the only "distribution" was HJ Lu's boot/root floppies.  You could boot up and run a few commands like "ls "and, if you were adventurous, compile a few C programs.  It was Erik who wrote the first of what would now be called a HOW-TO.  He wrote instructions on how to get Linux to boot from your hard drive.  It involved hex editing the boot sector.  This was before lilo or even shoelace, the very first bootstrap loader.  I always remember this when people talk about Linux being difficult to install.

 

              In 1992 I was running my own BBS and also was one of the few people around who had dial-up Internet access.  Between '92 and '96, I would go and get files from TSX-11.mit.edu and SunSite.unc.edu and post them for download.  At first there were just a few source files but then came the distributions.  MCC, which was the very first.  TAMU, the first that could really run X.  SLS, the first distribution that was made for general use.  These were the beginning.  I had callers from all over the world downloading Linux files from my board.  One person from Portugal called every week.  There used to be a list of boards where Linux was available for download.  Now it's ftp mirrors. Imagine downloading all of Red Hat or Debian over a 28.8K modem.

 

              Things picked up slowly in those early days.  I was at the very first Linux event in the US.  It was at the 1993 Fed Systems/Open UNIX expo.  They had a Linux track that was quite well attended, especially for the time.  Michael K Johnson was there.  This was long before he was editor of Linux Journal and his subsequent work at Red Hat.  Vince Petry gave a great presentation on how Virginia Power was using Linux systems to monitor power lines and keep the electricity flowing.  Dr. Greg Wettestine explained how Linux was being used at the Roger Maris Cancer Institute.  It's astounding to think that lives were being saved by Linux that long ago.  I'm reminded of this every time someone claims Linux isn't ready for production systems.  This was also the first time I met John "maddog" Hall.  Donnie Becker gave an overview of the fledgling Beowulf system he was putting together.  He explained why they had chosen to build with 486 nodes because of the Pentium math bug.  It was also the first time I ever contributed money to the Linux cause.  There was some real excitement about something called Corsair.  This would eventually become Caldera.

 

              The sessions broke for lunch and about 12 or 15 of us went to a nearby restaurant to get some food.  It was a good lunch and the conversation was very interesting.  Then someone remembered that a number of the people there were supposed to be on a panel discussion.  I volunteered to stay and handle the check.  After collecting some money from the people there they went off and I got the bill.  It turned out that the bill was over $50 more than the money I'd collected so I paid the difference.  That really was a good bowl of soup I got for $55.

 

              Businesses weren't as keen as they are now either.  I had a friend who'd retired from the Army and was working for one of the local contractors doing Oracle work.  One of the big targets back then was getting Oracle to have a port for Linux.  My friend told me that some team in England had succeeded in getting it running.  However, the party line from Oracle was  less than favorable.  IBM was ambivalent then, still doing its best to develop and kill OS/2 in the same breath.  Then there was Computer Associates.  I had a few friends who were working for CA in the development and services/consulting end.  The techies had been floating the idea of Linux but the official CA line was less than enthusiastic.  As one of the guys told me, "It wasn't just no, it was Hell No!"

 

              Over the next couple of years, an interesting thing happened.  Software Development East used to be held in Washington DC.  Some of us would go there, and to other shows like FOSE, and harass... "politely ask" vendors when they were coming out with Linux versions.  The sales people would snort and give us the "Be gone, you troll" look.  But when talking to the developers a different story emerged.  I can't begin to count how many would say, off the record, that they developed their products on Linux then ported them to the MS platform for sale.  I found more companies who were using Linux, and Open Licensed software in general, for their development because of the low cost and high reliability but they weren't actually making Linux products.  The sales people would always sing the same tune; there's no market for Linux products.  They'd always "prove" this by saying that none of their customers were asking for Linux versions.  No one asked because they didn't make one so no one asked because they didn't make one so...  Circular logic at its best.

 

              I mentioned Linux books earlier.  I remember the real exhilaration when the very first one was published.  Running Linux was, to be honest, a monumental breakthrough.  You can't imagine the jubilation of seeing Linux on a book in a bookstore.  Erik even authored a book.  He was working for Caldera at that time.  This is back when Caldera was a good company.  And there was Linux Journal.  A real, live magazine about Linux.  The first issue was just a couple of pages of stock paper.  No color, no glossy pages, very little ads and a cover story about Linux 1.x from Linus.  My subscription started with issue #2.  I even wrote an article that was published in June or July 1996.