A 35mm Kodak slide is a picture of a picture. On the first floor of Street Hall, Yale’s Slide Collection houses 320,000 pictures of other pictures. To hold them necessitates a library like a varnished jungle, a fortress of solid chest-high apothecary cabinets stocking hundreds of sleek beechwood drawers, each the width of a baby’s fist. Scanned from left to right, the labels read like obtuse verse:
But at 11 am on a rainy weekday, not many people are here to read them. A single professor idly raises a slide up to the light, squints, then snaps the drawer shut. “Most people these days would rather look at these images on the computer,” says Katherine Haskins, the plum-haired Library Director currently embroiled in a battle to scan the collection’s 380,000 slides and photographs. Essentially, Haskins is in the process of taking pictures of these pictures of pictures. The digitization of this entire library is a quest so vast that an end date cannot be projected. “We’re in a bit of a protean state,” Haskins admits, “I may have long retired by the time this has come to fruition.”
On this day in January, the staff here is silently celebrating. Today they reached the fifty thousand image mark, or 50,117 to be exact, according to the tracker on the online database that stores the scans. “Quid Tum,” Haskins shrugs, quoting Alberti, and the department’s motto. (Translation: “What next?”)
Of all the Yale libraries to undertake digitization, Visual Resources’ Slide Library has the least choice in the matter. This past October, Kodak manufactured its last 35mm slide projector. The clicking, dusty eyes holding vigil in the back of Yale classrooms will be gradually phased out, to be replaced with digital projectors.
The slide library is just one branch of the Yale library system that is currently transitioning from hard copy—videocassettes, magazines, books, specimens, oil paintings, Nobel prizes, papyri, death masks—to a digital, on-line database. While traditional on-line library searches yield only book titles and journal dates, over the past ten years librarians have sought to digitize complete holdings, so that what is on-line is a simulacrum of the entire collection in digital images. For example although a scanned Edith Wharton manuscript certainly does not retain the tactile value of the original, the digital form of the manuscript would still be very valuable to a scholar in New Mexico who is unable to travel to Beinecke Library. The slide library is creating a digital copy of itself in cyberspace—a second library that isn’t comprised of varnished beechwood drawers, but of ones and zeros.
Digitization is a gradual, painstaking process currently unfolding in the backrooms and basements of each of Yale’s disparate libraries. There is no central digitization headquarters, as it would be a nightmare to move and track all these objects. Instead, each library scans or photographs its own collection and uploads it on-line into a separate search engine. Yale’s digital library site lists twenty different searchable databases—from the Law School’s “Curiae Project,” which provides Supreme Court briefs in full text, to Beinecke’s, Yale’s largest digital collection boasting over one-hundred thousand images. Although separate teams toil in separate libraries, the library system plans to coordinate these databases so that in the not-too-distant future, a grand database that might integrate Haskins’ ambitious slide digitization project into Yale’s larger collections. The final product would be a search engine like Google, and a grand object in itself that would stand as a monument to the breadth of knowledge available at Yale.
This is the next step in a digitization project that is just over a decade old. In 1994, Yale’s Project Open Book was lauded for its ambitious aim to convert ten-thousand texts into digital formats. Today, the Beinecke library holds onehundred thousand digital images—including of 12,000 papyri scrolls and 660 exlorers’ maps. The Manuscripts & Archives collection, in comparison, works on a smaller budget to expand its current stock of ten thousand images. It has taken Haskins and her team eight years to scan the fifty-thousand slides, and with 270,000 to go, she often feels like an ever-idealistic Sisyphus as she clicks through the slides in a drawer. “We acquire more images per year than we currently digitize,” she said.
To translate a slide from a tiny plastic window into the language of pixels, Haskins sends orders downstairs where, amid the old stone of Street Hall’s 1870s basement, a bank of Eames chairs float around a coffee table stacked with issues of Wiredmagazine. This is Yale’s Media Services division, which hired Photographic Services Manager Joseph Szaszfai as a photographic assistant in 1964 to shoot Museum exhibitions for print catalogues and posters. Szaszfai has spent the past thirty years shooting the slides that comprise the Slide Library’s upstairs collection, and his new role almost seems like a renunciation of his old work. This past academic year, his team at Media Services was responsible for creating 39,000 digital images for use by faculty in class or publication, many of which were digital copies of slides he shot himself.
“This team has basically gone from silver to pixels,” says Media Services Director Peggy McReady with visible enthusiasm. “We still produce images, but five percent are requested on traditional film base, and 95 percent are requested as digital. That’s the opposite of what it was four or five years ago.”
“For some reason,” Szaszfai agrees, “It’s all happening this year.” This summer, Media Services sent out a pamphlet to Yale Faculty encouraging more sentimental professors to “go digital,” to request digital images for class and laptops to play them on. “There are folks that have been here for years and years and years, but they may have to change,” Szaszfai said. “These projectors are going to break, and there won’t be another way to do things.”
With changing technology, the layout of Media Services is also changing. The computer lab Szaszfai careens through was once a screening room for 16mm projectors. A darkroom down the hall is shuttered and empty. “Just today a guy came by and turned off the water.” By July 1, 2006, the division, which for the past forty years has printed photographs chemically, will be run entirely through a networked system of computers and printers. The basement’s pride and joy is Kodak scanner that can scan an entire carousel—at a rate of 85 slides per hour. On a Wednesday in January the entire basement is empty but for two technicians huddled in a windowless nook the size of a bathroom, gathered around a large computer screen. They watch as the machine whizzes through a carousel of images of Italian architecture.
Once these slides are digitized, they’ll first be handed over to the professor, either via e-mail or on a data disc. Secondly, they’ll be sent upstairs to Haskins. Her team will upload the images into one of several databases and will enter alongside what librarians call “metadata,” a string of information about the image—such as its artist and dimensions— that allows the image to be searchable in a larger database.
But few undergrads, unless strong-armed by an art history course, know that the resource even exists. “Currently, to find what they’re looking for, students are forced to wend their way by human beings,” says Katie Bauer, a former New Haven math teacher who was hired by SML under the daunting title of “Integrated Interface Systems Librarian.” Her job was to procure and coordinate electronic databases like LexisNexis from vendors. Today, Bauer, along with Jennifer Weintraub, a classically-trained librarian who fondly recalls learning Netscape 1.0 in Library School, is working on a project called “DPIP.”
“I was trying this whole time we were sitting here to think of what it stood for,” Bauer admits. “Digital Production Integration Program?” The program aims to integrate the images that Haskins’ team uploads with all of Yale’s digital collections. Currently, digital holdings in the Beinecke division of Sterling Memorial must be searched through a separate database from Manuscripts & Archives’ holdings. Though Yale’s libraries will always be a “decentralized system, no matter what,” Bauer says, “clearly we need to build something to coordinate groups.”
As far as librarians go, Bauer and Weintraub are particularly progressive. “Librarians are critical that students will only use what’s available digitally,” Bauer says, “They say it’s a bad thing that Google is so easy, and that people are getting spoiled.” Most librarians would agree that students should “wend their way” through the system that’s been in place for years—mastering card catalogues, toying with Orbis. “But when a student asks ‘Why don’t we just have something like Google, that searches all the information systems, or something like Amazon that recommends other books or sources,’ it’s hard to give an explanation. The answer is, we should be using these techniques. We have a lot to learn from what students already know.” DPIP hopes to integrate all digital holdings into one search engine. “You’ll type ‘Gertrude Stein’” Weintraub explains, “and you’ll get images and text from all libraries.”
Librarians are also wary of how simplifying the process may affect scholarship. “There’s a fear that if some images are on-line, those will be the only images students use,” Weintraub explains. “They’ll never go looking further.” Currently, Yale doesn’t have the capability to digitize every library holding, so libraries must prioritize what gets translated into a new format, instead of pursuing a wholesale conversion. They first digitize items requested by scholars, then scan items that receive the most use (freeing up librarians who would otherwise spend time pulling those items); their next order of business is to scan items that are monetarily valuable and thus more likely to be stolen—like the jewel-encrusted fraternity pins currently being digitized at M&A. Finally, they scan items that are so perishable, digitization serves as a mode of preservation.
Because of the extent to which digitization depends on demand, librarians worry that it will negatively affect scholarship by creating what Weintraub describes as “a self-perpetuating cycle”: The images that get more use will become the easiest to access and thus will get even more use. Just as an art historian fears for the integrity of copied images, so too librarians fear that a digital database will provide a false picture. “The worry is that this will become a ‘greatest hits’ collection,” says Richard Szary, the Director of Manuscripts & Archives, who last year was recognized by the Society of American Archivists as one of the “quiet giants of the profession.”
“What’s on-line is just such a small portion of the actual collection, and is completely unrepresentative,” he says modestly from a low-ceilinged back office of the gothic Manuscripts & Archives department of Sterling Memorial Library. “What we need to deal with is wholescale digitization, which requires a wholesale operation we’re not equipped to handle. But currently, this is a skewed picture of a collection.” A search for Eero Saarinen, the architect of Morse and Ezra Stiles College, brings up 179 hits. “It’s just totally skewed,” he stresses, “We probably have ten thousand.”
“Our collection is enormous,” says Matt Shirley, the Digital Imaging Project Manager for Beinecke, “but when you think of someone’s archives in a box as hundreds of separate images, this is a very small portion of our collection.” Shirley’s team photographs about fifty items per day, though he explains that “3D is much slower. Busts, swords, medallions, death masks, and Nobel prize medals take a lot of time.” Translating the image of a three-dimensional object onto a two-dimensional on-line database is its own challenge, and is the most obvious example of what is lost when students and scholars neglect the original object for the facility of a digital image.
Still, librarians are already noticing that resources are getting used by a broader audience. When Rebekah Irwin, the catalog librarian for the Digital Projects division, entered data to describe “Unknown Man from Alaska,” an 1890s photograph of an Inuit, she couldn’t upload very much “metadata” because she didn’t know very much about the image. Months later, she received an email from a Juneau, Alaska man who could barely restrain his glee at finding an image of his relative online:
“The image is of Rudolph and Daisy Walton and family, ca. 1890…This is the only known photograph of the Walton family and its discovery was met with much excitement among the Walton family in Sitka, who had never seen a photo of their grandmother, aunt and uncles (all of whom died young).”
This is one of the scholarly outcomes that exemplify why Irwin and her collaborators at the Beinecke laud the new digital database. “Correspondences like these are very rewarding,” Irwin wrote in an e-mail. “In this case, the digital reproduction of a rare photograph transcended the four walls at the library and as a result, we now have a deeper understanding… based on the shared insight of a far-flung user of our collection. Researchers often help interpret items in our collection,” she added, “that is not new. What is new, however, is the elimination of the space and time constraints of traditional libraries.”
By widening access to these materials without causing damage to the originals, the Beinecke digitization team is fulfilling a two-fold goal of deepening scholarship about the materials within the collections while simultaneously broadening the viewing audience for that collection.
Unlike Haskins, who is chiefly concerned with digitizing an analog resource—converting slides—Szary is building a digital collection from analog and other media with digital originals. “What do you do,” he asks, “when a writer’s correspondences that they donate to the library aren’t on a piece of paper anymore, but they donate a collection of emails? How do you preserve the order they were sent in? Where do you put the e-mails, how do you ensure that they aren’t tampered with, how many copies do you make and do you save them on a CD or on a server? Nothing lasts forever,” he says, “But it’s hard to know what will last at least a long time. We don’t want any 8-tracks or Word 4.0 files, because nothing we have can access them.”
Just last week in Manuscripts & Archives, a miniature robot named SAMMA (System for Automated Migration of Media Archives) finished the tough 18-month job of re-recording 2,000 hours of testimonials from Holocaust victims that had been in the library’s collection. Each night, a technician would load the robot with hours of blank footage, and each morning he would wake to remove hundreds of hours of exposed tape from SAMMA’s hungry belly. “When we applied for the grant, digitization was not even an option,” said project director Joanne Rudof at a recent conference, “because there was no acceptable digital format for video.” Now, the library has hundreds of copies of small DV tapes—a format that is ostensibly outdated. “In the case of converted materials,” Szary says, “I look at it as a manager protecting an investment. We put a lot of money into it, and if this gets lost, that’s a lot of work and a lot of hours that are lost, let alone in a preservation case the loss of the document itself.”
Such failures have occurred since the library’s 1994 inaugural digitization project, Project Open Book, whose 1992 proposal outlined a “vision of the research library of the future as an institution whose mission is to generate, preserve and improve for its clients’ ready access—both intellectual and physical—to recorded knowledge. The place of electronic tools in the library of the future will depend on how well they measure up against this mission.” Unfortunately, these books were recorded on a system that has since become obsolete. “No one can look at anything they did,” laments librarian Katie Bauer, “Cornell’s similar program had the same problem. They recorded everything on these giant silver laser disks, which of course nothing can read.”
The prevailing plan for keeping pace with constant improvements in media technology is to save digital objects to a server, and then back this server up periodically, on different forms of media. Any changes made in the conversion from one medium to another will be tracked using programs developed by the library.
In a skinny black turtleneck, Katherine Haskins stalks the hedgerows of slide cabinets with grim purpose, wrenches open a drawer and, between two careful fingers, pulls up a plate of glass as if gently lifting a strand of pearls. Flicking the sideswitch of an illuminated viewer with the other hand, she sets the glass pane upon the lighted square, revealing the ruins of a 19th century Italian palace in fragile miniature. The crisp, black-and-white pane is a magic lantern slide, the precursor to the celluloid 35mm Kodak slide that cycles through today’s art history classrooms. “These were outmoded in the 1950s or 1960s,” Haskins explains, pointing toward the image with a pinky. “But color wasn’t always spot on with the 35mm slides, so in most cases they stuck with lantern slides for a few years. In some ways,” she reflects, “this is how we think of digital media.”
Digital media teeters in the same purgatory, straddling a brink of technological progress where most scholars are unwilling to part with the old system, but would be hesitant to argue against the efficiency of a digital database, where slides can be rifled through by keywords instead of drawer knobs. Librarians describe the website as a kind of “electronic wailing wall,” referring to the database’s analog format: a long, studded white wall, showcasing the images students must memorize for exams that stretches the length of Street Hall’s second floor. Today, like the deserted slide library, the wall is frequented more by ghosts than by grade-hounds, but it is maintained out of a sort of sentimental interest in the preservation of an older world of scholarship. When the Cross Campus Library below Sterling is renovated this May, the enormous card catalogues in the first floor will be moved. Rarely used, they will be preserved as art objects, saved out of interest in their status as artifacts. Removed from the burdens of function, they can be appreciated more wholly for their form, for the human effort it took to create them.
Art history is a history of the artifact, and any slide or image memorized in a class serves, Haskins explains, as a “simulacrum with a sense of the value of the original, the ‘ding an sich,’ or ‘the thing itself.’ There are so many ways to understand the physicality of these objects.” Analog formats like the lantern slide are kept less for art history scholars interested in the artifact, but for scholars interested in the history of art history—what Haskins describes as the “historiography of art history at Yale,” as she returns the lantern slide to its drawer with a sentimental smile. Though the lantern slide serves as a simulacrum of its image, it is also its own historical object, its own ‘ding an sich’ worthy of study. The artifact presently being created by Yale’s hardworking librarians is a digital database, and in one sense is just another simulacrum— a sprawling two-dimensional image of the sprawling library itself.
Sentimental professors might be squirming in their tweed, but Haskins couldn’t care less. “Someone once explained it to me as ‘You go where the stuff is,’” she says, adding with a smirk a quote from a book she recently read by W. J. T. Mitchell, who saw digitization as an eternal battle: “The shock of new media is as old as the hills.” “New media” is the chief interest of art history, and society’s gradual progress is evident in every object we create and in every slide stocking these thousands of drawers, awaiting the new processes that will fight to preserve them for the next generation.
Adriane Quinlan, a Junior in Calhoun College,is a Managing Editor of TNJ.