An excerpt appeared in the print version of Stanford.
“I entered Stanford in the fall of ’64. In high school I had been very interested in physics and chemistry, and I guess must have signed up for EE even though I really wasn’t all that sure what that was. Before the first classes I met with my adviser, a James Angel in the EE department, and when he saw my interests, he said typically everyone’s required to take history of Western civilization their first year, but I think we could put it off and you could take introduction to computer science, which is a new course that’s just being offered. It was taught by George Forsythe.
That sounded really neat to me, so that’s what I did. You got to go over to Polya Hall and, using a keypunch, you punched up your card deck. All the computers were behind these big glass windows and the lesser people, the ordinary people, could not go behind and touch these things—they were special. So you would give them your card deck and you’d go away and you’d come back a few hours later and they would have run your program. There would be the deck of cards with a printout wrapped around it of your run—whether it was successful or, if it crashed, it would give you some indication why, and you would have to go back and debug your program and try again.
My other hobby on campus the first two years was breaking into buildings and exploring the steam tunnels. Back then the Farm really was the farm, and you could do things like that. You could have fun and it wasn’t a crime—although they certainly didn’t like you crawling around . . . the steam tunnels were dangerous; if a pipe broke you’d get cooked alive down there.
At the beginning of my junior year I was driving along Arastradero Road heading to Rosoti’s; I passed the D.C. Power lab and it had an official-looking Stanford University sign on it. By this time I had conquered everything down on campus. I’d even been up to the very top of Hoover tower. As I passed by the D.C. Power lab I made a mental note: I need to come back here tonight with all my tools. So I come back at 11 o’clock and I’m very disappointed because all the doors are open, the lights are on, and there are a fair number of people in there and they’re all sitting around this big PDP-6 computer. The PDP-6 was one of the first commercial timesharing systems, which meant that a dozen people could simultaneously be using the one computer, and it kept jumping around making each one think they had the whole computer to themselves.
So I saw this fantastic computer, and you could actually get up next to it and put your hands on it, and it was much neater than those ones down on campus. I started trying to get time. The machine was really for graduate students, but Les Earnest, the director of the lab, bent the rule in my case. He said if the machine’s not being used, and the resources are available, you’re a student so you can use them. That resulted in my adjusting my schedule where I’d sleep during the day and go up there all night. For the next two years, I just thoroughly loved going up to the AI project, working there rather intensively. I learned a lot and had fun.
While I’m doing that, a buddy from high school, Hugh Tuck, came over on several occasions to the AI Lab and we played Spacewar! He thought the game was great, I mean everybody did, and he said, wow, if we could build a coin-operated version of this, we could get rich. Three years later I’m working at Lockheed, where I’d been hired to be a programmer for a PDP-10, but Lockheed had never bought one so I was getting kind of frustrated sitting around without a computer. Meanwhile, DEC announced the PDP-11, which was a mini computer (it weighed 100 pounds) that you could purchase for $12,000. The next time I saw Hugh I told him now we could build a coin-operated version of Spacewar!
So Hugh and his brothers and sisters all chipped in, and together we had $20,000. Which is what it cost, by the time we were done buying the other components we needed, to get the first system built. Hugh and I knew that at that price it was too expensive, but we figured we’d build one and see the response and from that we could work backwards and see how cheap it needed to be to be successful.
We installed [the Galaxy Game] upstairs in Tresidder, where there used to be soundproofed rooms for listening to music. Outside two of the rooms was a kind of a lobby area and so the game was put there in the corner. Now in this first version, there wasn’t room in the cabinet for the computer. So there was a big cable that came out the back, went up the wall and then we had the computer sitting in the attic up above. The response we got initially was just incredible. People were packed around the machine continuously and standing on chairs trying to look over to watch. At that time, a game like that was just magical, to see these little things that you could steer and fire torpedoes.
After six to eight weeks of this enthusiastic response, Hugh and I decided to move forward; we didn’t really fully remember that this was just to figure out how cheap we had to make it. We made a little concession: We’d build another system and this time we’d have a more sophisticated display processor so it could drive two or even four displays and we could amortize the cost of the computer. The next system had very nice Fiberglas cases that Hugh had designed and we had a professional company, Peterson Products up in San Carlos, build. That was probably six months after the first one and that stayed for eight years and became a real mainstay.
The kids there developed their own rules. Friday, Saturday evenings there’d be a crowd of young guys around the machine, and there’d be people waiting to play. And when you’d come up, you’d take your quarter and put it on top of the machine—there’d be a line of quarters and that was the waiting line. When it came your turn, you’d take your quarter and drop it in and that got you three games. It was 10 cents a game, three for a quarter, and if you won it didn’t cost you anything. So people who were really good could put their quarter in and sit there for half an hour.
Maybe the best player was this young kid who was in 8th grade when he first came. This is ’75. And I made a deal with him that if you come by and clean the machine, then you can have some free plays. Then later, apparently, I gave him some instructions: If you’re driving in on the bad guy and you’ve got him in your sights and you’re firing torpedoes, if he drops into hyperspace—which just basically is a different space and he’s not visible but he’s still right in front of you, he hasn’t changed position—if you pop into hyperspace and keep firing and if one of your torpedoes hits him in hyperspace you’ll know because you’ll see the explosion. That was a very risky move because it took a quarter of your fuel to get in and out of hyperspace, and if you ran out of fuel in hyperspace you’re trapped. And if you ran out of fuel during the transition either way, then you blew up.
This kid, his name was Jonathan Rosenberg [son of economics professor Nathan Rosenberg], was enamored with the game and he would play all the time. Jonathan up until recently was a senior VP at Google, and about three years ago he wrote a post to their internal blog that was designed to kind of inspire the troops, and he told the story that I just told you. His point being that sometimes you have to take a big risk to get a big reward. And he printed out a copy of that blog post and sent it to me, saying do you remember me? I sent an email back saying of course I remember you, thank you so much for that story, and then I kind of changed the storyline into, by the way Galaxy Game needs to be saved right now, it needs a jump to hyperspace.
After I pulled Galaxy Game out of the coffee house in ’79, because it had gotten to the point where I just didn’t have enough time to keep it running—and by that time it had finally generated enough cash to pay back the principal investment—the pieces all sat around my house for the next 20 years. The Fiberglas cases were empty, but they were out under a big Monterey pine tree in my backyard. Then for whatever reason, I decided in ’98 or ’99, hey it would really be neat to bring these things back. So I called up Ted Panofsky, who was an engineer who worked up at the AI Lab maintaining all of their equipment. When we were building the second version of Galaxy Game I had hired Ted to build the display processor, and the two of us spent probably four months getting it running again.
Then I moved it over to the Gates computer science building; it was up there on the 4th floor, where it could really only be accessed by graduate students. The four years it was there it was very reliable, I never had a problem with it. But it wasn’t getting very much use, and my goal was to get it used as long as we could. I thought of the Computer History Museum, which at that time was still in some buildings near the big hangar [Moffett field]. I was a volunteer over there. They said, we’ll take it, but only on the condition that you give it to us, we don’t want any loaned equipment. They got the machine, and for the next year while they were still in the old facilities it could be powered up from time to time.
They hosted an event there, a Spacewar! contest. We had a four-person play-off: It was Nolan Bushnell, and I, Steve Russell and somebody else. And I’m happy to say I won. But I can also tell you that none of us was any good. It’s a game you really have to play for a while to get a feel for the timing, because you’re leading the other guy with your torpedoes. So for the first year it was powered up occasionally and used at the Computer History Museum. When they moved to the new facility they had a space called their visible storage exhibit, so they put it in there and it could not be turned on. That was kind of frustrating to me.
So when Jonathan sent me that blog post I very quickly seized on it. I just knew Google had to be one of the main benefactors of the museum, and if they asked for the machine they’d get it. That was the easy way to spring the machine loose. Jonathan responded back almost instantly to my email and said he’d love to. So it took awhile but we got the machine back. It took four months to get it running again; it didn’t like being left off that long. Finally it was moved over to Google in the spring of 2010. It ran there for a year, 15 months, then the building where it was was getting totally remodeled, so it’s in storage right now. Hopefully it will come out soon and find a new home where it can live out its days.”
