During this semester, we’ve understood data to mean numerical facts and statistics. Scholars can represent those descriptive numbers and aggregations along two dimensional coordinate systems and maps, communicating an understanding of the texts or medium under study by abstracted details into summaries. Contributions like Manovich’s Exploring One Million Manga Pages with Supercomputers and HIPerSpace “ use visualization and/or mathematical models to describe the space of possible and realized variations” by clustering the amount of gray scale or measuring the level of detail in illustration among 1,074,790 pages. Others employ methods like distant reading to process words and sentences in large corpora as way of confirm the coherency of writerly social characteristics, and ponder the examples which evade their models (see So and Roland’s stellar Race and Distant Reading). These flavors of analysis view data as a means push away from subjective knowledge of artistic output, proposing dimensions and measures as appropriate tools for grappling with humanities as scholarly sport. The quantization champions of digital humanities offer mostly rigorous and often legitimate intellectual insight to their subjects of study. Still, there is a more capacious views on data, a more digitally native observation, beyond the statistical definitions we’ve been exposed to this semester.

CUNY’s own Kevin L. Ferguson stands as starting point for many members of this minority group of data “analysts.” Looking at his frequently updated tumblelog Film Visualization, it’s easy to mistake his methods as art (in fact, he courted some controversy when an artist appeared to ape his methods). As a part of his scholarly work, Ferguson produces summed frame visualizations. Screenshots of a movie are superimposed on themselves by loading them into the open source medical imaging software ImageJ, provided by the NIH.

While comparisons on the corpus and categorical level are made by “automatic process which reveals otherwise unconscious information about film texts” (DHQ: Digital Humanities Quarterly: Digital Surrealism: Visualizing Walt Disney Animation Studios) what differentiates these summed frame ouputs from more common visualization methods is two fold. Firstly, this method takes digital data, in the form of images, as material from which to create a subjective understanding of the objects under study. While there are numerical transformations involved in the production of these images, the artifacts generated require context and knowledge of the specific films under analysis, and are not summarized by statistics. Contrast, color and framing of subjects can be determined for each movie, or for the movies of a corpus, but the visualizations themselves are only a component in the over all conclusions one can draw. Secondly, these summed frames suggest a mode of experiencing or “reading” that is more embodied than typical distance reading. The scholar may see a perpondance of close-ups in a given corpus rather than talling the number of such shots, or may observe the lack of vue variation in a color film.

Another digital humanist with an expansive view of data and its expression is Kim Brillante Knight. Her project “Danger, Jane Roe!”, described here, cleverly grounds itself in a tradition of feminist praxis. Knight fashioned embrodery of reproductive anatomy undergirded by LilyPads, a flavor of Arduino microcontroller for wareable technology. Little LEDs light up in the decoration based on how many #prolife hashtags are posted on Twitter over a specified period of time. Knight links the projects to many linages of practice.

The adoption of decorative textiles, an artform fogged by traditional modes of feminitity, for a politically potent topic is one example. Leaving the application stitching visible by using a contrasting color with the reproductive organs “can make explicit the workings of a circuit” more visible by implication, implying “polarity, connectivity, and flow” that most digital hardware products obscure. A key aspect emphasized in the wearing this visualization on the body is to “remov[e] data visualization from the screen or page…relocates discourse around reproductive justice onto the site of legislative inscription—the body that may be affected by pregnancy.” If you read the article linked above, you’ll be treated to a thick fabric of theory and practice that legitimizes what could be glossed as an non-academic pursuit if you aren’t considerate or don’t read the whole article. As it relates to an expansive sense of data, we can appreciate an embodied experience of data that has it’s origin in technology on the screen, but seeks to escape it. Once can imagine a version of this wareable, perhaps untenable for long periods of usage, that vibrates when a word from a hashtag dictionary is posted. Imagine that as an embodied experience.

Writing about a large sense of what data can mean in a digital humanities project leaves to invigored by how many potential new modes and methods are left underexplored in a field. Digital humanities, in whatever manifestation or metaphor of organization, seems to experimentation and not wedded to received ideas of scholarship, which I find appealing.

I’m proposing the creation of a digital archive for textual and visual material related to mainframe computers. As a part of the curation process, we will perform glitches or additional manipulations of the material as a method interrogate the assumptions we have about business computing and mainframe computers.

Both minicomputers and super computers are well understood in academia. Digital humanities emerged in a period when computing costs dropped precipitously through innovations related to the microprocessor. Scientific computing relied on supercomputers addressed on complex mathematical calculations. Although its birth place can be traced to the academy, mainframe computers mostly found application in business contexts. Given its pedigree, mainframe computing is an under explored topic in digital humanities. Yet concepts made popular by mainframe computing live on in today’s software infrastructure in the form of batch processing, multitenancy in systems, timesharing of compute / storage resources and transactional databases. The age of Big Iron also continued the contribution of women to information technology often made less visible than those made by men. For example, the theory of machine-independent programming languages and the creation of FLOW-MATIC language by Grace Hopper, which was extended as COBOL, the principal program language used program mainframe computers.

This project will handle subject matter germane to the many facets and specters of digital humanities. Using frameworks like Wax from the minimal computing group ensures some longevity of the archive in question. Mainframe computers continue to power the majority of credit card transaction processing. The prominence of this type of hardware in governmental payment process came to light when stimulus checks weren’t issued in a timely manner during the pandemic. Maintenance (or lack thereof) of our critical technology is a major concern of digital infrastructure scholarship. Historical research will be critical to reconstruct a cultural particularities of time and place. It would also be advisable to

I am also keen on inviting collaborators with different skill sets. Some may want to work through the issue of cataloging, while others write essays explaining how mainframes function, what work they do or what the outputs are like. If we want to take a deformative approach to the archive, we could have someone coming from a design perspective (Adobe tools, PhotoMosh), another from an informed naif (data bending with Audacity/converted text file), while a third collaborator could deploy an algorithmic approach (software libraries like glitch-this or pixelsort). Regardless of how we approach the archive, I expect curation and critical writing will be instrumental in guiding audiences through the material.

One of the hallmarks of Johanna Druckers’ scholarship is an interest in how knowledge is produced. In SpecLab, her experiments at the titular laboratory centered on “reconceptualization of premises and parameters, not a reassessment of means and outcomes.” This sentiment gives a generous acknowledgment of the benefits computer-assisted quantitive methods made in humanities research, while also recognizing the “invisible” constraints orthodox methodologies inherit from the disciplines and industries of origin (Drucker, 27). Recent publications build on some of her earlier experiments to demonstrate that what is presented to us as self-explanatory artifacts actually require more thought, context and interpretation than we’re trained to deploy in our every day encounters with, say, data visualizations. Along similar lines, Drucker draws our attention to the hidden costs of the transition away from physical media to digital media for academic publication this week. One small nuance present in Pixel Dust caused me to reflect on why I’m prone to accepting Drucker’s lines of augment.

Drucker demonstrates an understanding of relevant digital technologies outside of an academic context which views the “back-end” as cruical to the frontend. The thoroughness of comprehension merges in the specificity of one example used when description some misunderstandings of the transition to digital media for academic publication. The company named in “the misperception…that everything digital is available on Google” strikes a chord with me given the origin PageRank, the key algorithm in that companies competitive advantage in the 1990s IT tech ecosystem. I would argue that Drucker knowingly references Google because of the influence of citation analysis on this algorithm, borne out of finding a quantitative method of evaluating the importance of published scientific articles. As Google put describes the logic:

PageRank works by counting the number and quality of links to a page to determine a rough estimate of how important the website is. The underlying assumption is that more important websites are likely to receive more links from other websites (citation courtesy of the Wayback Machine)

Much like her example of the codex, Drucker emphases the continuities in the backend logic between older forms of media and their descendants. In this case, leaving discovery up to Google search results likely reinvigorates the same forces that constricted traditional academic publishing, both financial and institutional, through backlinking other SEO strategies that favor

Drucker’s implication of the backend as critical subject matter to understand the transition to digital formats reflects the true costs, rather that the price. One misconception has to do with the distribution of costs over the lifetime of an academic product. Whereas physical books and journals had large upfront costs, the cost of digital publication goes beyond licensing fees, or even the “author-pays” model mentioned in Fitzpatrick’s chapter in Generous Thinking. Drucker also counts the cost of maintaining the digital medium itself, often not appended to price tags. To draw the analogy using a physical book, it’s as if there were occasions where the pigments in ink stopped rendering, and someone needed to jostle them to recompose as letters and text on a page. Or the pages would occasionally fall out of the binding, become disordered on the ground.

I think there are reasonable responses to these problems offered in the motivations behind Gil’s Ed. And in general, the reading from this week and last about infrastructure shows a gradual movement towards systems thinking. As someone who’s being propelled rapidly into a more “backend position” in a professionally, I find a lot of sympathy with Drucker’s instinct to scrutinize the complex tradeoffs in the “invisible work” required to create and maintain digital platforms.