Paw Prints: Writings of the maddog

I just learned that Kenneth Olsen, one of the founders of Digital Equipment Corporation (DEC), died on February 6th, 2011 at the age of 84. From now on I will refer to him in this article as “KO”, as did most people who worked at DEC. At DEC no one had to ask who you were talking about when you said “KO”.

Readers of my blogs and columns know of my very long association with DEC. It started in 1969 when I was a college student, and one of the first computers I ever programmed was DEC's PDP-8. I taught myself how to program the PDP-8 in machine and assembler language based on the information from two paperback books that Digital published at the time. Each book cost five dollars, which was a lot of money back in those days for a college student, but the books were given to me by a DEC salesman.

As a college student I also joined DECUS, and I have often stated at conferences how the free exchange of programs from the DECUS library helped me on my path to computer science by exposing me to many programs that I could never have afforded if they were commercial software. "Free Software", in 1969.

I interviewed DEC at the end of college hoping to get a job, and while my career path took a different turn, I once again experienced DEC's hardware and software when I taught at Hartford State Technical College (HSTC).

At Hartford State we had two of DEC's machines, a PDP-11/70 with RSTS/E as a time sharing system (and 13 terminals), and a PDP-11/34 with RT-11 for “real time” work.

We had no budget for a hardware support contract, so we depended on “time and materials” for fixing issues that came along. Fortunately for us the machines were fairly solid, but one day the PDP-11/70 started crashing periodically, and over time we had to call in the local, the district, the regional and eventually a corporate field service person to try and find the problem. As I saw more and more effort being put into finding and fixing the problem, and more and more “time and materials” going into it, I knew that our minuscule budget would not pay for this.

When they finally did find the problem (a very intermittent problem in TWO circuit boards), I said to the corporate field service representative: “I do not know how we are going to pay you for this.” He looked at me, gave me a crooked smile and said “Yes, it was a hard one to find, wasn't it?” And the school never received the bill. DEC believed in supporting education.

After leaving HSTC I went to Bell Laboratories. There I had a chance to be a systems administrator for a CDC Cyber machine. I found out that they also had a set of DEC's machines running this strange operating system called “Unix”, and I fought to be the administrator of the DEC machines, to the point of almost turning down the job offer, both because I wanted to learn Unix and because I wanted to be working with the DEC equipment again.

Finally, when I had the chance, I left Bell Laboratories to work at DEC, helping to create their Unix products, first the line of products called “Ultrix”, then “Digital Unix”.

I did not meet Ken Olsen for a couple of years, but one day he arrived at the DEC facility where VMS was created. A few of our engineering staff decided to travel south a few miles from where we were stationed to hear the great man speak. KO was in the cafeteria, and he spoke for about a half-hour about how great VMS was, how great DECnet was, and how our philosophy of “One operating system, One architecture” was going so well. Then he paused, and told the audience that DEC could afford to step back and think about other things “like Unix....and of course we all love Unix!” Almost everyone in the cafeteria laughed at that statement, since almost everyone was from VMS. Then KO quickly finished his talk and people went up to shake his hand.

Our little group started back to our facility, and one person was griping that “KO had made fun of Unix”. I told them that KO had only been to “VMSland” once before, and now, on this second visit KO could have mentioned any of the other products instead of Unix...but he had singled out that one product. KO was sending them a very direct message.

My friends did not believe me, so I suggested that we go ask him. “ASK KO? TALK TO KO?” “Yes. He puts on his pants one leg at a time.” I said.

We approached him and I said “Mr. Olsen, we are from the Ultrix group, and we heard what you said. We would like to know exactly how much you love Unix.”

KO thought for a moment, then said, “Last year we were a 11 billion dollar company, and six hundred million dollars of that was Unix and Unix-related products, so I guess I love Unix six hundred million dollars worth.”

I thanked him for that assurance of his love, and that assurance kept me going for the next few years.

There were other things he said that caused controversy and consternation in our little product group, often taken out of context by the press (and by our competitors):

“They sell Unix like they sell Snake Oil” and “I find standards as interesting as a Russian Truck”

The first was an allusion to the fact that Unix vendors told their potential customers that Unix was “good for everything”, when Unix at that time was not real-time or highly available nor highly scalable. Today KO would probably agree that Unix (and Linux) have the facilities to be “good for everything” or at least “most things”, and today he would not accuse us of “selling Linux like people sell snake oil.”

The statement about standards was a personal statement. KO did not like working on standards committees because he hated to “nit-pick” (as he put it), but as an engineer (once the standard was developed), it was his job to implement it in the best possible way, and he enjoyed doing that. He did not mean by his statement that standards were not important, and in fact Digital supported many standards activities, but that he just did not like to be on standards committees.

Both of these he explained at the opening of the Open Software Foundation (OSF), and I rewarded him with a bottle fillled with mineral oil and labeled:

'“Uncle Ken's Snake Oil”: for refill see the Digital Unix Group'

KO's secretary told me that he kept that bottle on a shelf in his office until he left the company.

It was Ken Olsen who decided to fund a project to put Ultrix onto the MIPS processor, thus creating the first mainstream DEC system not based on a chip designed by DEC, that eventually turned into the “DECstation” line of products.

And it was KO who wore a button I gave him that said “I was Ultrix when Ultrix was not cool.”

It was to KO's private email that I sent a message complaining that an educational customer on the west coast had a system which had been crashing every day for a year, and the field service people were dragging their feet on repairing it. I told him that I was ashamed of that type of service, and (for the first time in my life) ashamed to be working for DEC. I never heard back from that email, but the next day there was a corporate "trouble shooter" flying out to the west coast on a corporate jet, and after landing he stabilized the problem in less than one day. I could be proud of DEC again.

There were many stories about KO and how he would wander through Digital's buildings at night, often looking into what the engineers were doing, giving suggestions on even the most minute things. There have been many things written about Ken Olsen and how he started DEC with a small amount of money which he carefully built up over time. Those people starting companies in the “dot-com boom” should have taken a page from how Digital make every dime of investment count.

KO and DEC gave many, many people the chance to contribute to computer science, to raise families and to have a good life. I always found him to be a fair and thoughtful man. If I have tried to pattern myself after any one person I have known in my professional career, it was after KO.

Rest in peace, Kenneth Harry Olsen.

Comments