While preparing to teach my spring graduate seminar, BookLab: Print to Programming, I’ve been thinking about why a student studying information science might want to take an experiential course at the intersection of book history, media studies and digital humanities. Writing several years ago in “How Not to Teach Digital Humanities”, I argued that DH practitioners should seek to “create versions of DH that make sense not at some ideal, universal level but at specific schools, in specific curricula, and with specific institutional partners.” I believed this when I was working in an English Department, and I continue to believe this now that I have moved to a School of Information Science. As such, it is incumbent on me to reconsider the frameworks of my teaching in light of new students with distinct skills, goals, and needs.

For English and History students, working directly with historical media offers insights about their primary texts, and can help them develop research questions to consider features of material culture, whether toward analyses of literature or social, political, or cultural discourses. The same will certainly be true for some iSchool Ph.D. students planning cultural analytics dissertations, while for many students in the iSchool’s MLIS program, the class’ archival visits, historical media labs, and digital humanities training have clear applications to the curatorial, archival, or pedagogical roles they might take up in libraries, archives, and other cultural heritage institutions. But why might a student studying data science, cultural analytics, or computation—or, frankly, training for some of the less archives-oriented roles in libraries—engage with media we might describe as “old-fashioned” or even “obsolete?” I have written before about the relationships, historical and theoretical, between letterpress printing and computer programming, and at some gut level it seems true that a historical understanding of technology would benefit professionals extensively using and building modern technologies. But the fact that practices such as typesetting and programming “share essential features of deliberation, planning, execution, industry, and automation” does not in itself argue that information scientists—or even digital humanists—should learn to compose movable type.

In this post I propose that a framework of cultural “dependencies” can help articulate the value of historical media studies to contemporary information praxis more broadly. Cultivating both practical and theoretical understanding of legacy technologies offers information scientists crucial perspective on the technical and social commitments, possibilities, and limitations of contemporary data design, human-computer interaction (HCI), computer vision, and machine learning, as well as a critical orientation toward the labor and infrastructure undergirding contemporary media technologies. Much as scholars such as N. Katherine Hayles have argued “digital media have given us an opportunity we have not had for the last several hundred years: the chance to see print with new eyes” (Writing Machines pg. 33), converse attention to historical media make visible and salient questions of infrastructure and labor that can be obscured by ubiquitous field practices and twenty-first century interfaces alike. To put this idea another way: to design and implement computational systems that serve human beings, we need to learn from the informational systems that have evolved over much longer periods of time to meet both human needs and desires.

To begin unpacking that idea, I recall a post by Ted Underwood—now my iSchool colleague—proposing that the movie Memento, in which the character has to reconstruct his memory and current situation from clues he left in his environment and on his own body, offers an argument for studying the past:

One important aspect of “our whole lives” is, of course, our media technologies: our computers, our phones, the ubiquitous devices scattered throughout our homes and offices. Despite the presentism of much technological discourse, our technologies are thoroughly “defined by instructions handed down from the past” that influence their shapes, their functionalities, their capacities, their limitations, and their communities of use. In 2021, we type on keyboards laid out to meet the mechanical needs of typewriters and influenced by the feedback of telegraph operators, who “found the alphabetical arrangement to be confusing and inefficient for translating morse code.” While more efficient keyboard layouts have been developed and proselytized in the intervening century, a combination of infrastructure, training, and habit has kept QWERTY as the dominant arrangement for the most central input device, perhaps, of modern life. Even modern touchscreens, in which the keyboard is not a device but an arrangement of pixels, arrange those pixels following the guidance of nineteenth-century telegraph operators. We text between mobile phones and post on Twitter using one nineteenth-century technology while the character limit of those tweets—even now that it has been doubled—stems directly from the informational limits of early mobile texting technology.

Twenty-first century computing is replete with such historical references, both in its hardware and software. Some of these references are metaphorical, such as the “files” and “folders” we navigate and manipulate in both graphical and textual operating systems. Some are aesthetic, such as the typography we use for both online and analog communication, much of which comes directly or is adapted from letterpress typography. Finally, some are material, such as the layout of keyboards referenced above, or the shapes of e-readers and even mobile phones, which correspond—perhaps inadvertently in the latter case—to common shapes and formats of historical media. Some of these elements can be identified as skeuomorphs, particularly the graphical elements of computer interfaces that emulate the visual characteristics of previous media. Designers endlessly the debate the value of skeuomorphic design, which waxes and wanes in popularity, with some arguing that skeuomorphs help users understand the use and value of new media, while others argue that skeuomorphs hold new media back, preventing designers and programmers from taking full advantages of new capabilities.

In the most strident calls to abandon historical reference points and fully exploit the affordances of current media—or better, what Bruce Sterling called “infant media”—I would identify a selective presentism, founded on a partial timeline that only records the most visibly successful transformations in technological history. By contrast, Sterling’s proposed Handbook of Dead Media sought to document

the failures of media, the collapses of media, the supercessions of media, the strangulations of media…all the freakish and hideous media mistakes that we should know enough now not to repeat…media that have died on the barbed wire of technological advance, media that didn’t make it, martyred media, dead media.

The working notes of that project remain an invaluable reminder that our genealogies of media technology ignore the vast majority of its evolutionary branches, thus making contemporary successes seem inevitable, while contemporary failures seem all the more unaccountable.

the printing press is a ubiquitous analogy for the computer in part because it readily summons narratives of seismic innovation. Even the language of “obsolete” technology and “new” media reaffirms these buried metaphors on a smaller scale…the popularity of these metaphors may best be attributed to marketplace pressures in Silicon Valley and the academic marketplace of ideas.

As a metaphor, rather than an embodied machine, The Printing Press™ becomes a sudden and singular “revolutionary technology” rather than the assemblage of diverse technologies, practices, and social structures that shifted over time within a constellation of other technologies and media. The metaphorical Printing Press™ has no dependencies, and conveys no dependencies on the contemporary technologies it is invoked to advocate.

Referencing only particular historical technologies as moments of sudden rupture obscures the interdependence and overlap of media technologies from the past through the present. Even the debate over skeuomorphism focuses only on the most visible connections between old and new media, when in reality media imbricate across visible and invisible registers. Alan Liu famously argued that narratives of new media “‘conversion’ connote…too right-angled a change” when a “better term is indeed ‘encounter,’ indicating a thick, unpredictable zone of contact—more borderland than border line—where (mis)understandings of new media are negotiated along twisting, partial, and contradictory vectors.” Consider the seemingly endless debate over electronic books. For more than a decade, we have seen no end of pieces claiming that e-books will replace print, declaring e-books dead when print sales rebound, mourning the lost potential of e-books, or defending the intrinsic value of print. As I write, a paper shortage imperils print book production even as many lose patience with ebooks entirely.

Just seach “will ebooks replace print” to see a vast range of opinions on a question that entirely misses the reality Liu and other scholars identify, which is the “thick, unpredictable zone of contact” between these two media. Ebooks have shifted our ecologies of reading, but have not supplanted printed books, and do not seem poised to in such a totalizing way. Instead, both print and electronic books now meet distinct readerly needs, desires, and preferences. Moreover, as scholars such as Whitney Trettien and Matthew Kirschenbaum remind us, in the era of print-on-demand, the line between electronic and printed books is not so easy to discern, while others, such as Élika Ortega, highlight new works that bind modalities together. Writing about electronic texts and libraries in 2006, Jerome McGann argued “We aren’t even close to developing browser interfaces to compare with the interfaces that have evolved in the past 500 years of print technology. A sophisticated, flexible, and stable system of graphical design and bibliographical codes stands ready for the scholar wishing to build a critical machine for the complex analysis of textual works.” I invoke this quote not to disparage current experiments with electronic interfaces, but to emphasize what a historically-informed praxis might offer information scientists, which is a deep well of knowledge about how people apprehend, respond to, and appreciate material presented through existing media technologies.

It might seem a tangent, but consider the movie Spiderman: Into the Spiderverse, which cites, through largely computer-generated art, characteristics of off-set and risographic printing processes, such as mis-registration errors. As explained on Sony Pictures Imageworks’ website, “One alternative type of camera lens focus (or DOF) was inspired by the comic book offset printing process where the (often accidental) mis-registering of the color passes can make an image look out of focus.” Mis-registration, an outcome of earlier information technologies and a mistake, at least originally, is now core to the visual rhetoric of comics, a kind of aesthetic dependency. We might call the art style of Into the Spiderverse skeuomorphic, but that seems too dismissive of a reference that is more deliberate, evocative, and beautiful. In an animated comic-book movie, those risographic elements ground the new cultural production in the history of its medium, providing a stable visual reference point that facilitates the other forms of visual experimentation also happening in the film. To oversimplify a bit, those registration errors have become part of what viewers like about comic art, and they cannot be “corrected” without sacrificing those attachments.

To return to my opening claim, we might instead acknowledge that new media technologies are shot through with social, political, and aesthetic dependencies on historical media technologies. In software development, as Russ Cox describes “a dependency is additional code that a programmer wants to call. Adding a dependency avoids repeating work already done: designing, writing, testing, debugging, and maintaining a specific unit of code.” When writing software, programmers use existing “packages,” “libraries,” and “modules”—skeuomorphic terms, all—to make use of existing functions, support existing hardware and software, or otherwise extend their code without unnecessary work: “For a program that needs the functionality provided by one of these packages, the tedious work of manually downloading, installing, and updating the package is easier than the work of redeveloping that functionality from scratch.” Dependencies are necessary for much modern software development, though when adopted without sufficent care programmers can find themselves in dependency hell.

Media dependencies provide a useful framework for thinking about the ways that historical media technologies weave through contemporary media technologies, by turns benefiting and limiting the latter. We might wonder, for example, why people spent so much money in the first decades of digitization designing page-flip animations for online and electronic texts or we might mock, as Clive Thompson did recently the “hilariously retro skeuomorphs” of Apple’s iOS and OS X applications. Thompson argues a strong position against skeuomorphic design, insisting that “skeuomorphs also wind up hobbling the new invention. Because skeuomorphs are based on the physical limits of an old-fashioned device, they get in the way of a designer taking full advantage of the new realm.” If “old-fashioned” media ideas interfere with innovation, why teach them to burgeoning information scientists at all?

I would suggest that considering media dependencies helps us apprehend skeuomorphs along a continuum of other historical entanglements among new and old media. In some ways, what we eventually identify as “skeuomorphic” are the media elements that fall away—I’ve heard few arguments to stop calling computer files “files,” for example, or to ditch Garamond entirely in favor of a fully digital typefaces—but the long history of media change indicates that humans are quite bad at knowing, at the outset of a new medium, how it will be used, by whom, and to what ends. Media dependencies bridge old and new media, helping users across the span reckon with both sides. If McGann is right, as I suspect he is, that it will take time to develop electronic interfaces that function as well as analog interfaces developed over hundreds of years—and if we acknowledge that the history of media change comprises mostly failed experiments, dead ends, and short-lived fads—then we should not expect to know at the outset what features are necessary or even wanted in a new medium. The experimental phrase of early printing, when the texts we identify as incunabula were produced, lasted at least sixty years, which would place us well within the incunabula period for digital texts, at least as a popular medium. We should expect false starts and recursions.

When Amazon released the Kindle, for example, it initially removed page numbers from its electronic books in favor of location numbers. The decision makes logical sense: what would a the skeuomorph of the page number even mean in a book that can be reconfigured on the fly as readers change the font, type size, and more? A Kindle book has no pages, so what good are page numbers? Soon, however, readers protested: particularly readers who wanted to share a book within a community, such as a classroom, in which different readers would access the book in different formats. Readers wanted to be on the same page, both literally and metaphorically, with readers in multiple modalities, and soon Amazon added page numbers back into Kindle books, though users can now toggle between page and location numbers. In classrooms, teachers and students alike are adapting to other ways of locating the same spot in a text, such as using CTR + F to find a key word or phrase. Perhaps page numbers will eventually disappear entirely in favor of search, but it will be a slower process than Kindle designers, who did not consider the dependencies required for certain forms of communal reading, imagined.

Liu’s essay reminds us that neat accounts of one media—or one technology—quickly and totally replacing a prior one are rarely true, or at least rarely so tidy. New media technologies shift an existing ecology, but old media persist practically, even spiritually. In 2021, analog technologies continue to shift and mutate alongside the computer, informing the evolution of computing—just as manuscript cultures mutated and thrived alongside while informing the evolution of print cultures. As Liu insists, “The déjá vu haunting of new by old media is clear enough.” Information scientists may not think of themselves as spiritualists or mediums, but they spend their professional lives immersed in haunted spaces, wrangling haunted machines. An ability to recognize, understand, sometimes call on, and sometimes consciously purge media dependencies can benefit information scientists enormously, both intellectually and practically. The goal of a dependencies framework for media history in information science is not nostalgia, but cultivating a richer technological imagination. We need information scientists who are not only able to use current media technologies, but who recognize the historical contingencies that led to this present; who are curious about the many branching paths that failed, were abandoned, or were actively foreclosed; and who can imagine more humane alternatives to our present that they can help create.