Interview 3: Warren Robinett

When did you join Atari, and were you specifically hired to program the VCS 2600 system ?

I worked at Atari from Nov 1977 to June 1979. I was hired as a games programmer for the Atari 2600, then called then Atari Video Computer System (VCS).

I believe there were four 2600 programmers before me, some of who were hardware guys who designed the system, like Joe DeCuir. I was among four more programmers hired in Nov 77 to write game programs for the system.

Where did you gain your programming skills, and on what systems before your appointment?

I wasn't very experienced as a programmer when I went to work at Atari, but I was educated. I had just gotten a Masters in Computer Science at Berkeley the year before. So I had done some programming in school, at Berkeley and at Rice in Houston where I was an undergraduate. I also had worked for a year in Houston before grad school doing Fortran programming for a company in the oil business -- Western Geophysical -- writing programs to keep track of where the oil exploration boats were when they did seismic explorations. (I applied to grad school about a month into that job.)

How long did it take you to become comfortable programming the VCS system, and how long was it before you began work on your first title SLOT RACERS (Atari - 1978)?

It took 6-12 months to get your head into the weird little chip in the 2600 console. My first game, Slot Racers, was a learning experience for me. It took 5 months. (It was a 2K game, written in assembler, like all games at that time. It didn't take long to fill up a 2K ROM at an average of 2-3 bytes per instruction.) Luckily, everyone else was learning what was possible at that time, too, so the state of the art wasn't very slick.

We're you (and Atari) pleased with the reception Slot Racers received by the game buying public?

It was all right as a first game. I don't know the sales numbers. At that time, Atari was trying to get a lot of games out on the shelves, because -- hard to believe now -- it was competing with a number of other game consoles that hooked up to the home TV to see which one would win over the consumers. Obviously, the 2600 won. At that time, Atari published essentially every game that the game designers produced.

The marketing department was not acting as a gatekeeper -- they wanted product.

How many people were employed to work within the VCS 'team' and can you give us an insight into the day-to-day goings during a typical project ?

The answer is different for each year I was there (77, 78, 79) because the VCS became a success and the number of programmers grew fast. There were about a dozen at the time I left in June 79.

The main thing to know is that it was a one-person, one-game situation. Each programmer came up with a game idea, wrote the program, created the graphics, did the sounds -- in other words, did everything -- and turned over a ROM image when he or she was done. My boss when I started, Larry Kaplan, told me on the first day of work: "Your job is to design games. Now go design one." That was it. There were very few meetings. You just went in and worked on your game program. Atari was not a corporate kind of place. People came to work at weird hours. I started about 11am usually, and often worked late into the night when I was on a roll.

It worked because people got into their games -- they were labors of love. (At least that was true for me.)

We hear about prototype game, and other that just didn't make it to release for whatever reason. Can you remember any of these 'lost' projects?

All three of my game projects (Slot Racers, Adventure and Basic Programming) got published by Atari.

Your second VCS game, Adventure - is hailed by many as one of the greatest games of all time. What do you think of this accolade?

Any designer likes to hear that his work was good.

Can you give us an insight into the development of Adventure?

I had played the original text adventure (created by Willie Crowther and Don Woods) at the Stanford AI Lab where my housemate, Julius Smith, worked. This was at the time I was just finishing up my first 2600 game, Slot Racers. I thought the adventure game idea was really cool -- rooms and objects that you could pick up and carry around -- and I decided I would try to do a video game version of that concept.

The original adventure displayed static text on the screen and had a keyboard that let you type text commands as an input device, whereas the 2600 had a joystick with one button as an input device, could not do text, and had two hardware "hi-resolution" sprites available for graphics, plus a low-resolution background and three other low-res graphics objects. So figuring out how to do an adventure game with graphics was a tricky design problem. I decided to show one room at a time on the screen, use the sprites for objects, use the background for walls that you could run into, and let you move off the edges of the screen with your cursor (your self) to move from room to room. So the joystick could move you around, both within a room and between rooms, but that still left the problem of picking up and dropping objects.

I used running into an object to mean pick it up, and I used the button for dropping it. I considered allowing more than one object to be picked up, and at one time, had an inventory screen, but I didn't like suspending a real-time game while you looked at your inventory, and I also didn't like letting you get eaten by the dragon which you couldn't see (if I let the game simulation keep running while you looked at inventory). So I finally decided that allowing only one object to be carried at a time solved some design problems (because now I could just show the carried object beside the cursor) and also it forced the player into strategic choices. (Which object do I carry, the weapon or the treasure?)

You can see more I said about this at this web site:

Did you plan a sequel?

Yes, but I didn't do it at Atari. I wanted to make an adventure game in which you combined objects to build machines to defeat the monsters. I was going to have monster-detectors that produced signals which could be plugged into cannons to make monster booby-traps. This idea got turned into an educational software program called ROCKY'S BOOTS (1982 - Learning Company) for the Apple II computer.

By that time I had hooked up with three women whom were interested in computers and education (Ann McCormick Piestrup, Teri Perl, and Leslie Grimm) and we were making educational games. We started the Learning Company in 1980. Rocky's Boots was one of the 6 products in our initial product line.

I recently read that Adventure contained the first Easter egg. For those who haven't already discovered it - can you let us in on the secret?

I created a secret room that was very hard to get to and put my signature in it. I didn't tell anyone at Atari about the secret room (including the people who wrote the manual for the game). The player had to discover it -- there was a hidden object, the Gray Dot, that was just one pixel in size (the smallest possible object) that could be used to get into the secret room.

When you moved from Atari, and worked on Rocky's Boots for the Apple system. How did this project differ from those previous?

48K of RAM seemed like a lot after the Atari 2600 (128 bytes of RAM and 4K of ROM).

The processor was the same -- a 6502 -- and you programmed the Apple II in assembler, same as the 2600, so the 2600 was a good training ground for doing real-time graphics on the Apple II. The project differed in that it was to do an *educational* game.

After this it seems you moved away form the Video Games industry. Was this the case, and if so why?

I should have done another game after Rocky's Boots and Adventure, both of which were hits. Starting a company (the Learning Company) nearly killed me from the stress (I got neck and back trouble) and it took me a long time to recover fully -- several years. I went to work at NASA on what we now call virtual reality, and never got back to doing another video game or educational game.

Can you give us a history of your occupations from then onwards?

I worked as a VR researcher at NASA Ames and the University of North Carolina, taking time out from UNC to start a company -- Virtual Realty Games Inc. -- which didn't make it.

Do you look back on these early days of the Video game / Computer industry with fondness, or are you happier in your current occupation?

Yeah, it was a golden age when people tried all kinds of ideas. There were no game genres at that time -- they were being invented. Nowadays, with multi-million dollar budgets for games, it seems that there is less willingness to try new concepts. I like my current occupation, too, working as a computer graphics researcher at UNC.

Finally, what are the highlights from your 20+ year career ?

For me, the highlights are the people I met along the way.