Game Chambers presents...

John Dunn was one of Atari's earliest video game developers. His groundbreaking Superman game remains one of the most enjoyable titles for the Atari 2600 (originally called the Atari VCS - Video Computer System). Since his Atari days John has continued to develop new software including graphics tools (such as SlideMaster, the "first ever professional paint program for a microcomputer") and music applications. He is also pioneering a branch of music called Genetic Music, with samples on his Algorithmic Arts website that prove to be graceful and mysterious. In this November 2000 interview conducted by Randy Ball, John takes time out of his busy schedule to discuss games, software development and art.

"Superman" for the 2600 is perhaps your most famous Atari project. What was your greatest challenge in creating that game?

The greatest challenge had more to do with bureaucracy than technology. It was to get permission to do a game that was more graphical and cartoon-like than the VCS games had been so far. This required the ability to use 4K ROMs which the VCS was designed to use, but which were not available to us because they were more expensive than the 2K ROMs currently in use.

Today 4K doesn't sound like much, but since we had been doing the entire games in 2K already, that meant I could use the other 2K entirely for graphics. At the time it was a big upgrade. I got permission to use the larger ROMs because Warner was about to release the first Superman movie, and they wanted Atari (which they had just acquired) to do a companion Superman game.

None of the VCS programmers really wanted to do it, so I agreed to take it on, on condition that I be allowed to use the 4K ROMs. Management agreed, but they were afraid I would "go artsy" and take a lot of time with it (I was the only VCS programmer with a background in art instead of computer science), so they insisted that I do it on a very tight schedule that would keep it in line with the movie release date.

A fellow VCS programmer, Warren Robinett, was working on an Adventure game that also needed 4K, so we collaborated, sort of. Robin gave me the Adventure kernel he had been working on as a quick start framework for Superman; and I broke the 4K barrier with management. And Robin was able to get the go-ahead on his very cool VCS Adventure shortly after.

Did this collaboration also influence the development of Adventure? For example, I can see how the bat in Adventure might be related to the helicopter in Superman.

Perhaps. Or the helicopter could have come from Warren's bat. It's been 20 years, details like that are pretty hard to remember. We all shared code then, so it might be said that every game had elements of every earlier game and not be far off the mark.

Why do you suppose the other programmers were reluctant to take on the Superman assignment?

That's an easy one. We all were very unhappy with the changes Warner was making. We felt (correctly, I think) they did not understand that game programmers were creative types, not engineering types, and needed to be treated accordingly.

There are some wonderful touches that add charm to "Superman", such as Lois Lane's subtle leg kick. Can you describe any ideas that you had for the game that didn't make it into the final product?

I wanted to make it more of a 2-player cooperative game. There is a cooperative mode, but it - not to put too fine a point on it - sucks. It would have been nice to have had the time to make that work better.

The existing 2-player feature that you refer to allows one player to control Superman's horizontal movement while a second player controls the vertical direction. Given more time, what might you have done?

It's hard to say. Probably give Lois more to do, and give control of her.

There is a Superman "cheat" that I learned about years ago where you can pause the game, wait for the "Daily Planet" screen to appear, and then maneuver Clark Kent into the doorway, effectively finishing the game in seconds instead of minutes. Was this a bug in the program?

Sort of. I hadn't planned it, but when I discovered it I decided it would be cool to leave it in.

Your programming credits include an Atari 2600 game called "Snark". I am not familiar with that title. What became of it?

That was my first game for Atari. It was not published while I was at Atari, and perhaps never was - I didn't track it. It had a video spin mode that caused the screen to color cycle really fast, and release was held up because there was some worry this would cause people to have seizures (I know, it's bogus - but this was the early days of video games, and that kind of intense color cycling was unknown territory).

How would you describe the game itself?

It was a combination Maze solver and shooter. Each game generated a new maze, and you were set upon by critters that you had to shoot in order to negotiate the maze.

Given the success of companies like Activision and Imagic, were you ever tempted to create games for a third-party developer?

While at Atari, I tried to interest management in producing games that had a positive educational underpinning. Not educational games, just games that carried positive subliminal messages. For example a D&D game where challenging a dragon to a puzzle would get better results than fighting it.

When it became abundantly clear Atari (Warner) was not ever going to take a risk on this kind of thing, I left the company with the intention of writing such games independently. I wrote a robot building block game based on the famous Lisp AI program; but it turned out the graphics editor I made for the game was more interesting than the game itself, and so I went in that direction instead.

And that's when you created the professional paint program "SlideMaster"?

Yes.

You mentioned that Atari (Warner) was not interested in pursuing games with the positive messages you envisioned. But didn't you accomplish that goal with Superman, where the player must rebuild a destroyed bridge and carry the thugs to jail, rather than blasting them away?

Not nearly to the degree I wanted. But it was, I had hoped at the time, a start.

Programming the Atari 2600 has often been described as difficult compared to programming other systems. What advice would you give to a hobbyist attempting to program an original 2600 game (feel free to be technical, if you like)?

Take the considerable time, talent, and energy it would require and put it into programming a DirectX based game, that can be easily ported to the XBox when it is released next year. This will give you a future instead of a past.

The recent release of Sony's PlayStation 2 caused frenzy in retail stores, which typically sold out of the console within hours. Aside from the obvious hardware advancements, do you see any key differences between modern video games and those of 1980?

Yes, a big difference: the advancing degree of sensory immersion. When you play one of the earlier games, no matter how absorbing it may be, you are playing AT the game, as you might play a board game. But with modern computer graphics, you play IN the game. So much so that when you think about the game, you think in terms of places you have visited. This trend will continue, of course, as the hardware continues to improve.

Throughout your career you have combined software with art. Do you see video game development as art?

No, I consider it craft, not art. Certainly there are artistic touches, and in modern video games there is of course a great deal of artwork. But a video game - any programming project, really - is more dependent on the attention to detail that we usually associate with craftsmanship, than the raw creative insight we associate with artistry.

You are currently the software creator for a company called Algorithmic Arts. What exactly is algorithmic art?

Algorithmic Art, sometimes called Generative Art, is art that is produced at least in part by algorithmic processes. An algorithm is the descriptive part of a computer program: if this happens, then do that, else do the other. I started writing algorithmic music and, to a lesser extent graphics, when I was using big analog modular video and sound synthesizers in the 70's in art school. So many knobs to control, as many as 300-400 in the setups I was working with, that I needed something to control them with. Microcomputers, which were just becoming available, proved to be the perfect solution.

The original idea was just to set up predefined sequences and step through them, so all the "knobs" could be adjusted at a time to some preset - in fact that was sort of a precursor to the presets that all synths have now. But as I started learning to use the computer, I discovered it was more interesting to have it make adjustments based on sets of rules, or algorithms.

In time, I largely discarded the synthesizer and kept the computer. But you can still see the influences of those old monster Moog modular synths in my software, such as SoftStep (which is a free download at url: http://algoart.com/).

What kind of application is SoftStep?

A MIDI step sequencer, a software emulation of the big modular step sequencers that predated the rhythm boxes. Also, since its introduction, it has become a fairly extensive general algorithmic music composing package.

Which gives you greater satisfaction: the craft of software development or the art that is derived from it?

Both. It is something of a feedback loop.

Your approach to musical composition is somewhat unconventional. How did your involvement in genetic music begin?

Genetic music was a natural extension of algorithmic music. I have always been interested in taking data from the world and looking for musical patterns in it, as part of the algorithmic process. It turns out that DNA and the proteins DNA describes has all kinds of patterns and motifs that lend themselves very well to music. There is an article by myself and my wife, a biologist, online that goes into this in some depth. It is at url:

http://mitpress.mit.edu/e-journals/Leonardo/isast/articles/lifemusic.html

What is your fondest memory from your association with Atari?

Before Warner acquired Atari from Nolan Bushnell, the VCS programmers had the freedom to design their own games from concept to finish. It was an intense, joyfully creative period that did not survive the takeover.

Intense in what way?

We were creating games. That is one of the most intense jobs there is, even today. And back then it was ALL new.