This project introduced me to the necessary—and often frustrating—preliminary steps involved in getting an archived-based DH project off the ground. The original plan was to use OCR to read digitized copies of the magazine. First question: are the materials already digitized? At what cost? Although a digitized archive was found to exist, it was beyond the budget of such a new project.
Our next attempt was to use microfilm, which we were able to procure via Interlibrary Loan. However, these materials turned out to be inadequate for our uses, since the magazine had been recorded at a very small scale: between 2 and 4 issues vertically per 35mm microfilm reel. Once scanned and viewed as PDFs, the text of the schedules was unintelligible to the computer, and often to us as well.
Next, we turned to the magazine itself, and thus to eBay, where we were able to purchase a selection of issues from each decade. After scanning the magazines, we were able to OCR most of the text, but many of the TV Guide-specific elements, such as times and channels, were not picked up. The evolution from a two-column layout to a grid and the corresponding shrinkage of text brought diminishing returns.
Given these setbacks, I decided for the rest of the semester to see what we could gather from the OCR files, rather than dwelling on what we couldn’t. I thus found myself turning away from time and toward more accessible questions of genre. From 1952, when the magazine was started, through the early 2000s, TV Guide included genres alongside many of its listings. These genres, however, were inconsistent, both in their identity and their distribution. Using text-searching software, I collected data about genre and published the preliminary results as a research report on the Mellon website. For me, the most interesting outcome was the idea of calculating genre distribution in minutes: how long, for example, one could spend watching all of the comedies on air in a given week (hypothetically or with the aid of recording technology). For 1998, this number—13,560 minutes, or 226 hours—far exceeds the total number of minutes in a week, being 10,080. An odd comparison, to be sure, but intriguing nonetheless.
TV in the Seventies
What drew me to Bernardi’s project was the dynamic between material and digital that is central to the archive’s mode of presentation. Her largest collection is that of postcards, printed images that—like photographs—are easily read as illustrations when digitized, rather than objects. Working with the team, I was able to see first-hand how collaboration between humanists and computer scientists can produce remarkable results, such as the high-resolution viewer plug-in developed specifically for the site. That this plug-in, which I can only imagine will be used by many others in the DH community, was developed in the specific context of a materially-concerned archive in some ways represents a reclaiming of the dangers of prescriptive software, as it will spread some of that same emphasis on materiality to countless other projects.
I joined REJ in the middle of the site’s transition from WordPress to Omeka. Following the stages of this transition exposed me to important conversations about best practices in file management, how to determine metadata fields, and other related issues that comprise the behind-the-scenes critical thinking of any digital project. In short, just as a two-dimensional image does not necessarily neutrally convey a three-dimensional object, an item’s metadata does not neturally convey the item itself.
When I first started with REJ, I knew little about Omeka except that its collections management and exhibition creation tools might be applicable to some of my own art history-oriented projects. My experience with REJ encouraged me to explore the platform more in-depth in regards to my own discipline, since Bernardi’s project is based primarily in the methodology of Japanese cultural history. In the process, I have become aware of ways in which Omeka is geared toward the discipline of history in which it was created and therefore motivated to focus on how art historical questions both work and do not work within the platform’s affordances (see Histories of Photography).
Mind and Metadata
Inspired by REJ’s engagement with Jules Prown’s theory of material culture, I am currently writing an article that explores these ideas in conjunction with his pioneering work with computer databases in the 1960s.