RSS

Tag Archives: lucasarts

A New Force in Games, Part 1: Fractal Dreamers

Lucasfilm Logo

There are at least two stories to tell about the way that George Lucas’s Star Wars movies changed the world. One is the tale of the impact the films themselves had on the culture of movie-making and movie-going. For better or for worse, the first Star Wars film ushered out the brief New Hollywood era of auteur-driven American film-making that had followed the collapse of the old studio system in the 1960s, whilst ushering in, with a strong assist from Lucas’s buddy Steven Spielberg’s Jaws, the era of the action-packed, escapist blockbuster that still persists to this day. And of course for the nerdier culture of 1980s gaming Star Wars became nothing less than a third great holy text to join Star Trek and The Lord of the Rings, which does much to explain why it keeps showing up around these parts.

Yet there’s also another Star Wars story that’s less appreciated. During the production of the first film and especially when the millions began to pour in after its release, Lucasfilm, Lucas’s production company, forever changed many of the technologies behind media creation and consumption. To say they did so “quietly” would be overstating the case. Some of the names associated with the technological side of Lucasfilm — Industrial Light and Magic, THX, Pixar, Skywalker Sound — are well known to just about everyone. But the actual nuts and bolts of the new developments, even of those pieces that hide behind one of those big names, can be difficult to appreciate for anyone who isn’t a professional working in one of the industries whose practices they revolutionized. I don’t propose to tell the full story of Lucasfilm the technology incubator here. (That’s actually already been done, and much better than I possibly could at that, in Michael Rubin’s Droidmaker.) I do, however, want to tell you about what it meant to the world of computer games. Like a surprising number of things at Lucasfilm, game development just seemed to happen of its own accord, as something the guy who made Star Wars just really ought to be involved in. It wasn’t initiated by Lucas or any of his cronies, but rather by Atari, who came for a visit in 1982, just as both companies were at the peak of their power, wealth, and influence.

That meeting would mark the beginning of Lucasfilm’s direct association with computer games, but their association with computers in general stretches back considerably farther, to the immediate aftermath of Star Wars‘s release and massive success. Made using traditional mechanical, analog techniques — scale models, stop-motion photography, blue-screening, etc. — Star Wars had been an exhausting film to shoot and edit, so much so that it had sent Lucas to the hospital on one occasion with a stress-induced panic attack. With plans already afoot to make many more films in the series, he was naturally eager to find ways of easing the burden of unglamorous, mind-numbing labor that still was film-making of any stripe — much less a special-effects-driven science-fiction epic — in the 1970s. He started collecting talented computer people to help with that process. This collecting was made much easier by the fact that Lucas, who had evinced a visceral loathing for the Hollywood machine since the days when their trade unions had denied him work as a camera operator whilst he was still a student, had chosen to center his film-making operation in Northern rather than Southern California, much closer to Silicon Valley than to Hollywood. He wasn’t much interested in computer-generated graphics in the beginning. For a guy like Lucas, who had never darkened the door of a computer-science department in his life, the notion barely existed. What he really wanted was a way to do optical printing — the overlaying of separate shots onto one piece of film — more easily and without the degradation that resulted from analog techniques; a replacement for the huge, hot, noisy machines that editors had to use in conjunction with razor blades, glue, and thousand-page notebooks to — literally in those days — cut movies; an easier way to mix sound; even a good accounting system to keep track of all his millions. The people he found to help with all that and much more would set the world on the path to a digital revolution in filmmaking, creating amongst other things the predecessors to modern digital-compositing software and video-editing programs.

More relevantly to our purposes today, however, his dragnet also scooped up some of the best pure computer-graphics minds in the country — people like Ed Catmull and Alvy Ray Smith, the eventual co-founders of Pixar. Soon joining them was one of the pioneers of fractal graphics, Loren Carpenter, whom they lured away from of all places Boeing after he brought down the house at the 1980 SIGGRAPH computer-graphics conference with the short film you see below.


Brilliant as they were, the little Lucasfilm Graphics Group that coalesced around Catmull, Smith, and Carpenter didn’t often have a lot to actually do in those earliest years. The second Star Wars movie, 1980’s The Empire Strikes Back, made no use of computer-generated imagery. Raised on traditional film-making techniques, Lucas could see the value of other new computer-based technologies that made things like editing and compositing easier, but wasn’t quite sure what to make of images that were born inside a computer. And so, like a number of research groups at Lucasfilm, Graphics tinkered away in benign neglect, refining their techniques and waiting for their big break, showing up at the occasional conference with something amazing, which prompted a steady buzz in magazines like Byte about the groundbreaking things that were apparently happening somewhere within the secret bowels of George Lucas’s Star Wars empire.

Their big break ironically came not from Star Wars but from that other big science-fiction franchise Star Trek. Industrial Light and Magic had been hired by the producers of Star Trek II: The Wrath of Khan to do many of the special-effects shots for that movie. This they mostly accomplished using their traditional models, composites, and stop-motion photography. But there was one big effect, an animation illustrating a terraforming “Genesis device” on an in-film computer screen, that stumped them.  The script read simply, “And then the planet transforms… EFX sequence here.” It amounted to a creative blank slate for the Graphics Group to show the world what they could do — or perhaps to show just one man. “This is a sixty-second commercial to George Lucas,” Alvy Smith declared, “to show him what he’s got.” They threw it all in: fractals, 3D modeling, texture mapping, fluid animation. Steven Spielberg, who tended to drop by the Graphics Group far more often than Lucas, loved it, saying it was a “great time to be alive” in a world that had such wonders; Star Trek II producer Harve Bennett was elated; Lucas himself called it a “great camera shot,” which by his laconic standards was gushing praise. Released in June of 1982, Star Trek II became in a sense Pixar’s public debut, thirteen years before Toy Story. Lucas was impressed enough to give Graphics some work to do on the third Star Wars film, 1983’s Return of the Jedi, although they weren’t given the opportunity to make any showstoppers like their Genesis sequence. Undaunted, a prescient Catmull insisted, “We’re going to be making entire films this way someday. We’ll create whole worlds. We’ll generate characters, monsters, aliens. Everything but the human actors will come out of computers.”


In the meantime, there were suddenly games. It was when the Graphics Group had just finished the Genesis sequence that Atari came to their nondescript offices in San Rafael, California, for a visit. Flush with even more cash than Lucasfilm at the time, Atari had quite a variety of research projects in progress, even if they would prove remarkably awful at turning them into finished products to supersede the aged Atari VCS games console. Thus it was natural for them to want to visit another company’s cutting-edge graphics research facility and see what they were up to. The two companies were hardly strangers; Atari had just released a licensed VCS game based on Raiders of the Lost Ark, Lucas’s hit cinematic collaboration with Spielberg.

What Atari saw in San Rafael blew them away. The delegation bombarded the bemused Catmull, Smith, and Carpenter with a million questions about how their stuff worked and, most importantly, how it might be adapted to videogames. Only Carpenter showed any real interest at all in such an endeavor. They were used to working on big workstations and minicomputers, not primitive micros, which they viewed with a certain contempt, dubbing their programmers mere “bit twiddlers.” They had high standards for their visuals: their graphics had to be good enough not to look out of place projected on a huge movie screen surrounded by other imagery shot on pristine 35-millimeter film. Their greatest enemies were what they called the “jaggies,” visibly blocky, pixelated areas that tended to lurk at the margins of what should be smooth, flowing curves. While the jaggies could be held at bay using the state-of-the-art, processing-intensive anti-aliasing techniques that the Graphics folks had spent years developing at Lucasfilm and elsewhere, those techniques weren’t much applicable to an 8-bit games console. Catmull and Smith at least wanted no part of that action; Carpenter was intrigued but also ambivalent, certainly not willing to entirely give up his film work for a game project.

Yet Atari persisted. Even Manny Gerard, the Warner Brothers executive who had orchestrated that company’s purchase of Atari, got involved, saying that “we ought to be in business” with Lucasfilm’s Graphics Group. Finally Atari offered to flat-out give Lucasfilm $1 million to set up a Games Group, for the products of which Atari would receive “right of first refusal” as publisher. An offer like that was hard to refuse. The deal was announced at the Summer Consumer Electronics Show in June of 1982. “We’ll be developing new forms of electronic entertainment,” said an excited Atari, “and the term ‘electronic entertainment’ is carefully chosen.” Don’t, in other words, just call them yet more videogames. Lucasfilm and Atari would continue to work hard to cultivate this rarefied image of Lucasfilm Games as artists of interactivity rather than mere game programmers over the months and years to come.

Had Atari not so aggressively forced their hand, one might be tempted to characterize the whole undertaking as something of a bait and switch on the part of Lucasfilm; no one currently in that Graphics Group whose work had so impressed Atari was earmarked to start working full-time on games. Still, the initial organization of an entirely new Games Group was tossed into the lap of the Graphics Group’s Ed Catmull, who found a fellow named Peter Langston at a Wall Street law firm to head it. Langston was a Unix hacker from way, way back who in 1971 had written Empire, an elaborate multiplayer strategy game played on a global scale, a forerunner to Civilization and other games of its ilk. (A later offshoot of the original game was even named Civilization.) An accomplished musician, he was very interested in the application of music theory to computers and vice versa and was, in the judgment of Catmull, just a visionary “star” in general, the perfect guy to take Games off his hands. When Langston, who was quite happy in New York, proved reluctant, Catmull continued to sweeten the deal, going so far as to offer to fly him back to New York for a couple of weeks out of every month if he liked. It was, once again, an offer that was hard to refuse. Catmull got his guy, and with a sigh of relief turned his attention back to film graphics.

Peter Langston surrounded by typically artsy trappings in his office.

Peter Langston surrounded by typically artsy trappings in his office.

Langston was an unusual choice for leader and administrator, a conceptual rather than an altogether practical thinker with a somewhat dreamy disposition. Working at the unhurried pace that would be typical of the young Games Group, he put a little team together to join him. He held the zap-em blast-em world of typical videogames in little more regard than did his colleagues in Graphics, and thus purposely avoided programmers with a lot of experience in the industry; he was after people like him, people who were “a little bit visionary.” He hired one David Fox because he admired Computer Animation Primer, a book Fox had recently written for which he had actually met and interviewed many members of the Graphics Group. It took him weeks more to settle on a very eager David Levine as a third team member. His biggest claim to fame was having designed the first add-on graphics board for the original Altair kit computer, which Langston thought was great. However, he’d also already done quite a lot of videogame programming, which maybe wasn’t so great. In the end Langston decided to give him a shot in spite of his surfeit of experience. A fourth employee, Charlie Kellner, late of the Apple Macintosh development team (another musician, he had programmed the pleasant little beep the Mac made at startup), would arrive still later. Like Graphics and many other teams benefiting from the Star Wars millions inside Lucasfilm at the time, Games walked a shadowy, largely unsupervised line somewhere on the intersection of a pure research group and a commercial proposition expected to deliver actual, tangible products. Certainly Games was nothing like the quickie projects being started by many other big companies to cash in on the videogame fad. Nor would they try to trade on the Star Wars name. In fact, they wouldn’t even be allowed to make Star Wars games.

Whilst negotiating with 20th Century Fox the deal that would lead to Star Wars, George Lucas had agreed to license the right to make Star Wars “toys and games” to Kenner Products. At a time when videogames still largely meant Pong, everyone interpreted “games” in this context to mean board games, which would be issued by Kenner’s subsidiary Parker Brothers, whose stable already included family perennials like Monopoly, Risk, and Clue. When videogames exploded a year or two later in the wake of Space Invaders (whose popularity was itself fed by Star Wars and the craze for all things science fiction that it engendered), Parker Brothers found themselves gifted with a golden goose for the ages, as was amply proved when they released an Atari VCS cartridge based on The Empire Strikes Back that spent many months in the top ten. The Star Wars videogame-licensing rights would be tied up in knots for years to come, the interested parties including not only Lucasfilm and Parker Brothers but also Atari, whom the latter agreed to allow to make standup arcade games and eventually console and computer games of their own under the name. (One way or another, everything involving videogames in the early 1980s always seemed to come back to Atari.) The end result was a circular tangle the likes of which only corporate America could create. Lucasfilm, the owner of Star Wars, had a games division that wasn’t allowed to make Star Wars games, while Atari had such a license thanks to kicking some money back to board-game manufacturer Parker Brothers, but chose to create those games in-house, even though they’d just paid Lucasfilm $1 million to set up the aforementioned games division for the purpose of making games for them. The upshot, however, was simple: no Star Wars for Langston, Fox, and Levine, nor for that matter for any of the others that would join them over the course of the rest of the decade. Many who were at Lucasfilm Games during this period have since remarked on what a blessing in disguise this really was, forcing the developers as it did to come up with original game concepts, original game fictions.

Even if they were barred from working directly with the Star Wars intellectual property, it was a damn good gig just to work for Lucasfilm, flush with cash, with all of the best equipment, with few or no hard deadlines, and right there on the close periphery of where the movie magic happened. If George Lucas himself seldom poked his head in the door, there was every day the possibility that he would. And there were other famous faces who were a more common sight, like Steven Spielberg, who spent quite some hours in between Industrial Light and Magic effects shots for Indiana Jones and the Temple of Doom camped out in the Games Group’s offices playing the Star Wars standup arcade machine Atari had been kind enough to send them. One day Spielberg recruited the team to become the screams of dying bad guys in Temple of Doom. Whatever else competing games studios might offer, they couldn’t offer experiences like these.

Of course, being constantly in the shadow of Star Wars and Indiana Jones could also be exhausting in its own way. David Fox tells the story of attending an early trade convention where a showgoer, excited by the games on offer, for once failed to make the connection, asking what else Lucasfilm had made: “I said, ‘We made the Star Wars films.’ Watching the look on his face was hilarious as he made the connection. And it was nice for once to not be in the shadow of the films!”

As for what they were actually working on in those early days… well, that wasn’t always entirely clear, at least for those looking in from the outside. They had unilaterally decided quite quickly not to develop their first games for the technologically antiquated Atari VCS games console, as Atari had anticipated, but rather to target the Atari 8-bit line of home computers and the new, more advanced Atari 5200 console that had largely the same internals as those computers — this even though neither platform was selling in anything like the quantities of the old VCS. Langston pronounced himself “amazed” when his research into the ways that games were typically made for home computers revealed most developers to be working “in a basement on a system with too little memory, too few floppy disks, no reasonable way to make a backup, and few if any debugging tools.” Determined to change all that, the trio spent months on an elaborate development system which ran on their big DEC VAX minicomputer. Conceptualized largely by Langston himself, it allowed them to write code in a LISP-like script, compile it, download it to an Atari 8-bit machine, and debug it as it ran there from their terminals. Similar systems, eventually capable of compiling down to a whole range of other microcomputers, would remain the core of the Games Group’s development methodology for years to come.

Ballblazer

Ballblazer

Their efforts to create actual games slowly coalesced around two projects, both of which had started as “throwaways,” learning exercises to work the kinks out of their development system and help the team gel. (That very description says much about Peter Langston’s academic style of management, and why Lucasfilm’s own management would in time start to find it kind of infuriating.) David Levine’s baby Ballblazer was the more artsy of the pair, a surreal 3D soccer game played between opposing spacecraft, with realistic if otherworldly physics and, most impressively, jazzy generative music provided by Langston that’s sometimes been compared to that of John Coltrane. The music was created using fractal algorithms pioneered by the Graphics Group that just seemed to be in the air in those San Rafael offices. Speaking of which: David Fox’s Rebel Rescue got a huge assist from Loren Carpenter, the only real games fan in the Graphics Group, who helped him to implement an admittedly jaggies-replete fractal landscape for his more grounded game that cast the player as a futuristic search-and-rescue pilot, trying to rescue downed pilots from a planet’s surface and bring them back to the mothership whilst fighting off hordes of invading aliens that are still swarming the atmosphere. Deciding they might as well have a sense of humor about the thing, they decided to call these aliens the Jaggies. Rebel Rescue would actually have made a darn good Star Wars game; its inspirations, including not only the name itself but also the X-Wing-like spacecraft you fly and the pilots you rescue in their distinctive orange flightsuits, are pretty hard to deny.

George Lucas, for whom games held little personal appeal, sat down with his Games Group exactly once during their first couple of years of existence. And yet his single visit had a huge impact on Rebel Rescue. The little team, idealistic as they were, took pride in the fact that both of their games were basically nonviolent. In Rebel Rescue you could avoid the enemy aliens or, if you were skilled, trick them into flying into mountains, but you couldn’t shoot at them. “Where’s the fire button?” Lucas asked. Fox explained. “Is there no shooting because of gameplay reasons or philosophical reasons?” Philosophical. “Great. Put in a fire button. I want to shoot at things.” It was also Lucas who suggested what would prove to be Rebel Rescue‘s second most memorable feature after the fractal terrain itself: some of the downed pilots are aliens in disguise, whom you have a split-second to zap before they kill you. This was deliberately left out of the game’s manual, thus nearly giving many players a heart attack when it first happened a few levels in.

Even as the team tinkered away with Ballblazer and Rebel Rescue, the Great Videogame Crash of 1983 was happening outside their ivory tower. Nevertheless, at year’s end the Games Group at last delivered working prototypes of both games to an Atari who had declared themselves bloodied but determined to fight on, who were “going to reignite the consumer’s love of videogames.” Just weeks later the Games Group was horrified to see their babies spreading like wildfire across the worldwide network of pirate BBS systems. It seemed that someone at Atari hadn’t been able to resist sharing these cool new games with a friend or two, and the thing had just exploded from there. Soon copies of the games started to show up for sale in flea markets and the less scrupulous software shops, decked out in homemade packaging invented by enterprising quick-buck artists. A still buggy Rebel Rescue in particular seemed to be in the collection of every Atari 8-bit owner on the planet, one of the most popular games on the system. Lucasfilm and Atari had a hit on their hands, but it was a hit they weren’t getting paid for. It was questionable whether Langston and his idealistic cohorts were more upset about the potential purchasers it was costing them or the fact that the games everyone was playing weren’t finished yet.

Most of the Lucasfilm Games Group, mid-1984: Charlie Kellner, David Levine, Peter Langston, David Fox, Loren Carpenter (visiting from Graphics), Gary Winnick

Most of the Lucasfilm Games Group, mid-1984: Charlie Kellner, David Levine (seated), Peter Langston, David Fox, Loren Carpenter (visiting from Graphics), Gary Winnick

Lucasfilm was understandably less than thrilled at Atari’s failure to protect their games. Atari, however, was also less than thrilled with Lucasfilm. It had now been eighteen months since their $1 million investment and, especially in light of their straitening financial circumstances, they wanted to see some finished, polished games in return. They were particularly unhappy that the Games Group had failed to deliver on a promise to give them something ready to show at the 1984 Winter CES. By now Lucasfilm management as well had decided that something had to give. In January of 1984, they hired one Steve Arnold to join Peter Langston as an awkward sort of co-manager. Arnold came from Atari, where he had been responsible for the Atarisoft line of ports of standup-arcade games to home computers, one of the few financial bright spots at the company during 1983. He’d rolled out an astonishing 49 separate conversions in five months as head of Atarisoft, so he certainly knew how to ship product. But a few Lucasfilm higher-ups wryly noted that his best qualification could be his PhD in psychology, or perhaps the time he’d spent years before as program director at a boy’s camp. Hopefully he could find a way to make the unruly Games Group toe the line without spoiling what made them unique in the first place. Also coming aboard around this time were two more team members. One was Noah Falstein, an established videogame designer and programmer late of Williams Electronics, where he’d worked on standup arcade games like SinistarThe other was the Games Group’s first full-time visual artist, a veteran commercial artist named Gary Winnick.

Rescue on Fractalus

With these new, somewhat more practical-minded additions, the Games Group did indeed start making progress more quickly; the new can’t-miss-it deadline promised to a still skeptical Atari was now Summer CES in June. Rebel Rescue was renamed Rescue on Fractalus! largely for legal reasons, to make it clear that it was not (officially) a Star Wars game despite the Lucasfilm logo on its box. The folks at Industrial Light and Magic built and photographed model spacecraft for the boxes, designed and built a cockpit model for the “Valkyrie” fighter the player flew in Rescue on Fractalus!, and even made flightsuits for the entire development team to wear in a grand photo spread. David Fox got to play the starring role, as a weary pilot trying to straggle home in his battered Valkyrie on the back of the Rescue on Fractalus! box. The whole effort cost at least $30,000. Yes, working for the company that made Star Wars did have its perks.

Final versions of both games, complete with packaging, were delivered to Atari well before the latest deadline. Lucasfilm held a lavish press conference to unveil them in May, presenting the games via slick videos created with the aid of professional voice actors and Lucasfilm’s general movie-making know-how.


A beleaguered Atari determined to press on (or still in denial) came to Summer CES with a new slogan: “The Day the Future Began.” The “Atari-Lucasfilm” games got a very positive response from press and public alike, and the partners put on the final touches for a July release. Yet practical questions still surrounded them. The Atari 5200 console had proved to be a flop, had already been discontinued, while Atari’s line of 8-bit home computers was still on the market but overshadowed by the cheaper Commodore 64. Meanwhile Atari themselves were still in financial free fall. And then, overnight, everything changed once again.

On July 3, 1984, Warner Communications announced that Jack Tramiel, late of Commodore, was buying Atari’s home-videogame-console and home-computer operations, surprising no one more than the people inside Atari who suddenly had their company sold out from under them. Looking on from the outside as Tramiel axed employees by the thousands in the weeks that followed, the Games Group wondered if July 3 should be called “The Day the Future Ended.” No one seemed to quite know in the midst of all the chaos where it left the Lucasfilm/Atari partnership. Tramiel himself didn’t seem to know much about their companies’ agreement and didn’t much seem to care. And so Steve Arnold set off to try to free the games from Atari’s exclusive clutches. He returned from his one and only meeting with Tramiel with a less than positive personal impression, saying that the latter reminded him of no one so much as Jabba the Hutt of Return of the Jedi fame. Sure enough, pictures of the two were soon hanging up around the offices of a very frustrated Games Group: “The Hutt Brothers: Jabba, Jack.” (Poor Jack just couldn’t win; the common comparison inside Atari itself was to Darth Vader.) But in this one case at least Tramiel’s bark was worse than his bite. Busy with other legal battles and the travails of rebooting Atari, he let Lucasfilm move on.

Arnold settled on Epyx as publisher out of a crowded field of suitors, signing a four-game deal that was announced with considerable fanfare at Winter CES in January of 1985. Ballblazer and Rescue on Fractalus! became widely available for purchase at last shortly thereafter. These, the Games Group’s first actual products, had taken two-and-a-half years to come to fruition, an eternity in an era when most videogames were still churned out in a matter of a few months.

Thankfully for the Lucasfilm brass, it looked likely that the next games wouldn’t be so long in coming. Peter Langston had bowed out at last in the fall of 1984, taking with him his rather abstract approach to game development and freeing Arnold to continue to refine the Games Group’s operations along more practical lines. And there was suddenly plenty of practical work to do. No longer beholden to Atari, they were now free to port Ballblazer and Rescue on Fractalus! from the fading Atari 8-bit line, every owner of which seemed to already have them anyway thanks to the leaked demo versions, to stronger platforms like the Commodore 64, where they ended up selling far more copies. Rescue on Fractalus! in particular became a hit, not a blockbuster and certainly not enough to justify the time and money poured into it absent Atari’s initial $1 million beneficence, but a solid piece of groundwork that established Lucasfilm Games as a maker of classy but accessible action fare. If it seemed just slightly underwhelming in light of the years it had been in production and all of the flashy promotion that surrounded it — seemingly every magazine in the industry published a big Lucasfilm Games feature article around this time; such was the cachet of the house that Star Wars had built — well, just about any game realizable on an 8-bit computer would have. It didn’t help that the year that had passed between the leak of those demo versions and the arrival of the finished games on store shelves had allowed lots of other programmers to start experimenting with fractal graphics, making Rescue on Fractalus! look far less revolutionary than it otherwise would have.

The next two games were very much designed to build on the technical as well as the commercial groundwork laid by the first two; both started with the graphics engine from Rescue on Fractalus!. Designed by Charlie Kellner, a newcomer who had been hired just before Langston left, The Eidolon had the player piloting a machine through networks of underground tunnels inside the protagonist’s own mind — fractally generated, naturally — full of dangerous “guardians of the id.” Koronis Rift hewed still closer to Rescue on Fractalus!: this time you were flying above an alien-infested, fractally-generated planet trying to collect technological relics rather than downed pilots. Both were once again well-reviewed when released in late 1985 after comparatively reasonable one-year development cycles, going on to sell modestly well, if not to match the sales of Rescue on Fractalus!. In the span of 1985 the Games Group had increased their catalog from zero to four solid action games, one of them a genuine hit, and were now largely self-sustaining.

At the same time, though, a slight sense of underachievement clung to the Games Group, who had failed to completely deliver either the revolutionary gameplay experiences for which Peter Langston had been hired or the blockbuster sales figures one might expect from the company of Star Wars. They were still something of an odd duck in the industry, their huge cachet still largely based on that name on their boxes rather than the actual contents of the disks therein. Yet, even after Peter Langston’s departure, the sense of artistic idealism he’d worked so hard to engender remained alongside Steve Arnold’s determination to actually ship games on a semi-regular basis. “I think in general we’ll be moving away from the concept of games,” said Kellner shortly after the release of The Eidolon and Koronis Rift, echoing some of the verbiage Atari and Lucasfilm had used when first announcing the new venture three-and-a-half years before. “We’re trying to produce an experience that’s like being part of a film, rather than just being part of a game.” The fact that they were still having to promise to move beyond mere “game” in the future could be read as an admission that visionary software had proved to be a bit more difficult to develop than expected. On the bright side, their next project would be by far their most audacious and, yes, visionary yet.

(Sources: the book Droidmaker by Michael Rubin; Byte of March 1984; A.N.A.L.O.G. of August 1984, April 1985, August 1985, March 1986; Antic of August 1984, December 1985; Commodore Power Play of August/September 1986, October/November 1986; Compute!’s Gazette of August 1985; Compute! of August 1982, November 1982; Creative Computing of March 1982; Enter of September 1984; Family Computing of August 1986; Game Developer of December 1994; K-Power of September/October 1984; Zzap! of February 1986, March 1986; Retro Gamer 27, 44, 116; the website LucasFans, now available only via the Wayback Machine; Peter Langston’s paper on the early Games Group’s development system, available from his website.)

 

Tags:

The Fractal Phenomenon

How long is the coast of Britain? Would you be surprised if I told you that it’s infinitely long? Infinite coastlines are just one of many byproducts of the strange mathematics of fractal geometry.

Why is geometry often described as “cold” and “dry?” One reason lies in its inability to describe the shape of a cloud, a mountain, a coastline, or a tree. Clouds are not spheres, mountains are not cones, coastlines are not circles, and bark is not smooth, nor does lightning travel in a straight line.

— Benoit Mandelbrot

For centuries mathematicians and philosophers were dogged by a nagging problem. Classic Euclidean geometry, with its regular lines and planes and solids, was elegant and useful, but fell down when it came to describing the natural world. It was ill-equipped to describe the seemingly endless spatial complexity of a leaf — or, for that matter, a section of coastline. Indeed, it’s largely the very presence or absence of such chaotic complexity that we subconsciously use when we decide whether something is natural or human-made. The world humanity has built is created in the image of our geometry, while the universe that birthed us, even the very forms that enclose us, is defined by its refusal to conform to our ideals of coordinates and shapes and angles. To believe that the universe is somehow corrupted or deformed because of this refusal would be the height of hubris. “These are not natural irregularities, but with respect to our fancy only; nor are they incommodious to the true uses of life and the design of man’s being on earth,” wrote Richard Bentley in the time of Swift and Pope. Yet how to reconcile the riot of natural beauty with the cooler constructions of Euclid? Until very recently we literally lacked a mathematical language to describe the world around us.

In 1975 the Polish/French mathematician Benoit Mandelbrot provided such a language when he published his theory of fractal geometry. Mandelbrot discovered that certain solution sets to certain equations generate what he called “fractals” — patterns that are infinitely detailed. Such sets were much better suited to describe natural forms than anything heretofore present in the geometric arsenal.

It’s this seeming paradox of infinite detail that’s the thorniest conceptual part of natural forms and fractal geometry alike. To begin to get a handle on it, consider the “snowflake” curve, an idea first proposed by Helge von Koch in 1904. Imagine that you start with a line forming a simple geometric motif.

a "snowflake" curve

Now you replace every straight-line segment with another copy of the original motif.

a "snowflake" curve

You do it again…

a "snowflake" curve

And you keep on doing it over and over and over, to infinity.

a "snowflake" curve

As you look at a coastline or a leaf at ever greater levels of magnification, there are always more details, more curves and changes to discover. This continues forever — or at least until you hit the level of individual particles of matter, the sort of resolution limit of the universe itself. Thus that aforementioned infinite — or at any rate unmeasurable — coastline. “Coastline length turns out to be an elusive notion that slips between the fingers of one who tries to grasp it,” wrote Benoit Mandelbrot. “All measurement methods ultimately lead to the conclusion that the coastline’s length is very large and so ill-determined that it is best considered infinite.”  What to make of such a logical monstrosity?

Mandelbrot found a way out in some theories of dimensionality espoused by the controversial German mathematician Felix Hausdorff more than half a century before. The easiest way to describe conventional notions of dimensionality mathematically is to say that an object’s number of dimensions equates to the number of numbers needed to describe a single point on it. Thus a line (X) is one-dimensional; a plane (X and Y) two-dimensional; a solid (X, Y, and Z) three-dimensional. Hausdorff, however, floated the counter-intuitive notion that dimensions could be fractions as well as whole numbers. Consider a line. As it gets more wiggly, ever more complex in its gyrations, its dimensionality increases, until it becomes infinitely wiggly, and completely fills — or, perhaps better said, becomes — a plane. The pattern labelled 1a below, for example, is made up of infinite repetitions (within the bounds of the resolution of the picture) of the hook-like line segment labelled 1b. According to Hausdorff’s formulation, it has a dimensionality of 1.8687.

A line with a dimension of 1.8687

This idea of fractional dimensionality is a strange one, but seeing complex lines in terms of dimension rather than distance does allow us to dodge the paradox of infinite coastlines, restoring a mathematical order in which the coast of Britain is again longer than that of, say, the island where I live these days, Funen.

The most interesting fractal equations are not those like the ones shown above that simply repeat a pattern, but rather those that echo themselves but never exactly repeat, to infinity. The classic example is one of the first discovered by Mandelbrot himself, the appropriately named Mandelbrot set, consisting of all solutions to (z(n) ← z(n – 1)2 + c) which converge toward zero rather than expanding toward infinity. (Whilst trying to avoid getting too far down into the weeds here: this notation represents a potentially infinite series of iterations, in which n represents the current iteration. The result of the previous iteration — represented as z(n – 1) — is used as input for the next. The variable c is assumed to be a complex number, meaning it can be plotted onto the X and Y axes of a Cartesian coordinate system.)

The Mandelbrot set produces the most famous and immediately recognizable visual pattern in the field of fractals.

Mandelbrot set

At this point I’d like to give you a chance to explore the Mandelbrot set for yourself, via the little program embedded below. You can pan around by holding down the left mouse button and dragging in the desired direction; drill down deeper into the image by holding down the right button and dragging to select the region you’d like to magnify; and undo each step you’ve previously made by clicking the middle button. Notice how this infinite space is made up of similar patterns that are nevertheless never quite exactly the same.

Now for the caveats: as you zoom in to ever-greater levels of magnification, the picture begins to break down and lose detail. This is not a reflection of the pure mathematics of fractal geometry, but rather a byproduct of the computing reality of limited number precision and limited processing power, especially given that this is implemented in the relatively slow language of JavaScript. Any computerized exploration of fractals is inevitably bounded by such considerations, as well as by the resolution of the individual pixels that form the patterns. Eventually my program will not let you zoom in further at all, a concession to these realities; when you reach the point that your zooms don’t take anymore, you’ll have to zoom out by clicking the middle button to continue your explorations. Due to its processing demands, not to mention the lack of handy mouse buttons, I don’t recommend that you try to play with this program on your phones or tablets. (If these limitations smart, you might want to consider how far we’ve come; a typical 68000-based computer of the 1980s would have required hours to generate each new image.)

Update: as of January 10, 2023, the website which originally hosted this toy is no more. What you see below is an archived version. You can still play with it, but it may be a bit slow to load, and the source code is no longer available.

As useful and oddly beautiful as fractal lines like the ones produced by the Mandelbrot set can be, they’re only the beginning. Fractal planes are also possible, existing in some fractional limbo between two and three dimensions. Just as a plane is in this formulation an infinitely complicated line, a solid is an infinitely “wrinkled” plane. And just as fractal lines are perfect for representing coastlines and leaves, fractal planes can easily become mountains. The striking pictures shown below were computer-generated by Richard F. Voss in 1983 using only fractal equations.

Fractal mountains

It’s also possible — mathematically, that is — to go further, to produce fractals that strain toward a fourth (or higher) dimension. However, representing them becomes a problem given our three-dimensional world. We can only show a three-dimensional slice of these fractals. Doing so produces some strange shapes indeed, like the “dragons” of Alan Norton.

One of Alan Norton's "dragon" fractals

Benoit Mandelbrot published his magnum opus, The Fractal Geometry of Nature, in 1982; it still remains the definitive book on the subject. It’s a strange sort of mathematics text, in which Mandelbrot devotes at least as much space to philosophical digressions into the implications of his discoveries as he does to proofs and equations, and that much space again to lots of beautiful color slides of the fractals themselves. With the book’s publication, the theory of fractals became an example of a rare phenomenon indeed: a development in abstract higher mathematics that was taken up and trumpeted widely and excitedly by not just the likes of Scientific American (which devoted many, many articles to the subject) but also mainstream magazines, newspapers, even television broadcasts. As is all too typical of any new media plaything, fractals were hyped as useful for everything short of curing cancer — and I suspect that some wouldn’t put that past them either. Not only could physicists use them to understand the motions of molecules and biologists to understand the growth of plants, but some researchers claimed that fractal music was a possibility, while others claimed that they could help you get rich by revealing the “hidden patterns” that govern the stock market. In a 1984 article for Byte magazine, Peter R. Sørensen waxed effusive: “their uses range from physics, biology, and sociology, to art and even motion-picture scene simulation.” While I’m hardly qualified to speak to their uses in the former three scientific disciplines, I do feel fairly confident in claiming that they have had the greatest impact on those latter two, mushier categories, especially if we preface “art” with the word “computer” and allow computer games to slip in under that label.

Tellingly, Mandelbrot was employed at IBM’s Thomas J. Watson Research Center at the time he wrote The Fractal Geometry of Nature; all of the images found within it were generated by the computers there. Many fractal equations, including some of the most beautiful, aren’t complex at all in themselves, but the need to iterate through them so many times to produce their solution sets means that the science of fractals could exist only as a theory without the aid of computers. The first primitive computerized fractal visualizations had been produced by Robert W. Brooks and Peter Matelski in 1978, just three years after Mandelbrot first codified his theories. Virtually from that time forward fractals were inseparable from the computers needed to properly generate and study them.

Many people got their first glimpse of fractals before the hype started in earnest, in a remarkable sequence of computer-generated special effects included in 1982’s Star Trek II: The Wrath of Khan.


The creators of the sequence were the Lucasfilm Graphics Group, a collection of software and hardware engineers whom George Lucas had started putting together in 1979. These folks, whose numbers included such soon-to-be computer-graphics legends as Alvy Ray Smith, designed and built all of the hardware and software used to create sequences like this one. In 1986, the Graphics Group would be spun off to become an independent company whose name you might know: Pixar.

When Lucasfilm elected to start a computer-games division shortly after the sequence shown above was created, it was natural for it to draw inspiration, technologies, and even personnel from the Graphics Group. One of Lucasfilm’s first two games, Rescue on Fractalus!, proudly bears its graphical underpinnings and its status as the first game to make significant use of fractals right there in its title. (By way of continuing the theme, the alien enemies you fight are called the “Jaggies,” a reference to the ugly pixelized artifacts that are one of a computer artist’s worst enemies.) The game sparked great interest from its first press preview in May of 1984, thanks to its endlessly varied mountainous terrain, generated on the fly using fractal algorithms.


In the wake of Rescue on Fractalus! fractals were suddenly everywhere in the computing press; I’m not sure there was a single magazine that didn’t publish at least one big feature article on the subject over the next few years. The appeal of fractals to programmers was obvious, and had little to do with Mandelbrot’s esoteric philosophies about geometry and nature. Irregular, natural-looking landscapes had previously been dauntingly hard to craft for games — hard not only because they were a pain to draw but also and predominately because they were so un-amenable to compression algorithms and, indeed, to virtually all of the many innovative techniques programmers had discovered to store graphics data using minimal memory. Thus the distinctly regular, blocky graphics that were the norm, and the shock that was Rescue on Fractalus! Much like the Fibonacci-derived galaxies of Elite, fractals let programmers create whole landscapes generatively, from only an equation and a few seed numbers. They had their limitations, not least the amount of processing power that had to be allocated to generating them — Rescue on Fractalus! runs at all of 5 or 6 frames per second — but for many applications they seemed like magic. It was not a coincidence that after 1984 virtual landscapes started to become markedly richer and more natural.

But fractals did more than just make games prettier; they opened up whole new realms of possibility for them. The bit of video below, from a remarkable game that will be the topic of another article I’ll be publishing soon, may not look like much in contrast to some of what you’ve already seen today. Yet consider that, thanks to the magic of fractals, the planet being landed on is one of thousands to be topographically mapped in its entirety, and that you can land on it absolutely anywhere you like, zooming in from space to touch down where you will just as you can zoom and pan and explore the Mandelbrot set via the toy embedded into this article. Each of these unique worlds is generated using just a few numbers, a bare handful of bytes of precious memory.


Fractals aren’t quite the media darlings they once were; many other shiny objects have come and gone since the 1980s. And while they remain a valuable tool in many branches of science, they’re no longer viewed there either as the revolution they once were. You certainly don’t hear much anymore about fractal music or using fractals to play the stock market. Likewise, they’re now just another item in a game programmer’s bag of tricks. Still, they retain a fascination and beauty all their own. In that spirit, have fun in any further explorations you undertake, and if you discover any interesting patterns using my little toy above, or if you create any enhancements on the Studio Sketchpad site that hosts it, by all means let me know.

(As noted in the article proper, old computer magazines are an embarrassment of riches when it comes to information on fractals. Particularly good articles are found in the Byte of March 1984, September 1984, and June 1986; the 80 Microcomputing of December 1984; the Ahoy! of April 1987; the Amazing of March 1989, July 1989, October 1989, January 1990, June 1990, and April 1991; the Compute! of January 1983; and the A.N.A.L.O.G. of January 1986.

And since we’re in a multimedia sort of mood today, check out this song about the Mandelbrot set, sent to me by reader Rick Reynolds.)

 
 

Tags: , ,