STS13: Videogame Typography and its Antecedents

09 Mar in a is for atari, conference, history, typography

This past week, I traveled to and presented a paper at the Society for Textual Scholarship. Although I’ve been interested in and working with Textual Studies for some time, this was my first time at the STS conference, and more than anything else, I found it an extraordinary learning experience.

Ironically, STS13 was just a few blocks away from SCMS 2013 — my normal Spring conference — but I think I made the right choice. Judging from the tweets coming out of SCMS, there were a lot of great panels showcasing some excellent game scholarship, but lately I’ve been feeling a need to contextualize my work a bit differently, especially since I’m in the beginning phases of a larger research project (more on that later). So STS was a good place to start building that foundation.

As is my custom, I’ll post the text of my paper below after just a few more words of explanation.

My 2008 dissertation was about videogame typography and textuality, and I’m making plans now to start writing the sequel. This is my attempt to start asserting or re-asserting why I think this topic has scholarly merit, and I definitely think it does. I have more to say and explain, but I think I’ll post those as comments instead of caveats. Just please bear in mind that this was written to be read out loud to a particular audience, and it’s only been lightly edited. Otherwise, please enjoy the paper that follows, and I definitely welcome comments, questions, or anything in between.

Videogame Typography and its Antecedents

In November of last year, I had a conversation with an administrator about the viability of a funded research project I was then proposing. This colleague, my associate provost was relaying his support of my proposal, but in characterizing the mixed, anonymous feedback of the reviewing committee, he offered some advice which has stuck with me. “In the humanities and social sciences,” he said, “we often face that question, ‘so what.’[slide] It’s unfair and reductionist, but it comes up. Well when your research is on videogames, maybe think of that question in ALL CAPS.”

This was good advice. I’d grown comfortable in writing and speaking to generally friendly crowds, so it was a helpful reminder that not all audiences will share my assumptions or what I value when I articulate just what is so interesting or important about the jots and tittles of classic video game systems from the 1970s like Fairchild’s Channel F, Mattel’s Intellivison or the Atari Video Computer System.

Today, while I think I /can/ assume that the present audience is generally friendly to my interest in typography, I do want to take this opportunity to begin articulating a “So What” for my project.

That video games can be productive objects of textual inquiry, I do take for granted, thanks in to folks like Steve Jones who have led the way on these questions. The number of journals, scholarly monographs and annual conferences focusing exclusively on games seems to be growing. And even more importantly, the conversation in the field now has enough coherence that it can be said to have moved beyond a formalist turf wars in the early years of the last decade, and now having grown up a little, game studies has moved toward what might be called a New Historical phase as influences of media archaeology and textual studies make themselves known in other forms under names like Platform Studies, Critical Code Studies or Software Studies.

Along this line, what I think characterizes or even anchors much of this critical work on videogames lately is the effort scholars are employing toward locating videogames materially, economically or socially as situated artifacts of expression reflecting or projecting the ideology of their contexts. This is the space in which an understanding of videogame typography has the potential to intervene. As I’ll hopefully show over the next few minutes, videogame type helps unpack the complex digital textuality of videogames, and in doing so, could support more nuanced practices of videogame study. Moreover, the typefaces used in videogames can be rich and nuanced, and their genealogy is all the more provocative of their social contexts because of their usually ad-hoc development, implementation and dissemination.

So to orient what follows under the “so what” that I began with, I’m going to propose three lessons videogame typography teaches has to teach videogame studies.

Now, if the phrasing of what follows sounds provocative, let me be clear that these lessons are simply some rhetorical pedagogy. In other words, they’re not entirely true, so if you challenge me on any of these, I’ll happily walk them back.

Lesson 1: There is no such thing as a video game.

Of course, if this statement were literally true, this could be the end of my paper. What I mean, though, is that the idea of a singular, static self-closed piece of software or hardware comprising or conveying a system of meaning or an experience is a fiction of consumption, cultural memory, and convenience. It is at most an anamorphic perspective on what is at best a more or less coherent confluence of many different systems.

Arcade games, for example, are ported, modified, hacked, upgraded, and bootlegged so much that the odds that any two of us could be referring to an identical set of images, sounds, and textures when we say “Pac-Man” is low.

Arcade games like Pac-Man were highly modular and subject to various improvements and modifications over time. Arcade owners could and would tweak their games to maximize profit or challenge. With an upgrade kit, they could swap out a few chips and some cabinet artwork to transform it into a different “game” entirely. [apparently this was discussed in a paper at sts11]

In the home console space, where we most strongly associate the idea of a game with a single physical artifact like a disc or a cartridge, it’s notable that most manufacturers in the 1970s seem to imply a distinction between a “game” and a ROM chip containing the software with which to play a game. Fairchild released “Videocarts” containing up to 4 games, Atari produced “game programs” advertising as many as 27 games, and Mattel sold “cartridges” for its Intellivision. Today’s gamers don’t buy “games” either; instead, we purchase licenses which allow us temporary, provisional access to we’ve downloaded.

So there is no such thing as a video game in the unilateral sense. And this is to say nothing of the social, economic and material antecedents of play that create the means, opportunity and spaces in which to experience these games. Erkki Huhtamo, for example, has explicated the Victorian roots of the domestic media center within which the home game console found its place in the 1970s, and Steve Jones has written of those spaces as meaningfully paratextual.

So far these complexities I’ve mentioned are basically no different than the textual conditions of literature. But one basic difference is that videogames usually blend audial, visual, haptic and typographic utterances, and since most videogames since the mid 1970s are also software, we can dig down into the processes which comprises the mix of these modalities — we can as Wolfgang Ernst paraphrases Foucault, understand software by “reconstruct[ing] the generative matrix created by mediatic dispositifs”. Understanding the ways by which linguistic and numeric signifiers move from code to machine to screen helps illuminate at least one vector within that generative matrix of the video game.

The Atari VCS or 2600 hit the shelves in 1977, but due to its lower price and, subsequently, its extensive library of games, it remained a significant market presence throughout the 1980s (1990). In 1980, George Plimpton spoke on behalf of Mattel’s Intellivision, touting its comparatively more realistic graphics compared to Atari. In this ad, the representational fidelity of the baseball players on either screen may involve some subjective interpretation, but more importantly to me, the typographic distinction between the two systems is evident and important.

Internally, the VCS is somewhat unusual, even for its time, in that it lacks a BIOS, or a simplified operating system. This is machine-specific software that, on a more advanced system like Intellivision, populates the programming environment with features like a screen buffer or an ASCII implementation for translating data into graphic representations of letters. The lack of a native text kernel for Atari VCS meant that each programmer who wished to use letters or numbers had to invent or borrow their own kernel, and the result is a good deal of diversity in how numbers (usually just numbers) appear in games. Intellivision programmers, on the other hand, could simply encode a string of ASCII values that displayed corresponding sprites from the machine’s memory.

Intellivision’s method may be easier to program, but it also means that every game for Intellivision uses this same font for its text, unless the programmer took extra effort to craft something unique. This Intellivision font is associated so strongly with the games designed for its systems that it persists into other environments where the constraints of its BIOS no longer apply, including the Atari VCS where Mattel ported several of its popular titles under the “M Network” label. Here in the Atari implementation, we can see an imported typeface, the presence of which is evidence for a standardized text Atari text kernel specific to Mattel, Atari’s primary rival.

So for lesson #1, what Kool Aid Man et al teach us here is that what we’re tempted to perceive as a single game corresponding to a single ROM image is in fact interpenetrated by many nearly invisible forces like corporate licensing agreements and programming practice, but these seams leave their traces in the typeface.

Lesson #2: Videogame history is wrong.

Again, this claim is not literally true, generally speaking. There are many excellent books about the who, what, and when of videogames. But this is kind of the problem: even the best, comprehensive books such as Leonard Herman’s Phoenix (boil down to) a list of things that happened, written in the order that they happened. Games, consoles and people are all points in a timeline moving inevitably from the past toward the present.

One intervention that a study of type could make is to instead locate game history somewhere in the unfolding microtemporality of the machine such that, instead of thinking of the 4 on the screen in terms of its presence, a moment of unilateral four-ness, we could instead recognize an active process of becoming-4 that manifests the progressive unfolding of four-ness through the progress of several microtemporal layers: including the loading of the bitmap graphic into the systems RAM, the switch to update the score counter, the timing of the scan lines as they move across the screen 30 or so times a second.

At a larger scale, typographic genealogies also intervene by opening up new affinities between games and systems, opening a path to alternate history for games.

MAME or the Multiple Arcade Machine Emulator is a remarkable resource for anyone interested in studying video games. I’d guess that most of MAME’s userbase uses it to illegally emulate arcade games on their computers, but MAME’s expressed purpose, which also acts as a legal disclaimer, does state that its intended use is for scholarship and preservation. Not only does MAME allow researchers to play (through emulation) a very large number of otherwise inaccessible arcade games, it includes some rich metadata as well. This data is in two forms: a large XML file containing, mainly, what MAME needs to know in order to play a set of ROM images, as well as a set of “driver” files encoding specific instructions for loading and executing that game. This functional orientation means that employing that XML for culturally relevant data involves some educated guesswork, but this information on some 18,000 different entities is ripe for for mining. Of the 18,000 or so entities understood by MAME, some 8000 are Fruit Machines or Skill With Prize machines, and another 2000 or so are pinball machines. Among the remaining 8000 identifiable ROM sets, excluding clones, bootlegs and other duplicates — MAME claims it can account for over 4000 “unique games.”

I’ve just recently begun working this data set, but it immediately reinforces lesson #1 when so many games share ROM files and drivers. But in reference to typography, MAME’s data is beginning to answer a specific research question I’ve been exploring.

There is a particular bitmap typeface used in video games. Its characteristics include a humanist axis, a left-leaning 8, S and 5, and a top-heavy R. You’ve seen it before, and if you see it somewhere other than a video game, it’s likely in some context where a designer wanted to make you think of video games for some reason.

You can see it here in Radar Scope (1980) and Donkey Kong (1981). If you know your game history, then you know that these two games have an important relationship since Shigeru Miyamoto at Nintendo built the latter using the hardware of the former. Given the task of “improving” Radar Scope, Miyamoto just transformed into something entirely different — different setting, different gameplay, a different (novel) genre — but one thing he left unchanged is the font.

This is also the font of Super Mario Brothers, of Pac-Man, Dig Dug, Bubble Bobble, and many, many other alliterative heroes of dark electronic dungeons. The font is ubiquitous and persistent, and it seems to originate in the video arcade, yet so far as I know, its story remains untold. I don’t even know what to call it. Type designer Ray Larabie has released a commercial TrueType version with the name “Joystix”, and I’m sure there are many others, but this is obviously a retroactive design and designation. I’ve also seen it called “The Namco Font”, and indeed the evidence I’ve gathered from MAME does show that nearly every Namco game through the 70s and 80s used it, but of course, so did most Nintendo games, as well as quite a few by Taito, and Konami — companies all based in Japan, by the way.

This might be important, because what I learned from MAME suggests that probably originated at Atari, an American company. [slide - sprint 2] You can see it in several games and prototypes from 1976 including Cannonball and the car racing game Sprint 2. By this point, Atari’s expansion into Japan had run out of money and Namco, which had already been distributing Atari’s games, bought Atari’s remaining assets and began designing their own games. Their first independent game was 1978’s Bee Gee, which does use the “Namco font” — 2 years after Atari, and (significantly) *after* Namco had acquired Atari Japan’s physical assets.

In other words, I think this shows some evidence (circumstantial, of course) that when Toru Iwatani built his Gee Bee prototype, he used parts and chips left over from Atari games, and these chips included what we now sometimes call the Namco font. Now, if this event actually happened, it illuminates an important juncture in video game history. Iwatani’s next game after Gee Bee was Pac-Man. But more so, what’s interesting is what happens next for the Namco font’s story: it proceeds down two separate timelines: first, into the further past for its originator at Atari, and even further back to its film or print antecedents. (For example, its thick and thin strokes are similar to the same distributions on humanist or Garalde faces like Garamond, Goudy Old Style or Bembo.) And the second timeline of course is into its future, where its possible to follow this typeface and its derivatives as it moves from one “unique game” to the next. Each time, that transition occasions a similar question about the moments of game development, and these can lead to some revealing inferences. For example, Taito generally used a different, less-interesting typeface until about 1981, which is also after they’d begun licensing and distributing internationally some games by Namco and Nintendo, implying once again that the convenience of available ROM chips, separately from its aesthetic superiority, let to its ubiquity.

Lesson #2 is that video game history is wrong, or at least that videogame history is more than a list of things that happened. Instead, it’s a story of the movement and interaction of non-human agents like typefaces, rom chips among a landscape of corporate licensing agreements.

So finally, bearing in mind the same qualifications and apologies as above,

lesson #3: Videogame ontology is incomplete.

This third lesson risk collapsing with the first, that there is no such thing as a videogame, but here I want to consider the form of the game as type of claim toward an ontological formalism that often happens when media studies contemplates the materiality of media artifacts. Yesterday, a few blocks away at SCMS (if I read twitter correctly), this inclination toward conceptual stratigraphy was characterized as part of the importation of media archaeology field, where the scholarly injunctive is to dig into the strata and substrata defining media and media systems.

With reference to game studies, though, I’m thinking here of the approach to treating games as participating in multiple “levels” or “layers” as Nick Montfort does, for example, in his influential “Combat in Context” essay. Montfort’s point is that videogames are part of systems beyond themselves, and at each point in these complex parfaits of affordances and constraints, “culture” obtains. Other writers also use this figure of levels or layers in writing about the nature of what video games are, and I think a similar logic is at work in the rhetorical gesture of referring to the inner working of software systems as taking place “behind the screen”. For instance, Matthew Kirschenbaum charactizes his turn toward materiality in _Mechanisms_ as responding to a “screen essentialism” then apparent in some digital media studies, a tendency to assume that the work of digital texts is bounded and completed by the screen and its interface.

In making my final point, typography wants simply to add back that layer of the screen, to reinvigorate the videogame screen as one of the key thresholds where meaning is negotiated graphically. Not to re-essentialize the screen, but to acknowledge its interaction with other, “deeper” layers of the system that contains the experience of the game’s meaning.

The screen is a particularly important domain when the representation of typographic characters is concerned, where the demands of legibility run up against the resolution of the raster grid of the screen and the phosphorescence of its pixels. This characteristic blurriness is such an essential aspect of pre-HD video game graphics that most software emulators now include filters to simulate that effect on our crisp LCD screens.

Game programmers crafted their graphics with that blurriness in mind, and in at least one case, Berzerk, that expectation creates an interesting association with a specific print font.

Berzerk is interesting for other reasons as well — its depiction of humanoid-on-humanoid violence helped launch one of the first moral panics about videogames, it also contributed to at least one player’s death here in Chicago — and its multiple versions from Arcade to 2600, 5200 and Vectrex showcase an array of accommodations for screen effects, including one variation that points to a noteworthy generic antecedent.

In the arcade, Vectrex and 5200 versions, the typefface is basically a segmented line font, basically reminiscent of Eurostile, but for the 2600, something different happens. The programmer responsible for this version, Dan Hitchens, chose and coded this face instead, the proportions and distributions of these numerals do invoke a specific antecedent in the novelty Letraset font “Countdown”, designed by Colin Brignall and published in 1965.

What “Countdown” does for Berzerk on the 2600 is reinforce the science fiction milieu otherwise relegated to the paratext of its box art. Lacking the robot voices of the arcade original, this is one of only a few generic signifiers of sy fy in this somewhat opaque and spooky game.

But notice how my screenshot here of Berzerk’s numbers isn’t completely correct. The inner corners lack the rounding evident in this Countdown sample, which Brignall no doubt intended as an echo of the blurring effects of the screen that Hitchens knew would be in place when the Atari’s Telivision Interface Adaptor chip generated one yellow pixel for each bit of the data he encoded in his bitmaps.

This is a simple example, but the lesson here that typography wants to teach game studies is that an ontology or conceptual stratigraphy which omits the screen is missing something important indeed, just as an emulator running on a computer with an LCD monitor hasn’t recreated or improved on the original, but has, rather, created a different new version with a different aesthetic.

In this talk, I’ve tried to speak on behalf of videogame typography to argue that there is no such thing as a video game, that videogame history is wrong, and that videogame ontology is a futile endeavor. Since these are all lessons in the negative, I will just close with a hope that I’ve also managed to at least hint at the richness and even elegance present in video game type.



Anonymous's picture

Efficacy, concision, focus

This presented negligible information using jargon nearly unreadable for its irrelevance, and in a sensationalist way to boot! With a title “Videogame typography and its antecedents” (presented in the “Namco font”, no less), what the reader expects is the story of where that and other commonly seen typefaces originated, but what he gets is an eyeful of egoism. I don’t read an article entitled “Videogame typography and its antecedents” to learn the author’s conference attendance habits, nor his absurdly eburnean positions on video game “ontology”, nor even his reactions to (apparently; finally) stumbling upon MAME. I read it to learn how video game typefaces came to be designed as they were. All the reader learns about this subject from this 3,500-word ramble are 1) that the “Namco font” originated before Shigeru Miyamoto and Nintendo ran with it, probably at Atari; and 2) the typeface used in Berserk was likely modeled after Countdown.
I am reserving anonymity to avoid a flame war and make you concentrate on this message: This essay has earned you an F in tenth-grade English. Maybe you’re overworked, overly burdened with teaching and output, I prefer not to speculate, but I hope for your students’ sake your lectures are not like this. Feel guilt-free to delete this comment after reading; it is written for you, not for any other readers.

zach's picture


Obviously, I don’t appreciate the personal insults, but I’m actually glad for any kind of engagement on this paper and this topic. Honestly, it’s been hard to find others in academia interested in the same kind of questions I’m approaching here and that you were seeking answers to, and the defensive, dare I say navel-gazey tone you’re noting in this paper is partly a result of my having to make that argument to a room full of textual scholars.

If you’re genuinely interested in the “Namco font” — which should probably be called the “Atari Font” or something else — it seems to have been created by Lyle Rains at Atari when he was working on their first CPU-based arcade games in the mid-70s — “Sprint 2”, for example. So Nintendo, Namco, and others must have picked it up from modifying, copying, and reverse engineering the Atari cabinets they were getting.

So it’s an interesting font because it’s emblematic of that crucial shift to CPUs, and it’s had staying power because of that influence. It also just looks nice; I think so, anyway.

I am, by the way, working on a book on this subject, and while I’m writing to address the wider situation of digital textuality, I’m also orienting it to try answer the kinds of questions you came here seeking answers to. The challenge is that once you get past a certain layer of widely known information, there’s just not a good record of this kind of thing.

May I ask, since you seem to be interested, what other fonts you’d be interested in reading about?