Veteran game developers reveal their childhood creations

When Oliver Franzke was eight years old, his parents bought an East German knockoff of the Commodore 64 called the Kleincomputer 85/4. He started to learn the programming language BASIC on it in between sessions on a friend's C64 playing LucasArts' adventure game Zak McKracken. He recalls now that this pairing changed his life forever. He learned how to make BASIC draw lines, and he used those lines to illustrate his first text adventure game.

"It was just a room," he recalls, "and the player had to figure out how to leave it." If the player's text input didn't match any preprogrammed phrases, the game would reply, "I don't know what that means" in German. It was a simple game, but it was a start. Franzke found the experience "incredibly fascinating" — so much so, even, that he knew right then and there that he wanted to make games for a living.

For many game developers — like Franzke, who achieved his dream and now works as a lead programmer at Double Fine Productions — making games is in their blood. It's part of who they are, how they were shaped as an individual. And often it's something they began doing at an early age.

These are the stories of how more than a dozen game developers got their start in the business — of not their first step in professional game making but rather of the childhood passion or hobby that drew their interest years before they reached legal voting or drinking age.

The thrill of learning

Nowadays young and aspiring developers have a bevy of options for their tentative first steps as game creators, but from the mid-1970s through the early 2000s there was little other than BASIC — short for Beginner's All-purpose Symbolic Instruction Code. Its simple syntax had enough flexibility that kids could be creative and expressive and ambitious with their ideas — and discover the thrill of learning a new language along with all the other skills of game development. And, for at least a few of the developers included in this article, it enabled game making to become a passion.

Franzke continued making games in BASIC for a few years after that first attempt at a text adventure, then he got an IBM PC with a 256-color monitor. "This was around the time when Myst was huge," he recalls. "I was so blown away by its graphics that I started to learn how to create 3D graphics using POV-Ray. I quickly figured out how to combine primitive shapes — box, sphere, cylinder and so on — into more complex models." Before long, he ran into its limitations and stepped up to a 3D modeling and rendering program called Reflections, which made it possible for him to make more impressive graphics.

He took that as an opportunity to make a Myst-like game called Human Shock, which was about waking up in the near future to find the Earth mysteriously abandoned. "While I never finished the game, I implemented a pretty polished demo," Franzke says, "which included multiple puzzles [and] featured a main menu that let you load and save your progress."

He also started and abandoned efforts to make a real-time strategy game inspired by Command & Conquer and a top-down space shooter, then later put newfound knowledge of C++ and Microsoft's Direct3D graphics library to use on a first-person shooter demo. Franzke suspects he never finished any of these projects because his goal was to learn new skills rather than create a polished product. "Each new programming language — BASIC, Turbo Pascal, ASM, C++ — allowed me to do more things or to do them faster," he says. And each new art tool similarly allowed him to create better graphics.

Every time he learned a new skill or technique, he experienced a burst of creativity. "I just couldn't stop creating cool things and I loved it," he says. "The feeling of realizing your ideas in the form of a game (or part of a game) was (and is) very addictive."

Myst co-creator and Obduction writer/director Rand Miller learned of this creative thrill at a young age, too. He discovered computer games while he was in junior high when a family acquaintance took him to a computer lab at the local university to show him an IBM mainframe. Miller was instantly hooked by the magic of Lunar Lander, a text-based game in which the player had to decide how much fuel to burn every second in order to safely land a lunar module on the moon's surface.

He peppered his acquaintance with questions until he learned that computer games were "programmed" by people in special languages, the easiest of which was BASIC. Miller then bought a book on BASIC and set about developing a computerized version of tic-tac-toe. It would be played solo, human against computer. But the machine wasn't cooperating. Miller programmed it to output the state of the tic-tac-toe board at every turn, and he got player input working, but then he couldn't figure out how to tell the computer to make its move — how to make it think. It wasn't until he asked his family acquaintance for help that the teenager realized just how studious and thorough a programmer has to be; the computer wouldn't know how to make its move until he wrote the step-by-step procedure by which it could do so.

From then on, he rode over from his school every day to play and write games for a couple of hours in the university computer center using passwords swiped from students there.

Miller singles out two other early game development efforts as milestones, each inspired by other entertainment products. Starsky and Hutch, "a text-based race to drive a time bomb as far from the city center as possible," was based on the "Starsky and Hutch" TV show about an odd-couple pairing of streetwise cops who rid the streets of crime from their red and white Ford Torino. Swarms, for the PDP-10 minicomputer, was inspired by the book "The Swarm" by Arthur Herzog — which weaved a terrifying tale around the possibility of swarms of killer bees attacking North America.

"Swarms was an interesting leap for me, game mechanics-wise," he says. "It was the first time I added in a temporal element. There were various 'defenses' you could attempt to combat the spread of the bees. Some of them worked immediately but with only a small effect. But time-based defenses might take four or five 'turns' to see if they would work. The idea was to spread various defenses around to try to save as many lives as possible — but the 'time' element added an extra piece to juggle that worked really well. The complexity of the mechanics — effect, timing, collateral damage — made game balancing something I had to teach myself about."

Swarms was also Miller's first taste of professional game development. He was still in high school when it was accepted into Creative Computing magazine, its BASIC code reproduced in full in the May-June 1977 issue — alongside a note from Miller praising the Albuquerque public school system for its student-accessible use of a DEC-10 computer system, which he'd benefitted from in developing his game.

For Apogee Software/3D Realms founder Scott Miller (no relation to Rand), forward-thinking educators were instrumental in exposing him to computers and games. He was living in central Australia in 1975 when his high school got a Wang 2200, possibly the earliest commercially available preassembled personal computer. "At first, I didn't think it was interesting," he says. "I just saw it as a big calculator. But some of my friends had copies of a magazine called "Creative Computing" and they were typing in games written in BASIC."

"This was a revelation," he continues, "and I realized that computers allowed you to create worlds with your own rules. I was immediately hooked."

He was never able to get enough time on the computer, he says, so he started sneaking back into the school after hours via a window he unlocked before leaving. This allowed him to work late into the night — sometimes until midnight — on his own little games in BASIC. "The screen was 60 characters across and something like 12 or 15 lines high, and it was just a green CRT with nothing but ASCII text," he says. "So I made a tank that looked something like (###)== and then an enemy tank facing your direction on the same line 50 spaces away. You'd then type in a number between one and 1,000 and the game would print out the line with the two tanks again but show where your shell hit. Then the enemy take would take a shot at you."

"You kept taking turns until one of the tanks was hit and destroyed. Believe it or not, this game was a big hit with all my friends, and it was actually quite tense!"

The simplicity of the technology — particularly the capacity only for bare-bones or nonexistent graphics — gave Miller a gameplay-comes-first perspective that he retains to this day. "As a game maker back then," he recalls, "I had to focus on rules, on fairness, on interesting decisions that involved trade-offs for the player to weigh, and all of those background considerations."

Another of his games, Space War, was a two-dimensional space battle for a teletype display that he modelled after the Star Trek mainframe games. Every turn, the playfield had to be reprinted on paper. "I was quite proud of the AI in that game," he recalls. "Most players could not beat the AI-controlled spaceship."

Eventually the school discovered what he'd been doing when a cleaner walked in on him coding. The next day he was called into the principal's office. They had an inch-high stack of print-outs of the code from his games. His work was so impressive that they agreed to overlook his transgressions if he would help the computer teacher whenever needed and start coming in the front door for his extracurricular programming — using a key the principal slid across the desk.

Falling into games, then falling out

As with any other childhood passion or hobby, game development can capture a young mind whole and then suddenly vanish from the kid's life for months or years. That was just what happened to Sande Chen, whose credits include writing on The Witcher and design on many mobile and social games. Chen was in middle school when she first took an interest in making games. She was inspired by Infocom's many text adventures, choose-your-own-adventure books and role-playing games like Wizardry and Ultima. "I knew from an early age that I wanted to be a writer," she says. This was a natural extension of that, especially since she already knew how to code in BASIC.

She'd plan out a branching narrative and draw blueprint maps of locations, then use these notes as a reference point as she wrote if-then conditional statements to cater to player decisions. Chen can't recall the exact details of her first game, but she says the protagonist was a young boy. He'd been kidnapped or otherwise dropped unexpectedly into a medieval fantasy society. "After every block of text, the player could select one out of five choices to progress the story," she says.

"What I remember the most is that I had my brother play it through, and when he kept on getting bad endings, he didn't want to play anymore," Chen continues. "I realized that the game I made was too harsh because it was very easy for a player to fall into clever traps and get bad endings. So, the next time around, I began to think more about improving the player's experience."

Chen's game-making adventures didn't last much longer, however. While she continued to write and to learn computer programming, she soon lost interest in games — which she believes was common at the time among girls entering high school. The thought of working in games didn't come to her until after graduate school when a friend returned from a conference and said she'd met some people from a game company (Vicarious Visions, the first studio Chen worked with).

Dyad developer Shawn McGrath also dropped out of game making for a few years after starting early. His problem was less one of passion, however, and more of the difficulty of finding reliable collaborators when you're developing games as a hobby. "I stopped [at] around 16 because I couldn't draw and I couldn't get graphics into the game," he told Gamasutra in 2011. In a new interview, he says that he'd started out in competitive chess when he was four. He was exposed to game design as a discipline in second grade. He'd stay back from school a couple of times a week to play and discuss a Viking-themed, grid-based, chess-like game his teacher's boyfriend was developing.

McGrath later attempted to design his own chess-like games, but playing Tetris held more sway with him than game making (even as it planted a seed in his head that he could maybe be a game developer when he grew up). It wasn't until he was 12 or 13, when he started learning programming, that he really got into development as a hobby. "I would just try to clone games that I was playing at the time — top-down shooters and maze games, mostly," he recalls. "Nothing crazy. They were all kinda shit."

For a while he and Jonathan Mak (of Everyday Shooter and Sound Shapes fame) attended the same school, but Mak was of a far more advanced skill level at the time, and they became minor rivals because nobody else at their school was making games. That left McGrath with scant few potential collaborators, and his would-be artists kept flaking out. And that, ultimately, got McGrath out of the hobby until Sega's cult hit, the on-rails shooter Rez came out and showed him that he could maybe use computer-generated graphics instead of relying on artists. He'd been working for a couple of years as a professional programmer outside of games by that point, but his heart wasn't in it. He soon began the long journey to full-time indie game development.

Other developers, like SpyParty creator Chris Hecker, who previously designed and programmed the procedural animation system in Spore, never have that falling-out phase. They start early and never look back.

"The first game I wrote — or tried to write — was a clone of Tempest, my favorite game at the time, on Commodore 64," Hecker says. This was around 1982 or '83 when he would have been about 12 years old. "I had this program called SuperBASIC or UltraBASIC or something like that, and it added all kinds of nifty sprite routines to regular C64 BASIC." He likens it to a kind of game engine, like a Unity or Unreal for the Commodore 64. Except bad.

He didn't know any relevant math, so he hardcoded the lines that define the lanes of the simple flat-plane level at the beginning of the game — before it gets into curves and steps and more exotic 3D shapes. Then he added the player's claw-shaped spaceship. "[But] as soon as I tried to get the enemies in, Super/UltraBASIC ran out of steam and died," Hecker recalls.

He decided to take matters into his own hands after that. He set about teaching himself Assembly language, which is just one thin layer of abstraction above machine code (the language that a computer's CPU uses to actually execute tasks). "I guess I never looked back, since SpyParty is using all custom code and not an existing game engine," he says, "and people think I'm crazy for it — maybe it's just childhood trauma at its core!"

Starting young

While it's common for developers to first dabble in game making during their teenage years, some start much younger — nudged into it by an older sibling or a parent with an eye for technology. Spelltower and Typeshift creator Zach Gage credits his first step into game development to his mom, who bought the family's first computer, a Macintosh LC, in the early 1990s while he was still in kindergarten. "I was only allowed to buy two computer games," he says. "I bought Prince of Persia and Lemmings, and then when I got tired of those, I ended up just spending most of my time playing with Kid Pix and sort of using the stamps to design games."

(Kid Pix was a completely open-ended drawing and painting program for kids that included "wacky" brushes and customizable stamps.)

At this point, Gage's games were extremely simple. They were mazes or had some kind of world and characters crudely drawn out with the rules and systems entirely in his head. A few years later he discovered HyperStudio — a multimedia authoring tool similar to Apple's HyperCard software, which allowed users to create "stacks" of interactive, virtual cards. With this, he focused on making his own adventure games, layered with increasingly greater complexity — despite the absence of any scripting or programming languages — by using hyperlinks to create branching paths. One time, he made a game called Morzag in Time and Trouble with help from an older friend who would sometimes babysit him. Gage can't remember much about the game, but he knows he built an actual, physical box for it.

Dwarf Fortress co-creator Tarn Adams recalls a similar marriage of technology and creativity influencing his and his brother Zach's lives from an early age. "We've been working on game-like programs for as long as I can remember," says Adams. "Our father knew that programming and computers would be important, and he thought that games would be a good way to get us interested. It stuck a little too hard."

Sometimes the brothers worked on different things — Zach wrote adventures in the style of Infocom's games and Sierra's Quest series with his friends, while Tarn wrote "a bunch of strange ASCII animations and generators." But all of their larger projects — the ones that seemed really interesting — involved both brothers.

When Tarn was in fifth or sixth grade, they made a John Woo-inspired, text-based assassin game where you keep having to complete contract hits to pay for family accidents — first you need to pay for your girlfriend's corneal transplant after accidentally shooting her in the face, then you have to pay for treatment for various other family injuries (until a final stage where you need therapy for your dog, for reasons he can't recall). The game included eight factions who would put contracts on each other's members and also on you (and you had no say as to whether these attempts on your life succeeded). For another game, they asked their mom to scream into a microphone to get a sound effect to play when people are shot to death by Uzis. They also had a World War II strategy/arcade hybrid called Planes that included a simple, clumsy pathfinding algorithm for moving ships in the ocean. In the absence of the internet or any relevant books, Tarn devised the algorithm through brute-force self-reliance.

There was also an arcade-style space game with ships that had chunks of pixels blowing off them, along with a game called Cybercop that neither brother remembers but Adams suggests was likely less a "clean up the streets" cop game "so much as slowly degrade and get wounded and die a horrible death within five minutes" game. And they found backups of some 400 or so others they made in BASIC before they moved up to C in high school. "We don't even remember making half of them," Adams says. He also notes that some of these games have corrupted source code files because of frequent computer crashes — a notable example being Tarn's ASCII monster designer, almost all of which was overwritten during a computer crash by a memory dump of the code for an interactive fiction game that Zach had just started.

Levelling up to better games and tools

For all of its ease of use and versatility, there was only so much BASIC could do before it would run out of memory or crumble under the complexity of a sophisticated and ambitious game design. Sooner or later, these kids would inevitably have to move on to something more advanced in order to continue learning and growing as game makers.

For the Adams brothers, that breaking point was a BASIC game called Dragslay. It was a text-only fantasy role-playing game in which the player would fight a monster, get loot and swap equipment, then repeat. Sometimes, instead of a monster, you "let out your true self" and had to fight it. If you beat the 10th monster, a dragon, you'd advance to the next region (of five), which had different special attacks and harder monsters.

Tarn and Zach Adams

When they entered high school, Tarn and Zach remade Dragslay in C — now with a world map, goblin tribes and goblins that could mock the player for past failures. That led directly into Slaves of Armok: God of Blood, an isometric-view fantasy game they worked on just before Tarn started graduate school, which also got remade as their skills improved. Both versions of both games would later provide the groundwork for Dwarf Fortress, which added Civilization- and SimEarth-style procedural world generation to their fantasy and role-playing systems. "It became pretty clear there was a lot of potential, revisiting previously played worlds and so on," says Adams, "so we always felt compelled to work on it."

Gage also stepped up to more complicated tools and larger games as he got older. He thinks that both Kid Pix and HyperStudio were invaluable in teaching him to conceptualize ideas, execute on them and then give the finished product to somebody else to experience. But it was Cocoa, a kid-friendly visual programming language that Apple made in the 1990s, that established his practical game development skills before he moved on to actual coding in C++ and Java in high school.

With Cocoa, Gage remembers he made a space game called Escape from Pluto, which he put up for sale through a website. "I sold one copy," he laughs. "I was very psyched." Another game, which turned out to be too ambitious for Cocoa to handle, had a battle engine similar to Dragon Quest's.

At one point he read an article about a 13-year-old kid, Greg Miller, who'd made a game starring Cocoa's mascot called Wecman's Big Adventure. Miller had sold 100 copies at $10 a pop at the Macworld Expo and had over 1,000 downloads on AOL. "I emailed him and we ended up working together on the sequel," Gage recalls, "which was really exciting for me because I got to work with this person who had some success in games. I felt like I was working on a real game, which is a pretty amazing thing for a 10- or 11-year-old to feel like."

Around the same time, Apple sold Cocoa to a company called Stagecast. They were impressed by Gage's work, so he made a little game for them in which the player controls a UFO and tries to pick up all the cows on the screen. "That was included with their standard set of demos," he says. "They paid me to go to some science museum and basically demo it to kids and teach them how to make games."

He also contributed to a book about Cocoa, which led to a magazine article for which Gage had to do a photo shoot. "They put a bunch of wires on my face and I totally hated it," he recalls. "It made me very early on kind of be like, 'Oh man, I don't ever want to be famous. I don't want to be near magazines where I have to do photo shoots.'" But all the same, it encouraged him to keep going with games. It showed that people other than his mom and friends cared about what he was making — that there were people around the world to whom his games mattered.

Gage wasn't the only one to get attention thrust upon him for early achievement in games. Interplay founding member Rebecca Heineman was a teenager lying about her age when she started working as a professional programmer on Atari 2600 games and when she won the National Space Invaders Championship in 1980. Naughty Dog co-founders Jason Rubin and Andy Gavin were 15 years old when they sold a skiing game to the company Baudville for $250 and 16 when their graphic adventure Dream Zone sold 10,000 copies. (They started contracting for Electronic Arts the following year.)

Romero Games

Brenda Romero also stumbled into the industry as a 15-year-old. When she met a girl her age in her high school bathroom who worked at Wizardry developer Sir-Tech, she'd already been making her own board games for years, often taking the pieces from multiple existing games and figuring out a rule set to combine them together. The girl got her a part-time job there, at first doing whatever nondevelopment tasks were needed, then helping out on minor design and programming tasks as she honed her talents. Romero's husband John, one of the co-founders of id Software and a co-creator of Doom, likewise started young. Something of a child prodigy, he wrote code for the U.S. Air Force when he was just 15. His dream was to make games professionally, but he didn't think he had the skill or knowledge to match his idols in the industry — so for years he just sent his games off to magazines and publishers rather than applying for jobs.

Now both legends of the industry, they've helped their 12-year-old son Donovan make and release his first game, Gunman Taco Truck, on Steam, to what has so far been a positive reception.

Thinking like designers

Game design is a skill and a discipline, but also a mindset — a way of thinking about rules and systems that favors engagement over order. Often a passion for game development emerges side by side with an intrinsic understanding of this idea — that games combine audio, visuals, touch, rules and systems of interaction into an enjoyable experience. Eyram Tawia gained this understanding quickly as he wrapped his head around the concepts of code. He found programming through a book, the Secret Guide to Computers by Russ Walter, that he read at age 13. Now the co-founder and CEO of Ghana and Kenya-based games and comics company Leti Arts, Tawia recalls well how he stumbled at this early stage of technology discovery. Walter's book had included information on programming in the languages C, Logo, Pascal and QBASIC, but only the last of these loaded when he tried to use them from a DOS prompt. "I had no idea software needed to be installed before run," he explains.

"My first game was a semi text-based game for my comic book 'Sword of Sygos'," Tawia continues. "This game had a green map with a smiley that represented you and a dark-green box that represented where the sword was. You navigated by responding to prompts as to which direction to move (north, south, east, west) and you'd randomly meet enemies, which you'd fight by responding to prompts to attack or block. It was fun and spanned about 10 pages of code."

He then joined forces with two of his friends, whereupon they collaborated on lots more games. They made racing games, top-down shooters, fighting games, betting games, maze games, financial literacy games, 3D flight simulators "and a whole lot more" — all in QBASIC.

He singles out Jake, a maze game, as one of the most enjoyable to make, as it introduced him both to maze design and to artificial intelligence bots, which he made to chase the player around the mazes. "Another one was Crazy Speed 2000," Tawia continues. "I remember I woke up from a dream thinking about how I could manipulate sounds in QBASIC and tie it to a speedometer radar or a loading bar which could simulate an accelerator effect of a car."

Former Google chief game designer Noah Falstein, whose past roles include project lead on Sinistar and Indiana Jones and the Last Crusade, remembers that he was dabbling in game design years before he saw a video game for the first time. (Nolan Bushnell's Computer Space was his first exposure to the medium at age 15. A few years later, in 1975, he started making his own mainframe computer games at college.) "I remember playing with wind-up toy cars and repurposing them by adding cardboard 'skins' to make them into boats or spaceships at an early age," he says. He also made hundreds of elaborate "blueprints" of science fiction vehicles and spaceships as part of a "not very imaginative," (probably) Star Trek-inspired future sci-fi world in his head.

He soon developed a fondness for board games. In particular, ones with colorful plastic counters and props, and also 3M's Bookshelf Games series — which packaged family-friendly board and card games in boxes made to resemble leather-bound books. At 10 years old, he saw the film "Sink the Bismarck!" — a dramatic retelling of the events leading up to the destruction of the famous German battleship during World War II. "Not long after," he recalls, "I made a game out of cardboard, marbles and pennies to play with a friend." Much of the cardboard formed a mock version of the ship with different point values assigned to each of the compartments. A cardboard torpedo bomber had a slot through which a penny could be rolled, mimicking the flight of a torpedo, as the plane approached the Bismarck. To keep it challenging, players had to release the penny at least three feet from the ship.

"When I found it was boring to wait until the other player had done taking their turn," Falstein notes, "I added a defensive component: The second player could defend the Bismarck by rolling a marble perpendicular to the path of the penny to try to knock it out of the way."

"That proved to be too hard, so I made a chute with about five marbles so all of them could be released at once," he continues. "That made it possible, but still hard, to intercept the penny 'torpedo.' There was absolutely no logical or historical reason to do so; it was just fun."

Years later, at age 16, Falstein made a board game that he recalls had startling similarities to the popular Nintendo DS turn-based strategy title Advance Wars: Days of Ruin. The board was a map of four countries, one in each corner, their territories colored black, red, blue and green, respectively, with a large sea around an island in the center. Players could build factories that would produce military units, such as infantry, tanks, artillery, planes and ships, and vehicles had ammunition and fuel counters that depleted as they fired or moved. And there were spots on the map that could be mined for uranium, which enabled the production of nuclear bombs.

"You could win by capturing or destroying the enemy capitol or eliminating all forces," Falstein says. "I don't recall playing an entire game with friends, as it took many hours to complete, but mostly played all four sides on my own, which seems pretty sad in retrospect." (Falstein also notes with some regret that the board is long gone — he threw it out in 1995 while helping his dad move from his house and didn't think to photograph it first.)

Monkey Island co-creator and Earplay's chief creative officer Dave Grossman says that he also took "a couple of stabs" at game design around the time he was in junior high. "My friends and I played various kinds of paper and board games," he recalls, "and it seemed like a cool thing to try. How hard could it be?"

One time he tried to make a space game. "Each of the players was the captain of a star ship," he recalls. "They sat in different rooms from each other in big, padded swivel chairs, getting reports and issuing commands, James Kirk-style. In yet another room, I had a great big map on cardboard the size of a large kitchen table, where I could track the ships and space torpedoes and stuff like that."

Grossman went from room to room telling players what their sensors were picking up. It was a miserable experience, he says. The fight mechanics were "the opposite of fun" and the captains spent much of their time sitting alone in separate rooms struggling to locate one another's ships.

Grossman briefly dabbled in Commodore 64 game development, too. "I got a little Coca-Cola can bouncing around the screen," he recalls, "and you could drive it with a joystick, or maybe arrow keys, and I don't think I got much farther than that." The project didn't hold his interest long enough to finish.

Another time, he decided that he could improve the mainframe game Hunt the Wumpus, a text-only cave exploration game in which players endeavor to find and kill a monster called the Wumpus before it ate them or they fell in a pit. Grossman didn't like that running out of arrows meant the game was over. He felt that it should continue until someone dies, so in his version, you could fight the Wumpus with your bare hands. "I also made some new maps for the system of caves, including a random generation option to keep the exploration part fresh," he says. This was hardly reinventing the wheel, but it got him thinking like an editor — a skill he'd need time and again in his professional life.

Looking back

Grossman is quick to downplay the significance of his childhood game-making efforts. "The lessons learned I would have learned quickly no matter when I started designing," he argues. But he concedes that at least some of the ways in which he thinks about games — his personality as a game designer — were already present at this early stage. "How they got there, I have no idea," he says.

Noah Falstein, by contrast, believes that the time spent crafting his own games during his childhood were critical to preparing him for a career in the business. He singles out the large military game, in particular, for helping him develop an understanding of how to make games fun (or boring) and how to iterate on or modify rules and concepts to see if they improve the experience. It's not something he thought about consciously at the time — "I didn't analyze motivation or consider core loops at 10 in the 1960s," he notes — but these fun experiments were at the root of a lifelong fascination with game design.

Sande Chen similarly sees her early experiences writing text adventures as foundational on her path as a professional. "I took to branching narrative and nonlinearity easily because it was already in my way of thinking," she says. Oliver Franzke makes much the same assertion. By teaching himself programming and 3D graphics, he was able to specialize in computer graphics at college and then to work in the industry as a tools and graphics programmer. And by being able to speak the "language" of artists, after all the art he made for his own games, he found that he could bridge the gulf between departments at professional studios.

For Tarn and Zach Adams, there is no clear division between making games as kids and adults. "We've been going more or less nonstop since we started, so it just feels like one long arc," explains Tarn. "It's hard to say how much it matters in terms of developing good intuition and so on compared to how it would be if we had started in, say, our twenties. [But] it certainly feels like a lot of practice."

That practice was invaluable to Eyram Tawia, too, who says the key for him is to hold on to the passion he had for programming and games as a kid and to remember how special it felt at the time.

"I really don't want to lose that zeal," he says, "so I always make sure I code all my games with that excitement."