From Independent to Indie: Cave Story Revisited

1

Hunt the Wumpus (Gregory Yob, 1973^[1]) is a remarkable game.

You're dropped into a cave, in pursuit of the Wumpus, a great pungent suckerfooted monster who would gladly eat you if it could get close to you...and if it were awake. The cave is split up into a network of rooms; you are warned when the Wumpus or some other hazard (like a bat or a pit) is one room away, but not exactly which room it is in, which is a problem when you always have several possible rooms to move into at any point in time, and an even bigger problem when shooting an arrow into the darkness ahead will wake the Wumpus if it lands in the wrong room. So a good deal of deductive reasoning—and a little bit of luck—is needed to form a mental map of where the hazards are, while keeping a safe distance from any of them, probing on, noting where you get a warning, recording which rooms are safe and which are adjacent to danger, and deciding when it is—or isn't—safe to probe just one more room ahead...

For a game written in 1973, it holds up remarkably well. Games are short but tense, and the mental calculus involved in each game gets more engrossing with each room safely entered, as the likelihood of the next room containing a hazard—or a Wumpus—approaches 1. Yet the random assignment of hazards to rooms allows the game to maintain tension even after you've managed to get your head around it.

It's also historically remarkable^{†Cogger 2005}, as it establishes a couple key conventions of the text adventure game before that genre would be established a couple years later. Think about it: The machine narrates your progress in first person and present tense, as you navigate a hyperlinked map and solve puzzles, albeit with a protracted goal (hunt the Wumpus) and a protracted set of verbs (hunt the Wumpus). The map is even a cave! A cave full of pits! Not a colossal cave, but we'd get there eventually.

Hunt the Wumpus is remarkable most of all because its program takes up a single page of text. Here is that page^{†Yob 1976}. If you can dig up a BASIC interpreter, or more likely set up an emulator for an old computer that had BASIC in its boot ROM (this is the hard part; to make it easier, try this TRS-80 emulator,) then you can punch the game in and play it right now. You don't even need BASIC, if you don't care for it; if you're just learning to program, it's a worthwhile exercise to translate the game into your language of choice (many examples exist). In fact, you don't even really need a computer, just another person to keep track of the locations of the hazards, to hide them from you, and to move them around according to the program and your choices.

And this last point about Hunt the Wumpus is itself remarkable, for the analysis of another cave story soon to follow, because in 1973 this did not make it so.

2

In the 1970s, most computer games were small, and most computer games were made by either an individual or a handful of people. In the period stretching from the birth of the medium in the mid-20th Century^[2] to around sometime in the '80s, "independent game development" would not have been a meaningful term, as the term today is an intuitive shorthand for small-scale production, and the majority of videogames then were made by individuals. There were only so many Ataris, Konamis, Sierras and Nintendos to go around, but there were lots and lots of amateurs and upstarts submitting lots and lots of listings to lots and lots of magazines and books. Thumb through the pages of the People's Computer Company (pick one) or Micom BASIC Magazine (no, really, pick one!) or even Digital Equipment Corporation's 101 BASIC Computer Games^{†Ahl 1973} and you'll find evidence of a thriving hobbyist gamemaking scene with a low barrier to entry (many of the games within those pages were made by kids, especially those in DEC's gamebook). And even when a game was produced with the resources of a large company—those resources including access to flashy technology and tools^{†Video Games Densetsu 2018}—it was still typically made by an individual or a small group. Just one person made 1981's Tempest; just two people made 1980's Missile Command. Even The Legend of Zelda was made by just 7 people—6 if you don't count executive producer Hiroshi Yamauchi, who was president of Nintendo at the time and was credited on all games by default.

So what changed? Hardware changed. For one thing, it improved. Better hardware meant you could write more elaborate games. Or, to put it another way, more elaborate games required better hardware—storage devices that could hold more code, processors that could execute more code. But more elaborate games also required more resources—time, or skill, or people, or expensive digital tools—to fill all those empty kilobytes, then megabytes, then gigabytes. Only a sliver of individual programmers or software houses had the resources to leverage that power in a way anyone would care about, and as those games became more complex, players' expectations for the complexity of future games, particularly retail games but ultimately all games, went up for everyone, everywhere, so that sliver got smaller with each cycle of Moore's Law.

3d graphics, whether they were done on a Voodoo card or a Super FX chip or by a humble software renderer^[3], accelerated this process. It's no surprise that Yūichi Toyama, who began programming in MSX assembly as a child and went on to write Herzog, the first real-time strategy game, in 1988 as an adult, completely fell off of game programming when 3d graphics emerged^{†Ka 2021}—they're inherently more complex than 2d graphics, especially the mathematics involved in scaling and rotation, to say nothing of the 3d collision detection that comes with them.

the way gamemakers talked to that graphical hardware would change, too. in the '60s, the TX-0 could plot a point on the screen with a single instruction (MIT c. 1960); in the '80s, microcomputers had VRAM or display lists at fixed addresses to PEEK and POKE; but more sophisticated operating systems, in which several programs could run at once in separate windows, and in which memory protection meant VRAM and I/O devices were not directly accessible, required you to navigate a graphics API. In fairness, memory protection and windowing systems are actually rather old technologies, if you include implementations running on the large, timeshared university and business "minicomputers" (miniature only compared to the colossal mainframes for which they were a relatively compact, inexpensive alternative) that predated the desktop microcomputer, which took a while after its introduction to catch up to the state of the art. But on the other hand, there wasn't as strong an expectation that every game be graphical back then, and in fact several important works of interactive fiction were made during the minicomputer era, as the genre was still mainstream (well, as mainstream as you could get before Commodore, Apple and Radio Shack made computers themselves mainstream). So player expectations still turn out to be critical in the end.

As this complexity mounted from all fronts, so too did the technical skill floor for game production. Yet this did not immediately become unmanageable, and in particular the period from the '80s to the mid-'90s—the height of the shareware era—was a hotbed of amateur game production. After all, 3d graphics were still more novelty than norm.

And then DOOM happened.

3

It's obvious on the surface that DOOM (id Software, 1993) is a historically signficant videogame. It was to first-person shooters what Super Mario Bros. was to platformers (note, I'm sure I'm not the first one to make that analogy, or the second, or the third,) and is the punky, gory aesthetic predecessor of every other AAA action game. But I don't think it's fully appreciated how historically significant it is.

What does that mean? Well, what do I mean when I say that a game is "historically signficant" in the first place? One widely-held, intuitive idea of a game's signficance to the history of its medium centers around formal influence, that is, formal concepts invented (or, failing that, popularized) by a game that were then copied by other games. How many such concepts? How many such games? Did it invent a whole genre? Was it a popular genre? Did it invent a mechanic that transcended genres? Did it combine existing concepts into something more than the sum of their parts?

By this logic, Colossal Cave Adventure would obviously be considered historically significant, being the common ancestor of all modern interactive fiction (though technically not the first work of its kind). Grand Theft Auto III would likely also be considered signficant for codifying the now-ubiquitous open-world sandbox game; Herzog could take its place in history as the progenitor of the real-time strategy game; and even something like Battlezone could claim minor credit for its technical achievements, being the first arcade game to use 3d graphics. Videogame history, taken this way, becomes a unending string of famous firsts.

But another way for a game to be significant is for it to have ripple effects on the way games are made and played. Its innovations could be methodological, for example, or it could affect the way the public engages with games as cultural artefacts, or the material conditions under which future games are made. Certainly, many of the developments that affect the context in which games are produced have nothing to do with individual games. The emergence of the microcomputer was one example. The emergence of the hobbyist computer magazine and of online payment services, as we'll see, are two other examples. But sometimes a game sets a strong enough example or so dramatically changes player and gamemaker expectations that it can claim the same level of significance as a wider structural shift (the thesis I am working toward is that Cave Story is one such game).

DOOM is significant in all four ways described. Its influence on the FPS has already been mentioned, and in terms of methodology, its approaches to modding, portability (DOOM famously runs on anything) and low-level game structure (DOOM was the origin of the term "game engine," and all that term implied ^{†Wrist 2018}) have become de rigeur for other gamemakers since. furthermore, the controversy raised by its violence and satanic imagery, in addition to other violent games of the period like Mortal Kombat and Night Trap (which wasn't even that violent,) would precipitate a defensive shift in gamer self-identity whose diffuse fallout we're still dealing with today, for better and for worse.

As for its effect on the conditions of game production (a highfalutin phrase the account to follow should elucidate through use,) that comes down to the domino effect it had on shareware.

Richard Moss gives a meticulously-researched, comprehensive account of the history of the shareware scene in Shareware Heroes: The Renegades who Redefined Gaming at the Dawn of the Internet (2022).^{†Moss 2022} The account is generally upbeat and celebratory, though it was hard for me not to read it as an autopsy, as analyses of landmark games and gamemaker success stories were interspersed with hints of the scene's gradual professionalization and economic saturation. DOOM gets its own chapter, and just as well; DOOM was the biggest thing to ever hit the shareware scene. It was so big, in fact, that the scene would implode under its weight.

Not at first. At first, DOOM was a slice of heaven on a magnetic silver platter. A slice of hell, if you like. Any way you slice it, it was an overnight success. Moss writes that BBSs and FTP servers around America crashed under the immense load of hundreds of thousands of people clamouring to download the game on day one. (p. 209) The traffic wasn't one-way, either; player-designed WADs would be uploaded almost as soon as the game came out, and packets of deathmatch data would ping-pong across streets, cities, continents, in every direction.

But the most significant effect of DOOM, for our discussion, and this was no doubt due as much to how comfortably it controlled as to how pretty it looked, was that a lot of people with a lot of money and a lot of influence decided that all blockbuster games really ought to be 3d from now on. Writes Moss:

Now anything that wasn't a 3D shooter was old hat. Everyone was either talking about it or playing it (or likely both), from the low-level workers at games publishers and big software companies to the elite developers working on the next big commercial games, all the way up to the top executives at the biggest games and technology companies—many of whom spent the next Computer Game Developers Conference discussing how they should respond to DOOM's success with their own marketing and product development strategies.^[4] (p. 210)

Any game that wanted to compete with DOOM in particular had to be 3d, and since DOOM was the upcoming industry leader, the new state of the art, it followed that anyone who could afford to do 3d, had to do 3d. This meant all the big industry game houses, and anyone outside the industry with the brains—or the resources—to keep up. You have a NeXT workstation too, don't you?

Think about it. Media industries had been adopting 3d computer graphics in many little steps throughout the '80s and early '90s, but the period from 1993 to 1996 saw a slew of landmarks in quick succession. Consider the following.

• 1993: The real-time 3d videogame triple threat—DOOM on PC, STAR FOX on console, Virtua Fighter in arcades (which would famously cause a minor crisis of faith in noted 2d animator hideaki anno, when he first enountered it). And in theaters, Jurassic Park.
• 1994: ReBoot, the first computer-animated TV show. The first spec for VRML, bringing real-time 3d graphics to the World Wide Web. Donkey Kong Country on console, Killer Instinct in arcades, and Myst on computer all set high water marks for the use of prerendered graphics in videogames, with the Sony PlayStation pushing polygons the same year.
• 1995: in the US, Toy Story is the first computer-animated feature film, and the 3dfx Voodoo is the first desktop GPU to gain widespread adoption. In Japan, the Sega Saturn appears, and Kyoko Date becomes the first virtual idol. And around the world, Active Worlds, the VRML-driven virtual reality network, goes online for the first time.
• 1996: the Nintendo 64 launches with Super Mario 64, the first game to really get 3d movement right in third person, with help from the 64's pioneering control stick and a camera that could be freely pivoted around the avatar, zoomed in or out, and fixed in place. Tomb Raider's Lara Croft, an accidental virtual idol, blurs the cultural line between character and flesh-and-blood celebrity, taking a slew of interviews in character in the ensuing years, and even posing in the 1999 issue of Bikini magazine (alongside the likes of noted flesh-and-bloods Rebecca Gayheart and Malia Jones,) to say nothing of the Lucozade partnership.^{†Parkinson 1999}

And that's not even everything. It didn't have to be DOOM that did it, really. id had managed to turn the clock forward a few years, achieving the appearance of 3d on hardware yet unsuited to contemporary polygon rendering techniques. But the technology was going to get there at some point. Sooner or later, something was going to land the decisive blow, and if it hadn't been DOOM, it would've been some other game. It could've been Duke Nukem 3D. It could've been System Shock (that one wasn't shareware, but it could've had the same effect on player and gamemaker expectations). But in our timeline, it was DOOM.

It was all downhill from there. The technical constraints set by DOOM would be exacerbated when graphics cards became widely available a couple years later, and CD-ROM drives became standard PC hardware, especially to hold the (relatively) high-resolution textures mapped to the polygons pumped out by those cards. Shareware authors would continue to appear for a time, but they'd lose relevance with each passing year. You'd think the CD-ROM would work in their favor, as it allowed for hundreds of shareware programs to be collected onto one disc, instead of just a few per floppy (moss, p. 40,) but it was as difficult for any one author to stand out on such a compilation as it was for any of them to make any money off of them. (p. 74) By the late '90s, the shareware market would be increasingly overexposed and oversaturated. (p. 243) The last shareware game to be widely covered in the games media, according to moss, would be Tread Marks in 2000 (p. 293). And by 2004, game production had totally bifurcated.

The period between these years would mark a relative lull in amateur game development. Sort of. Let's put a pin in that. Let's mark that one with an asterisk.

The conclusions reached by Moss are slightly off. He argues correctly that the reason we don't hear shareware used by that name to describe games anymore is that the term's industry equivalent, "demos", would be used instead (and this was around the time "demo discs" became a popular industry practice) for the practice of presenting a free, self-contained game segment to promote a full paid version. (p. 293) But the issue is that he suggests we began using the term "indie" to describe the kinds of people "shareware" used to describe simply because it was more accurate to define such people by their relationship to the industry than by a particular distribution method that not all of them used anyway. (p. 293) While such people are certainly better described as "independents" or "amateurs" than "shareware authors," the term "indie" did not emerge out of the blue one day for clarity's sake. Instead, it belongs to a very specific social movement with architects, manifestos, and well-defined goals. The first wave of "indies" were not the same people who were writing shareware games a decade earlier, although they would create games for the same platforms and in several of the same genres.

It's simple. Shareware authors weren't "indies," although some of them were independent. Some shareware authors and some indies were amateurs. Not all independents were indies, but all amateurs were independents. Still with me? No? Of course not. I lied. it's not really simple. At this point, to head off any further confusion, we should define our terms.

4

Social Energy

Any social practice needs energy to support it, and that energy comes in the form of people. Theatergoing survives because there are people willing to enter movie theaters. Terminal-based videogames have largely disappeared outside of programming tutorials because no one wants to play them, though exceptions, like Dwarf Fortress and Nethack, certainly exist. A chicken-and-egg problem emerges if we take social energy as the only relevant factor in a social practice's existence; gladiator matches did not cease simply because people got bored of them. But it is a critical factor nonetheless; if people did get bored of gladiator matches, no one could afford to hold them—or, for that matter, to train for them. The amount of energy needed to support a practice varies; part of the reason terminal games survive at all is because they take so few resources to make that you don't need a lot of people participating in the practice to support it, as opposed to the cost-intensive movie theater, which needs a steady flow of attendees daily to keep its lights on.

Just because you can afford to make the infrastructure to support a social practice (and, by implication, a social scene centered around that practice) doesn't mean that people will magically appear around it. Some kinds of infrastructure, however, are important not just to facilitating a scene but to sustaining the flow of people into and through it. A scene needs places for its members to gather; it needs channels for communication; it needs accessible repositories of its collective knowledge; and it needs motivations, intrinsic or extrinsic, for people to participate. That is, sources of interest. Interest is not enough if the former elements aren't there; and although the former elements can sometimes help generate interest, they can't by themselves make up for its absence.

6

If you're reading this, you probably already know what Cave Story is. If you don't, what matters at this point is that Cave Story is an action-adventure game originally made in 2004 by Pixel. You find yourself in a cave. You try to escape the cave. Along the way you meet, help, fight, befriend, and otherwise mingle with a colorful cast of cave dwellers, including a friendly robot named Curly Brace, a not-so-friendly bar of soap named Balrog (huzzah!), and the scattered members of a former human expedition into the caves, who have discovered something disturbing at their center, something dangerous, something that should never leave those caves, but that, thanks to one of them, is about to—at which point your own escape becomes second-priority.

Cave Story is a Metroidvania with run-and-gun gameplay reminscent of a super-streamlined Metal Slug^[7]. Your weapons are pared down to the bare minimum needed to cover the maximum situations, and half the fun of using them is discovering new ways to apply them. The bubbler, technically a "hidden" item that everyone learns from the grapevine how to find, is a prime example of this, working as a shield, a wide-range shotgun, a machine gun, and a means of shooting around corners—and those are just the uses I myself could find. There's also Curly's machine gun, which she can optionally lend you after you meet her. While it has a narrower range than the bubbler, aiming downwards allows it to function as a jetpack. It's a testament to how carefully the game was balanced that this ability to fly freely through stages keeps the game from becoming too easy, while keeping the game from becoming too hard without it. It's also balanced against the game's dedicated jetpack, the booster, which offers a shorter thrust that is faster and that can be aimed in four directions.

But to stop at what the weapons do is to miss the game's core mechanic, one yet to be seriously revisited by later designers: the weapon energy system. Besides health, enemies drop golden, glittering triangular crystals, likened by fans to Doritos. collecting one adds "weapon energy" to a gauge tied to your currently equipped weapon, and when that gauge tops out, your weapon levels up, becoming stronger and gaining new features, up to the max level 3. But if you take damage, then you lose weapon energy, and when the gauge bottoms out, you level down, until the minimum level 1. In practice, you'll often be maxed out in easy sections, or while backtracking across areas you're already familiar with. Meanwhile, there's a special kind of panic that sets in during a boss fight, when you see the "Level Down" message appear, and have to switch from spamming to strategizing, forcing you to switch to one of your other, still-maxed weapons to keep your defenses up. Furthermore, bosses typically fire off objects that can be shot at to get crystals, so you're compelled to juggle damaging the boss with restoring enough of your lost weapon energy that you're in a position to fight at all. The overall effect is that you can't rely on any one weapon too heavily, even if it seems optimal for the boss; leveling down forces you to think creatively about how you can apply the rest of your arsenal.

And as an added twist, leveling up a weapon doesn't always make it better. Some will argue that the blade is more powerful at level 2, when it does a consistent amount of damage to a single target, then at level 3, when it does a variable amount of damage that can spread to other, nearby targets, even if the cap on the latter's total damage output is higher—and both shots travel a shorter distance than at level 1. Then there's the nemesis, a special late-game weapon, arguably the most effective in the game besides the ammo-limited super missiles, that actually becomes weaker as it levels up, so that in exchange for added firepower, it turns the energy chips into hazards to avoid, and even prods you into getting hurt on purpose to level it back down again. This makes every boss faster to kill but, for the reasons mentioned above, makes every boss riskier.

On the one hand, the weapon energy system adds excitement and pleasurable unpredictability to bosses and minibosses, and recovering lost energy isn't too tedious once you've topped a weapon out. On the other hand, it leads to a period whenever you first obtain a new weapon where you're gradually grinding it up to max level in an easy area, which stops you awkwardly in your tracks. Where the mechanic really shines is in areas where your weapons are suddenly dropped to level 1 and you're unable to backtrack or even reload the area at all by going in and out of a door. (In other words, in the last couple sections of the game). Now every bit of energy you collect becomes precious, as you carefully consider which weapons you want to get to which levels—you can't level all of them up to max—in order to get yourself in optimal shape for the next boss. All of this is to say that the system is somewhat hamstrung by the structure of the rest of the game, which beckons you to grind.

That structure, however, has benefits that make occasional grinding an acceptable trade-off. Cave Story is a "Metroidvania", but in a looser sense than that's usually meant. A Metroidvania tends to have a pretty consistent core gamplay loop involving colecting upgrades and using those to reach new areas containing new upgrades, etc., with some bosses in between. Whereas your objectives in Cave Story are looser, and can be summed up as "whatever is best for where the story is headed right now." That might mean hunting down a unique enemy. It might mean scouring an underground desert for lost puppies. It might mean donning a disguise to pry intel out of characters who otherwise wouldn't give you the time of day. Or it might just mean getting from one end of a stage to the other. it could be argued that part of the reason Metroidvanias are so prominent in indie game development, besides the influence of Cave Story itself, is that it is the easiest genre of action game to weave a story into, since it doesn't employ the same rigid separation of "where the gameplay happens" and "where the story happens" as do action games with a more regular stage structure and more fixed boundaries between stages.

The teleporter from the Plantation to Grasstown, for example, breaks the game's hub-and-spoke map design for no reason other then to tie the plot together. Oh, so that's how Kazuma escaped. That's why he wound up where he did at the beginning of the game. But it speaks to the next point that there is added dialogue for one of the NPCs, in the event you decide to see if anything has changed, because the game expects you to check.

Cave Story was a game that took a long time to make, and it's one that expects you to take a long time turning over with your controller (or keyboard). A lot of that time was spent adding easter eggs and hidden features and optional events in every place there was room to add them, and the game is designed around a hypothetical player who goes everywhere, speaks to everyone, and investigates everything at least once. At least. This means that the game, on dropping you in a new area, often doesn't tell you what to look for. It certainly doesn't place a blinking beacon on your minimap (besides the beacon representing your own position). This can catch off-guard players who aren't used to having to take their own notes and mark off places they've already been in their minds, and it can make reaching the true ending without a guide frustrating, especially since the main route will guide you from place to place almost without you thinking about it, with only the Plantation requiring some thought about where to go or what to do. It's the least "modern" aspect of a game that otherwise shares more in common with games that would follow it than games preceding it (in the US; see below,) and part of the challenge is that, in general, "check everything" styles of puzzle-solving work best when the gameworld is laid out on a grid (think most top-down NES games) or a tree (the menus of a visual novel or many an old-school japanese adventure game,) and you can reason about which cells or nodes you've checked off, rather than the tile-based but wider and more naturalistically sloping environments of CAVE STORY.

But the satisfaction you get from chancing upon a critical hint left on a computer in a remote cabin, or on a bookshelf in a makeshift shelter, is impossible to recreate any other way. The intended effect can be summed up with a quote from the Gun Hermit, whose prototype Polar Star you "borrow" while he's sleeping as your first weapon, and optionally revisit late in the game:

In this world, there exists a balance between those who are creators and those who are users. I knew that, of course, but it took your help for me to experience this firsthand. From now on, I vow to dedicate myself to the side of creation. The labor involved becomes joy when I know there are those who will enjoy my work to the utmost.

	00	02	04	16
	01	03	05	17
	02	00	07	19
	03	01	06	18
	04	06	08	00
	05	07	09	01
	06	04	11	03
	07	05	10	02
	08	10	12	04
	09	11	13	05
	10	08	15	07
	11	09	14	06
	12	14	16	08
	13	15	17	09
	14	12	19	11
	15	13	18	10
	16	18	00	12
	17	19	01	13
	18	16	03	15
	19	17	02	14
	      

	// the function below produced the table above
	dodec(u,a) =:: u ^ a << 1, if a = 0 or 1;
	               ((u ^ (u & 2) >> 1)
		        + 4
		        + (12 * (a & 1)))
		       % 20, if a = 2 or 3.
	      

7

Gameplay aside, the artwork and music were a major part of the game's appeal when it was new. They were also an example of the game's doujin context being lost in translation. While indie devs would often use pixel art and chiptune to call back to specific "retro" gaming eras, and Japanese games sometimes did this too, the pixel art in Cave Story was in a style that couldn't be tied to any specific era, because it was pixel art for its own sake. Pixel was a pixel artist prior to Cave Story, going back to the mid-'90s, when the style was still "contemporary" rather than "retro". If we include the earliest work attributed to him, In Search of Spring, made in shmup construction kit Dezaemon, his experience with pixel art goes all the way back to the SNES in 1995. Though Pixel mentions being inspired by Super Metroid in designing Cave Story in a GDC talk^{†Pixel 2011}, tying into stock indie narratives about using one's earliest gaming experiences as a personal north star, such a narrative is complicated when you remember that, when Cave Story began development in 1999, Super Metroid was only 5 years old and one console generation removed. It would be less accurate to characterize him as the archetypal indie dev drawing on an era he had only ever encountered as a player in his childhood, than as a member of the old guard himself, who emerged around the boundary point between two eras, and held on to the techniques of his own era as the world changed around him. He's paraphrased in this video as saying that whatever kind of trends might happen to come up, he says he's gonna keep making these games from here on out[...]^†

This is signficant, because the indie ethos would eventually get exported back into Japan, and have the effect of changing the way many Japanese amateur gamemakers viewed themselves, ultimately splitting Japanese amateur game production into "doujin" and "indie" camps ^{†Grau de Pablos 2024}, defined, among other things, by two different orientations to the past; the latter positions themselves as recovering lost traditions, the former as conserving and building on traditions that had never faded to begin with.

You can see shades of this, if you take a closer look at the game's art. Although backgrounds are tile-based, sprite art is not. Nor does Cave Story emulate the palette limitations of any particular 8-bit or 16-bit system past; instead, 16-color palettes are carefully chosen from a modern 24-bit color space. Vivid hues are combined with delicate, brush-like shades, soft textures with sharp highlights, fine mechanical and xenoanatomical details with hyperflat faces and expressions, Sanrio critters with Kanada thunderbolts and Itano missile scrambles, monsters with smiles full of skulls, machines with scanner-eyes like jewels. I sometimes open up the game's spritesheets and just gaze at them because I love the pixel art so much.

Someday, in a future where the last computer has been broken beyond operation, some gray-haired geezer is going to be found in a grotto somewhere, carving the shapes of King and Misery into the earth, in little dots of varying depth to emulate shade, with pigments of berry and bugblood to emulate hue. And as onlookers scratch their heads the geezer will laugh like mad and gesticulate like mad and say, abortively but no less earnestly, "They shone. The beam ran across the surface of the glass and it made them shine! Just the right shade of yellow, just the right amount of space between the little dot-eyes. Decades of techniques and half-dreams getting slung around over the wires, and sedimenting in living minds, and converging on that very spot, a tribute to, to what? A terrible time, I know, I know. The old days of king and misery, ha! I'm lucky to have survived them. But by Lain, you just had to be there. This thing in the ground, it's nothing. It's not the same. You just had to be there. It almost made it all worthwhile."

The music is also distinct from any particular platform, having been made in a custom sound driver called Organya, using a custom format called Org. The resulting sound is at once rigid and warm, with a distinctive buzz-drone sawtooth waveform as a common leading instrument, like a floppotron or PC speaker but less harsh, and makes a versatile instrument, from oppressive boss themes to suitably gestative, cavernous ambiences. Though more historically significant than tone or style was an emphasis on strong melody that would become a mainstay of indie game music going forward.

In general, there's a tendency, when it comes to talking about why Japanese games are "like that" (whatever "that" happens to be in a given conversation,) that underemphasizes the specific history of the field, and overemphasizes cultural essence. This is not a new tendency; commentators in the '00s^{†Kalata 2007} analyzing why Japan hadn't picked up FPSes and dropped its fondness for 2d games tied these trends to quirks of Japanese culture, rather than the reality that, for most of gaming history, these regions had played their games on incompatible hardware and therefore only rarely communicated directly with each other, leading to a kind of marsupial effect compounded by the fact that the language barrier left local schools of game design analysis and theory likewise region-locked. To be sure, offshore developers did get a taste of Japanese console and arcade gaming through Nintendo, Sega, et. al., but this was a one-way relationship—Japan didn't import offshore consoles to the same degree, because its console market was already healthy enough, and the offshore console market, originally led by the US, would not truly recover from the US crash until the Xbox, after the turn of the millennium (and even the first Xbox landed with a thud in Japan!).

The use of so-called "retro aesthetics" by doujin gamemakers in particular have to be placed into context, since they're more than meets the eye (or ear). The japanese independent gamemaking scene had largely skipped over the technical arms race led in the states by id, Epic, etc., instead honing mastery of bread-and-butter techniques like pixel art and real-time synthesized music (aka chiptune), and working within established genres like RPGs, ADVs (adventure game), as well as arcade staples like STGs (shooting games, or shoot-'em-ups,) FTGs (fighting games,) and the like.^[8] As a result of the accumulated expertise of the doujin scene, their work was, by and large, competitive with that of the industry at large, within the genres present.

That such expertise had accumulated is not in question. the testimony of the earliest indies to encounter it speaks to this. In a tweet, Kayin, designer of I Wanna Be the Guy: The Movie: The Game, says of the Touhou Project:

When us western indie devs were realizing things were possible lol ZUN was already laps ahead of us, like a space civilization watching earth from a distance. We didn't even realize HOW popular it was until image sharing sites started popping up. ^{†Kayin 2024}

but how and why that expertise had accumulated is not as clear. F_T_B, tracing the history of doujin soft ^{†F_T_B 2025}, cites the development of Cho Ren Sha 68K (Famibe no Yosshin, 1998) as a turning point. In particular, he cites the analytical deconstruction and reconstruction of the STG genre, a process detailed in development docs posted to his website and later translated for Shmuplations^{†Famibe no Yosshin 2014}. These docs show yosshin breaking the STG down into its atomic components, design decisions that are mutually exclusive, selection of those components with an eye to their psychological effect on the player, and recomposition into a new, original, and meticulously balanced design, logging the process publicly on a web-based BBS^[9] and folding feedback from players of early versions, distributed at Comiket, into later versions. it was a technique that would catch on immediately; F_T_B notes the influence of the game in titles like Dive On and Hellhound.

There are other factors that may have contributed to the level of design sophistication doujin soft had reached by the turn of the millennium. The early and widespread adoption of the World Wide Web as a form of knowledge exchange by doujin gamemakers, particulary the BBSes used by yosshin and, later, Pixel himself (both in Cave Story and in his 2007 shmup Guxt,) would have had an effect analogous to the effect the emergence of websites like TIGSource had on early indie game production, allowing insights to spread rapidly through a scene, and allowing for relatively fast feedback on works in progress. Only this was over a decade eariler.

It also tracks with similar developments that were happening in broader Japanese gaming fandom outside of doujin shmups, such as the development of programs to extract and analyze the contents of visual novels, and the growing adoption of game engines^[10] by amateur gamemakers for full-size products, lowering the barriers to entry and increasing the speed at which ideas could be experimented with, iterated on, and turned into finished games. These engines included the aforementioned Dezaemon^†, RPG Tsukūru Dante98 for RPGs and general top-down adventure games, known offshore today as RPG Maker, and responsible for such games as Azusa 999 and Peret Em Heru: For the Prisoners; VN engines like KiriKiri and NScripter, the latter of which was used to create doujin darling Tsukihime. lessons learned from the development of these games would affect even those building from the ground up, like Pixel did for Cave Story. In an interview with Derek Yu (who we will return to), Pixel would in fact go on to mention that his biggest regret in developing Cave Story was not taking more time to write his own utilities up front.^{†Yu 2005}

Besides, to some extent, the Yosshin technique was in the water. The fan culture in which doujin soft and the broader doujin field was situated had been becoming increasingly deconstructive, analytical, and self-referential since the 1980s. Those within it in the mid-'90s could point to countless accumulated examples of deconstructed anime, character tropes, costumes, visual effects, as well as works, both professional like DiGiCharat, Neon Genesis Evangelion (along with a decent amount of Gainax's output in general,) and the pop art of Takashi Murakami and his collaborators, and amateur like asset-remixing game mods and MAD videos (later MAD movies,) that took the results of such analysis as a starting point.^{†Azuma 2001}

On top of all that, there was also a better financial incentive. Doujin gamemakers could, in fact, sell their games, and the infrastructure allowing them to do so was much more robust than the scattershot operation shareware gamemakers had to rely on before online credit card processing made selling manageable, and aggregation sites made discoverability manageable (both changes coming too little, too late) (lam, 2010).

Cave Story brought all of this development with it when it arrived on the shores of the players who would become indies. Its emphasis on depth over breadth in design can be considered a tendency of the doujin sphere as a whole, down to the handcoded sprinkler effect seen in just two areas of the game. There are essentially two ways a program can be impressive. It can either be a very large system with lots of moving parts that nonetheless hold together well, or it can be a very small, very subtle thing with a million different facets that become apparent only after you've seen it from as many different angles. Doujin gamemakers of the '80s and '90s, without the resources to make the first, tended toward the second, and by the time Cave Story had come out, they had become very good at it.

Pixel art and chiptune music (and, later, low-poly models) have a low barrier to entry and require few resources other than access to a computer, but they're flexible and infintiely masterable. Even tiny sprites can be done in countless styles, all unique, as can be seen in the recent wave of sprite redesigns for princess peach. And when working in established genres, as doujin tend to do, the collective knowledge gathered about those genres, in the form of design tips or programming techniques or even tools like RPG Maker or KiriKiri, make it easier for the doujin gamemaker to concentrate on depth. On this note, there is considerable overlap between the doujin soft scene and broader otaku subculture, from distribution sites like Comiket, to a certain obsessive streak in the design of games like Cho Ren Sha 68K and Hellsinker, rooted as they are in encyclopedic knowledge of the minutiae of their genres. Not all doujin gamemakers are otaku, but on account of their historical ties to fandom spaces, there is a kind of otaku background radiation that permeates the scene, and which comes to the forefront from time to time, and recedes from time to time. There are shades of this in Cave Story too, for better (you can find a pixel-high man in the graveyard and return him to his pixel-high family, as an easter egg,) or for worse (you can enter a secret passage in Curly's room and make off with her panties. As an easter egg.)

8

When I look back on the parts of Cave Story that stood out most to me, outside of the action, the main thrust of the plot, although charming, is actually the least important. But it did matter to players in 2005 that Cave Story had a plot. One of the things that made the game stand out among others in the same scene is that the game wasn't in one of the arcade-style genres—shmups, fighting games, beat-em-ups, and the like—that represented much of what those discovering the doujin sphere first encountered, but also wasn't a visual novel (like Tsukihime,) or an RPG, or some other genre that leaned toward "pure story". In terms of gameplay, it filled a niche that hadn't been properly filled in the PC gaming landscape since the days of Jill of the Jungle and Duke Nukem 2.

In fact, I did feel compelled to poke around in this universe and I did feel a need to follow along with the miniature drama playing out among these characters. But what exactly drew me was something about the world itself, about being in there; most of the intrigue of the setting is in the gaps left in the plot where hidden pieces of the story suggest themselves. The architecture of the cavern labyrinths, the various robot models, all a little like you, all a little different, the strange lifelike designs—eyes, owls, sun-faces—etched into the foreground tiles. the way every piece of the game's setting seems to be cut from some whole cloth, although you never learn what the cloth is or where it came from. it's aesthetic-as-narrative by way of Xevious, the magnetic sekaikan of it all.

PUSH: sekaikan.

"sekaikan." "database consumption." "narrative consumption." these terms have been shrouded in mystification since they first reached the english-speaking world decades ago. but the process they describe is actually very simple, and each of these ideas elucidates the others.

the process begins with narrative consumption, which goes like this. someone creates a series of stories. these stories, though interesting in themselves, aren't what is really being presented to the audience. what is really being presented to the audience is something else, for which the stories are just entry points. this "something else" isn't just a backstory or background narrative, it extends beyond that. it's more like a proto-narrative system, or a system which underlies and precedes ("proto-") the production of these narratives, and allows the production of new narratives of a certain type. eiji otsuka, who developed the theory of narrative consumption^{†otsuka 2010}, refers to this system as the "grand narrative." it could be conceptualized as emerging from the connections between the individual narratives, which otsuka calls "small narratives," taken as a network of signs; it could also be conceptualized as a procedure for creating those narratives—"here's how you make a gundam show," "here's how you make a sonic fangame," etc. the "something else" is what people generally mean when they say "sekaikan."

when otsuka originally described sekaikan in the context of narrative consumption, he gave the example of someone taking the 772 extant bikkuriman chocolate sticker characters and deriving a plausible 773rd character by induction on the lore and character design of the previous characters, implying that sekaikan could encompass structural and stylistic elements of a work, as opposed to just narrative, characterization, or other kinds of content. likewise, zack wood, a game designer exposed to sekaikan while studying abroad in japan, includes aspects like game controls and mechanics in his definition of the term^{†wood 2013}.

for instance, the reason the quality of sonic fangames is so high is because the fandom has thoroughly absorbed the sekaikan underlying the classic sonic games, and have extended that sekaikan to include a certain style of 3d sonic game markedly different from the "official" 3d games, but which hews to what the fandom considers important about sonic platforming.^[11] this is an example of how such a system, which is not localized in any one work, can therefore be gradually modified by creating new works that fit within that system, according to group consensus, but stretch it just a bit, much like how languages evolve.

this proto-narrative, linguistic potential extends beyond complete narratives even to characters and individual design elements. it's what allowed a game like Xevious—considered one of the first japanese videogames to integrate sekaikan into its visual design—to bring a sense of place and history to its setting through the architectural style of its level backdrops and enemies, among other things^{†endo 2015}, despite not having a plot given at the title screen.^[12]

the flipside of this is that these elements—story beats, voices, character design elements, and the like—take on associations based on how they're used in other stories, and this forms a kind of language in itself that allows those elements to be used to create new elements and, eventually, new narratives themselves, which produce new plots and characters to attach new associations to, ad infinitum, by way of the system they originally belonged to.

eventually the associations, the charge held within these elements, becomes strong enough that one can consume the elements themselves, using the narrative as a delivery mechanism for them, just like one uses starch as a delivery mechanism for sauce—database consumption indeed.

hiroki azuma argues in otaku^{†azuma 2001} that database consumption replaced narrative consumption, but it is the author's hypothesis that "narrative consumption" and "database consumption" are just successive stages of the same process. in the first stage, standalone elements, like the xevious spaceship or the bikkuriman sticker character, are used as a point of entry into a suggested larger narrative; in the second, those elements have become so charged with the context in which they were originally encountered that they can become stand-ins for that context and those related ideas. if database consumption took longer to emerge, this is only because it required a critical mass of images charged with context before the language that came out of that mass was intelligible. you could've drawn a twintailed tsundere in the '60s, but nobody would've understood you; even in the early '80s not enough people would've understood you quickly enough. not enough had calcified.

this is similar to how words pick up connotations over time through the different contexts in which they're used, and it's a very basic phenomenon once you strip away the technical language—it's what allows us to see a character wearing a white cap with a cross on it and say "oh, that character's a nurse," even though nurses haven't worn that sort of hat in decades and the red cross is a completely different institution from a local hospital. we already use a kind of cultural language when reading most stories; what really makes the case of otaku fandom interesting is how a separate, parallel sub-cultural language emerged from interactions within a tightly connected community, and how much higher-context this language is compared to the one used by ordinary readers, and most of all how aware they are of how this process works and how deliberately they push it along and integrate it into their own art, and into otaku culture more broadly.

an ability to nail sekaikan appears to be an inevitable result of fan cultures becoming more sophisticated over time, involving not just intuition but careful analysis and debate. consider the unofficial sonic physics guide, or the unofficial level design standards surrounding mario levels and fangames, no matter how brutal (to the point that even the seemingly chaotic kaizo mods have rules of their own). or the style guides for community-led projects like SCP or omegaverse, or the design guides made for authors of interactive fiction.

the significant thing about sekaikan in videogames, for our purposes, is that it extends to game structure and mechanics. in this regard, it had clearly been mastered by the doujin scene by the time the rest of us would encounter them. it would be a gross oversimplification to suggest that this directly influenced the design sense of the early indies becoming as spohisticated as it did. (sekaikan had been known by some offshore gamers to some extent prior to the consolidation of the indie scene, as can be seen from this forum thread from 2003^{†kitsune 2003} though likely very few.) but there's a thread worth following here. POP.

9

The most important takeaway from Cave Story, to an offshore player encountering it for the first time in 2005, was that an amateur gamemaker could keep pace with the standards of design and presentation set by the industry, and even exceed them—if they shifted their priorities. perhaps they couldn't compete with the latest Quake, but so what? they could still compete with the latest Metroid or Castlevania, and just as well if those were the sorts of games you were really into. indeed, Cave Story was on par with Nintendo and Konami's own 2d metroidvania offerings at the time (which would've meant Metroid Fusion, Metroid: Zero Mission, and three Castlevanias, all on the Game Boy Advance).^[13]

Another part of what made Cave Story feel so new at the time was, again, its emphasis on story (including a wealth of unskippable cutscenes which may annoy players today, but which people sat through with popcorn back then). Prior doujin games in genres like RPG and ADV certainly existed, but were less often translated, so the precedent for independent-scale storytelling hadn't been established, and its combination with action gameplay was especially fresh after a decade of DOOM-era narrative minimalism. (perhaps some other game could've come along and been translated first, kickstarting this process years earlier.)

Cave Story was one of the earliest games to benefit from the burgeoning field of online games journalism. While it never had a chance of getting covered in print, or even noticed, by an EGM or Game Informer, there was intense cheerleading for it online as soon as it was released, being mentioned on Insert Credit^{†brandon 2005}, and called GOTY material in the Insert Credit forums^†Chaz, before it was even translated (again, being an action game helped its discoverability). News of the translation appeared on bigger sites like Inside Mac Games^†Park, and also on smaller blogs like Gameba, with much of this groundwork being set before TIGSource had been established. Cave Story was being referred to as The greatest game ever made by a startstruck Kotaku writer as early as 2006^{†brownlee 2006}, and the game was recommended as an entry point to doujin soft in an article by the Escapist that same year.^{†Szczepaniak 2006} A tribute site and fan forums went online months after release. Even Johnny Greenwood of Radiohead was playing it.^[14]†Lucas By the time news broke that Nicalis was porting it to the Wii in 2008, Cave Story had already passed into legend.

Cave Story didn't singlehandedly bring the modern indie forth. The indie scene is already full of inflated origin stories as it is, and what is so often presented as a unified culture with a single history is more like a series of semi-independent threads with their own individual origins (the earlier thread by Kayin touches on this elsewhere). In a way, what happened over the course of the '00s is that the doujin, Flash/Newgrounds, and TIGSource threads gradually overlapped, and finally collided. By the time Indie Game: The Movie had come out, most of the groundwork had already been set. By the time Braid had come out, most of the groundwork had already been set.

Cave Story was nonetheless an important and widely-held touchstone for the burgeoning indie scene, whose players would include Derek Yu, who in 2005 would head TIGSource, an indie hub whose alumni would include Terry Cavanagh, Konjak, Phil Fish, and a host of other so-called indie darlings. The game would often be cited by forum members as one of the games that first turned them onto the potential of indie games, and characters like Quote, Curly and Balrog would often appear in memes and signatures. The most vivid testament to the game's penetration of the forum's culture is Joe no Geimu, a (sadly unreleased) Cave Story hack whose existing footage (and accompanying rap) is shot through with indie in-jokes.

The ambivalent flipside of all this subcultural prestige is that Cave Story would contribute to dramatically raised expectations for development time, effort, and overall work-life balance in indie game production, on account of the mythology surrounding its development—Pixel had famously eked the game out over the course of 5 years, going back and redesigning its core mechanics midway through.

But these expectations may have been set in error. We can gather from his interview with Yu that pixel was balancing art with school during college, and with work afterwards. He mentioned having children to take care of, and doing most of his development late at night, and there's nothing to suggest he had an unhealthy work-life balance—certainly not the kind of monomaniacal focus that would be promoted in his wake. Pixel didn't set out to be the Romantic Indie Dev—we made him fit. (Similar to this is the 7 years spent developing Silksong at a reasonable pace rather than a frenzied one.)

What's more, as indie games improved dramatically in quality over the course of the ensuing decade, the technical skill floor lowered by the first wave of hits would get raised back up in the end. You could get away with pixel art and chiptune now, but it had to be the pixel art and chiptune of a professional.^{†Ryerson 2013} Tools would be released to help control the ballooning scope of new indie games—Unity, Unreal, GameMaker Studio, and the, like.^[15] But somehow this would just move the goalposts further. When these were released, we didn't make Iji faster and more comfortably—we made Hollow Knight.

On top of that, we never solved the original problem of the shareware era: Nobody could make any money off of any of this. Download stores like Xbox Live Arcade and digital platforms like Steam would open their arms to indies in the coming years, but they would prove difficult to survive in, and only a relative few would truly thrive.

The scope issue didn't really get under control until the creation of tools that forcibly limited the scope of what was made with them, and until responses to either the indies or standards they'd set—altgames, notgames, scratchware, zinesters, glorious trainwrecks, and so many other threads—emerged to encourage the use of those tools. Fantasy consoles would later emerge as a kind of compromise between the need to control scope and the need for flashy graphics and energizing music.

And yes, there were other threads. It was inevitable that other threads would weave their way into the spotlight as the limitations of this would-be avant-garde came into view. There was nothing within the orthodox ethos of the TIGSource-era indies, as sophisticated as it was, that could have produced something like Depression Quest, or Problem Attic, or even CORRYPT. So, a "mainstream" having emerged even within amateur game production, that mainstream would have to be expanded, critiqued, and revolted against in equal measure.

Just as well. The "indie community" of the present day may lack the unity of history and purpose presented by Cave Story and its playerbase, by IG:TM, even by something as recent as '00s-era throwback UFO 50 (from a team helmed by the same Derek Yu,) but that unity had always been an illusion to begin with. Besides, the vibrant field that resulted from all these threads coming together to put amateur game production back on the map is more than big enough for all of us.

10

These days, you really can find Cave Story anywhere; it would eventually breach containment and find its way onto consoles, taking the indie banner to an ever wider audience. By now it's been ported, officially and unofficially, to everything from MacOS and Linux to the 3DS and PSP to the Amiga to the Sega Genesis, to give just a fraction of them. This is in large part thanks to the game's vast worldwide fandom, who have been accumulating terabytes of fanart, mods, hacks, tools, covers, remixes, cosplay photos, and every other kind of transformative work since 2004, and who have translated the game into over a dozen languages. The point of entry to this corpus is the fandom's home base at https://www.cavestory.org.

Among the better examples of the game's legacy are those games that took direct inspiration from it, from Underside to Undertale. It's undoubtedly the reason Metroidvanias are such a bread-and-butter genre in indie gaming. If you're not riffing on Cave Story, chances are you're riffing on someone who is, or someone who's riffing on someone who is, and so on.

But perhaps the best example of the game's legacy is the indie scene itself, in that you can see in Cave Story a microcosm of that scene's hopes and dreams, if also its original sins.

To conclude, here are some tips for new players:

1. The Bubbler's wide spread makes it good for recouping weapon energy during boss battles. The maxed-out Blade is useful for the same reason.
2. When in doubt, investigate the furniture—bookshelves, computers, fireplaces.
3. If you're going for 100%, for your sake, save the life pot for the last section of the game. You'll know when you're there, because there won't be any chances left to save.
4. If you ever meet Pixel, give him my regards.
5. Get the tow rope.

NOTES

1. An article linked below gives this year as "1972". I'm going off the date the game published, rather than the date it was written, and could find no mention of it before the 5th issue of PCC in May 1973.
2. The 1950s, assuming your cutoff point is the oft-cited Tennis for Two in 1958 or—looking at you, Ahoy—Draughts in 1952; or the early '60s, if you want to hedge it and start with Spacewar!, as Arcade Idea suggests we do; or the '40s if you want to be really obtuse or, like me, you have a secret consipracy theory that a punch-card program for the Whirlwind bouncing ball game exists buried in an archive somewhere at MIT.
3. Even a software renderer shambling along at 640x480 and borderline vomit FPS was banking on chips getting fast enough to meet it halfway. This is code, not magic.
4. The GDC's enthusiasm for DOOM can be seen in a report on page 8 of the priemiere article of their Game Developer magazine, in April 1994.
5. That's not fair. It's more like shoujo Kaizo Mario.^[16]
6. There are also cases of mods and fangames that predate the indie movement but have been folded back into indie as it's intuitively understood, like Sonic Robo Blast 2.
7. Pixel would eventually reveal Metal Slug to have been a major influence on the game during a Q & A in 2022.
8. There was a DOOM port for PC-98, but it was...not the most flattering introduction. That there was, as far as I'm aware, no shareware release to prime the grapevine certainly didn't help. A Japanese DOOM community would eventually emerge (check out the Japanese Community Project if you want to see what they can do,) but this would be predicated on the same pivot to Windows that would make smoother and more sonorous ports of DOOM—not to mention existing map editors—accessible to Japanese players, and by then the game's window as a potential trendsetter would have closed.
9. It needs to be clarified that when I say "BBS", i mean something different from the dial-up systems referred to by that name offshore ( although those did exist in Japan too, in the '80s and '90s). In this case, the BBSes used by Yosshin, Pixel, et al. existed on the World Wide Web, and the early adoption of the World Wide Web by the doujin soft community was a major point in their favor. In particular, we have records (last link on the page) of the BBS messages being sent back and forth between Pixel and his playtesters during the final year of development.
10. I say "engine" but many of these tools predate widespread adoption of the term as coined by id Software. They might better be referred to as "editors," evolving out the simple level editors that were included with games like Wrecking Crew and Eggerland Mystery, but expanded to whole genres rather than specific games.
11. some examples of the fandom's notion of good 3d sonic platforming can be seen in their fangames. games like sonic utopia (video here,) sonic: project hero (video here with a mod featured here to emphasize how movement works,) and sonic calamity all combine the spacious, racetrack-style levels and high speeds of the modern sonic games with the emphasis on physics and momentum that marked the classic series games. (a "3d classic sonic," if you will.) notably, none of them include a boost mechanic, and all of them implement some functional version of the super spin attack. meanwhile, games like sonic overdrive (video here) and sonic rush 3d model a kind of idealized modern sonic game, tricks, boosts and all. as for the fandom's notion of good 2d sonic platforming, it sufficies to say that the line between "official sonic sekaikan" and "fan sonic sekaikan" is porous—creators like christian whitehead and tee lopes started out making fanworks before getting hired by sega, and brought everything they learned in the trenches with them. a game like sonic mania has one power-sneakered foot in both worlds, and was lauded for being more accurate to the spirit of the classic series than sega's own in-house efforts at the time (the obvious points of contrast being sonic 4 or the classic stages in sonic generations).
12. compare with "diegetic narrative".
13. This is no small claim, as early indies and their allies put a lot of stock in the Game Boy Advance. The resentment for the "state of the art" bubbling under the surface is palpable in this article on freeware game recommendations, touting the GBA as that hallowed last bastion of 2D gaming, and calling one the games covered something to keep you busy while you boil ramen or wait for Doom 3 to load. There were battle lines being drawn around games like these.
14. The link to the Radiohead blog is broken; click here instead.
15. That's the other thing about about advances in technology: it means better tools could run on cheaper machines. It used to take a lot of good tooling (a lot of money) to fill a megabyte well (see: Metal Slader Glory). Cave Story added up to just under that, because it was made in 2004, and as much as we take it for a throwback, it was a lot easier to write Cave Story for Windows XP in 2004 than it would've been to write Cave Story for the Mega Drive in 1994. in fact, it was easier to write Cave Story for the Mega Drive in 2022 than it would've been to write Cave Story for the Mega Drive in 1994. (Don't believe me? Two words: 87.2% C.)
16. Not to be confused with Shoujo, who plays Kaizo Mario, on YouTube.

BIBLIOGRAPHY

• ahl, david h. 101 BASIC Computer Games. Digital Equipment Corporation, 1973. https://archive.org/details/101basiccomputer0000davi/mode/2up. accessed 15 august 2025.
• bourdieu, pierre. The Field of Cultural Production. orig. 1962. ed. randal johnson. The Field of Cultural Production. Columbia University Press, 1993.
• brandon. News: Doujin updates. insert credit, 2005. https://web.archive.org/web/20050205020822/http://www.insertcredit.com/archives/000117.html. accessed 17 August 2025.
• brownlee. Cave Story Hell Speedrun. Kotaku, 2006. https://web.archive.org/web/20090629232210/https://kotaku.com/174827/cave-story-hell-speedrun/. accessed 26 July 2024.
• Chaz. F.O.N.T. (mostly about Doukutsu). insert credit forums, 2005. https://web.archive.org/web/20050205015856/http://insertcredit.com/forums/viewtopic.php?t=2423 accessed 17 August 2025.
• Cogger, Kevin. Crucial Classics: Hunt the Wumpus. 1UP.com, 2005. https://web.archive.org/web/20051228134442/http://crucialclassics_10.1up.com/ Accessed 19 August 2025.
• dir. ka, alex. 芸夢 [gei·mɯ] File #1 - Yūichi TOYAMA ~Pioneer of Modern Real-Time Strategy Games~ (English subs). Game Preservation Society, 2021. https://www.youtube.com/watch?v=EnBZvyxTVaU. accessed 15 august 2025.
• Doujinshi: Japan's Most Misunderstood Medium. translating japan, 2021. https://www.youtube.com/watch?v=5N0VerI9DxE. accessed 28 december 2023.
• EIJI, Ō., & Steinberg, M. (2010). "World and Variation: The Reproduction and Consumption of Narrative." Mechademia, 5, 99–116. http://www.jstor.org/stable/41510959 accessed 13 August 2025.
• endo, masanobu; trans. blackcloak. Xevious – Developer Interview Collection. shmuplations, 2015. https://shmuplations.com/xevious/. accessed 17 August 2025.
• famibe no yosshin; trans. blackcloak. ChoRenSha68k Development Notes. shmuplations, 2014. https://shmuplations.com/chorensha68k/. accessed 17 august 2025.
• fiadotau, mikhail. "Indie and Dōjin Games: A Multilayered Cross-Cultural Comparison." GAMEVIRONMENTS, ISSUE 10, 2019. https://journals.suub.uni-bremen.de/index.php/gamevironments/article/view/95. accessed 8 december 2023. [2]
• grau de pablos, tomás. "Between Indie and Doujin: The creation of the Japanese Indie." game studies: the international journal of computer game research, volume 24, issue 3, 2024. https://gamestudies.org/2403/articles/grau_de_pablos. accessed 12 November 2024. [8] [9]
• Independent Innovators: Doujin, Technopolis, and the Sharp X68000. F_T_B, 2025. https://www.youtube.com/watch?v=u2gqoxaUjN8. accessed 17 august 2025.
• justice, kim & ashens. "Eugene Evans, Imagine, and the Bedroom Coder Dream." 2019. https://www.youtube.com/watch?v=O9pTlMvdjaM. accessed 13 November 2024.
• kalata, kurt. Clash of the Cultures. 1UP.com, 2007. https://web.archive.org/web/20150518151011/http://www.1up.com/do/feature?pager.offset=0&cId=3155815. accessed 17 august 2025.
• kayin. untitled tweet, 2024. https://twitter.com/KayinNasaki/status/1747530601828114845. accessed 18 January 2024.
• keogh, brendan. The Melbourne Indie Game Scenes: Value Regimes in Localized Game Development. Independent Videogames: Cultures, Networks, Techniques, and Cultures. ed. ruffino, paulo. Routledge, 2021. https://eprints.qut.edu.au/227028/. accessed 26 june 2025.
• keogh, brendan. THE VIDEOGAME INDUSTRY DOES NOT EXIST. the MIT press, 2023. [4] [5] [6] http://direct.mit.edu/books/book-pdf/2085066/book_9780262374132.pdf. accessed 8 july 2023.
• kitsune. Sekaikan. Quarter To Three Forums, 2003. https://forum.quartertothree.com/t/sekaikan/7222 accessed 17 August 2025.
• kobayashi, nobushige & yuhsuke koyama. "The Early History of the Hobbyist Production Field of Video Games and its Impacts on the Establishment of Japan's Video Game Industries." REPLAYING JAPAN, VOL. 2, 2020. https://researchmap.jp/hichibe/published_papers/25369891. accessed 8 dec 2023. [1]
• koshiro, yuzo. "Talk About Micom Basic Magazine." 2019. https://www.youtube.com/watch?v=T5lbX-qk7rY. accessed 06 November 2024.
• lam, fan-yi. "Comic Market: How the World's Biggest Amateur Comic Fair Shaped Japanese Dōjinshi Culture." mechademia: second arc, vol. 5, fanthropologies, pp. 232-248, 2010. https://www.jstor.org/stable/41510966. accessed 29 December 2023.
• lucas (yougavemehope). Radiohead digs Doukutsu. doukutsu.livejournal.com, 2007. https://doukutsu.livejournal.com/80887.html. accessed 17 August 2025.
• van de Weyer, Martin. Shmup_Lyf - Dezaemon Dimensions. Magcloud, 2022.
• moss, richard. SHAREWARE HEROES: THE RENEGADES WHO REDEFINED GAMING AT THE DAWN OF THE INTERNET. unbound, 2022. [3] [8]
• npckc & sdhizumi. douZINE. 2022 https://web.archive.org/web/20221118014158/https://douzine.crd.co/. accessed 16 dec 2023.
• park, eddie. Doukutsu Monogatari Freeware Available. Inside Mac Games, 2005. https://www.insidemacgames.com/historical/news/story.php%3FArticleID=12206.html. accessed 17 August 2025.
• ed. emma parkinson. LARA: The Official Lara Croft Magazine. Future Publishing, 1999.
• picard, martin. "The Foundation of _Geemu_: A Brief History of Early Japanese video games." game studies: the international journal of computer game research, volume 13, issue 2, 2013. https://www.gamestudies.org/1302/articles/picard. accessed 8 dec 2023.
• Pixel. The Story of CAVE STORY. Game Developer's Conference, 2011. https://www.gdcvault.com/play/1014621/The-Story-of-CAVE. Accessed 20 August 2025.
• Pixel. Gero Blaster: Frog Shooter, not Boombox. GameSpot Gameplay, 2013. https://www.youtube.com/watch?v=a6ak1kaMhzo. Accessed 21 August 2025.
• rankin, joy lisi. A PEOPLE'S HISTORY OF COMPUTING IN THE UNITED STATES. harvard university press, 2018.
• rock. A story in a cave. Gameba, 2005. https://web.archive.org/web/20231125024228/http://gameba.blogspot.com/2005/01/story-in-cave.html. accessed 17 August 2025.
• ryerson, liz. The Talk of Magicians. ella guro, 2013. https://ellaguro.blogspot.com/2013/08/the-talk-of-magicians.html. accessed 17 August 2025.
• sršen, mate. untitled tweet, 2023. https://twitter.com/MateSrsen/status/1740766129105314140. accessed 29 September 2024.
• szczepaniak, john. "Doujin A Go-Go, Baby!" the escapist, 2006. https://www.escapistmagazine.com/doujin-a-go-go-baby. accessed 4 March 2024. [7]
• Designing 2D graphics in the Japanese industry. Video Games Densetsu, 2018. https://vgdensetsu.tumblr.com/post/179656817318/designing-2d-graphics-in-the-japanese-industry. Accessed 19 August 2025.
• wood, zack. Characters and Worldbuilding: Analyzing the Strength of Japanese Games. Game Developer, 2013. https://www.gamedeveloper.com/design/characters-and-worldbuilding-analyzing-the-strength-of-japanese-games. accessed 17 August 2025.
• wrist, tara. All their Engines. Bullet Points Monthly, 2018. "https://bulletpointsmonthly.com/2019/05/01/all-their-engines-best-of-2018". accessed 15 august 2025.
• yob, gregory. "Hunt the Wumpus." The Best of Creative Computing, vol. 1, pp. 247-250, 1976. https://www.atariarchives.org/bcc1/showpage.php?page=247. accessed 24 October 2024.
• yu, derek. "Interview: PIXEL." the independent gaming source, 2005. https://archive.cavestory.org/tigsource. accessed 4 March 2024.
• yu, derek. untitled tweet, 2018. https://twitter.com/mossmouth/status/987105285503238144. accessed 29 September 2024.