Pressed into the Service of Cinema: Issues in Preserving the Software of Hollis Frampton and the Digital Arts Lab

The landscape of the epic attempt in art is of course dotted with ruins, and some of the ruins are as noble as completed work. In fact, some of them, of course, are things that one cherishes very much.

—Hollis Frampton

From 1977 to 1984, filmmaker, photographer, and writer Hollis Frampton turned his creative attention to the aesthetic potential of the computer. Frampton is best known for his conceptually rigorous films and writings that investigated the origins of media forms with a focus on the relationship between language and image. Despite now being seen as a preeminent practitioner and theorist of the photochemical arts in works such as (nostalgia) (1971) and his 1965 essay "Some Propositions on Photography," Frampton expressed numerous reasons for incorporating the digital into his artistic and critical efforts. The computer in its very nature of operating algorithmically had clear resonances with his avant-garde practice based on analytical dissection and repetition. It represented the realization of his passion for bridging science and the arts. Furthermore, Frampton hoped to introduce a humanizing contingency into the overregulating impulse of computer technology.

"UB Professor Hollis Frampton at computer console . . . making art with computers." Photograph from a Courier Express article on the Digital Arts Lab, dated October 24, 1980. E. H. Butler Library, Buffalo State College Archive.

Frampton took advantage of the new accessibility to digital technology offered by the introduction of microcomputers such as MITS's Altair 8080 (1975) and Processor Technology's Sol-20 (1977). In 1977, he cofounded the Digital Arts Lab (DAL) with video artist Woody Vasulka, under the auspices of the University of Buffalo's Center for Media Study (CMS), where both were teaching at the time. Under the direction of Gerald O'Grady, the CMS provided students an education in experimental media practice that was technologically up to date, theoretically advanced, and media agnostic. Though the DAL added digital media to the CMS's pedagogical training in film and video, Frampton's interest in computer art should be contextualized within the intermedial explorations of both the CMS and the larger upstate New York electronic media arts community.¹ Additionally, Frampton was by no means the only media artist exploring the computer at the time. Vasulka's efforts in digital art were perhaps more successful than Frampton's and are equally deserving of preservation and scholarly investigation, though such an effort lies outside of this article's focused ambit.

At a time when personal computers came in kits requiring a great deal of technical skill to assemble, and there were few, if any, off-the-shelf programs to manipulate sound and images, Frampton and his students constructed a functioning computer system that allowed artists control over digital media at their constitutive elements. With original DAL software such as DEMON (1979) and IMAGO (1981), audio and video could be acted on at the level of the sample or pixel. This accomplishment, advanced for the time but quotidian now, was made possible through the creation of one-of-a-kind hardware such as the DAL's frame buffer, GOLEM (1981-83).²

Owing to Frampton's untimely death in 1984 at the age of forty-eight, and outside advancements in computer technology overtaking the DAL's technical and aesthetic achievements, their software never attained widespread implementation. As such, this significant portion of Frampton's late career has remained underexamined in the invigorating critical reevaluation that his writings, photographs, and films have received over the last decade.³ This article's initial attempt to redress that oversight was engendered by the collection of Frampton's eight-inch floppy disks and printouts of computer code held at Anthology Film Archives (AFA) in New York City.⁴ A discussion of conserving those archival objects is inseparably linked to a consideration of Frampton's intentions toward computer technology.

In that light, the work that follows crosses between the fields of archival science and film history. To replace the artist interview, which is an integral part of any preservation of new media artwork or computer software, an exegesis of Frampton's relevant writings, interviews, private letters, and artworks enables recommendations on preserving this unique set of digital artifacts. Frampton never published a definitive article on computer technology in the manner that he did for photography, film, and even video. As such, this narrative account of Frampton's digital turn is assembled from traces spread throughout his other works.

Going in the other direction, this essay constructs a history of the DAL's technologies and pedagogy to extend the investigation into Frampton's larger creative practice that converged multiple media toward his artistic aims. While the following historical narrative might skew toward the teleological, implying that incorporating the computer meant discarding the analog, for Frampton, the opposite was true: the digital operated alongside and in tandem with the photochemical.⁵ A close examination of the praxis of Frampton's digital turn affords a historical view of the interrelationship of film and the computer decades before the subsuming of analog forms into new media. This suggests the need to more closely integrate film and new media studies, perhaps most especially in the fields of experimental film and video art. While this article's focus is exclusively on Frampton's practice, it is hoped that it can serve as a model for the larger study of the effect of the digitization of filmic styles.

Frampton's Turn to Computer Technology

Overall, the Hollis Frampton Collection at AFA contains 146 obsolete eight-inch floppy disks, around two thousand pages of computer code and related documentation, and numerous manuals for purchased software and hardware. From the information written on the floppy disk sleeves and contained in the printed-out code, it is possible to re-create a sense of the number of different programmers involved at the DAL (twelve), their years of highest productivity in terms of numbers of programs (1981-83), and, as can be deduced from the programs created to operate various parts of the DAL's computer system, the large amount of effort spent in getting what would now seem entirely basic functions to work correctly, such as having the disk drive communicate with the computer or delimiting the pixel size of the screen.⁶ What is also obvious from examining the floppy disks and code printouts is what the collection is lacking: original hardware, including one-of-a-kind devices designed by the DAL on which to run these programs.

Where the other printouts of programs in Frampton's digital collection include only code and programmer comments, one titled TREE.ASM contains a crude image of a Christmas tree visualized with ASCII characters, including asterisks, backslashes, and hyphens.⁷ The tree is depicted growing out of a flowerpot emblazoned with the letters D-A-L. No explanation of what the image refers to is included in the printout other than the subtitle, "A D.A.L. Pictorial Extravaganza (for Nam June Paik)." Though a clear example of Frampton's wry wit—the bare image is anything but an extravaganza—the connection to the pioneering video artist Paik is unclear from the printout itself.

However, the meaning of the program's joke becomes obvious based on a possibly apocryphal tale Frampton recounted in his 1974 lecture "The Withering Away of the State of the Art," delivered three years before he began programming. The lecture is one of a number by Frampton in which he considers the implications wrought by the eventual demise of film due to the rise of electronic imaging technologies. Frampton ends the lecture by considering a meeting in which Paik exhibited the features of his analog video synthesizer for filmmaker Stan Brakhage. In the anecdote, Frampton places both artists as the paragons of their particular media—Brakhage for film and Paik for video. Frampton muses,

I can imagine Brakhage as he watched Paik elicit from the contraption, at the turn of a wrist, visions of his inner eye that he had labored for twenty years to put on film, feeling tempted by a new luminous apple. "Now," said Brakhage to Paik, "can it make a tree?" I can imagine Paik's ready smile, which seems

Detail from a paper scan of TREE: A D.A.L.EXTRAVAGANZA (for Nam June Paik). Hollis Frampton Collection, Anthology Film Archives.

to come out of innocence, a little slyness, and the pleasure of feeling both ways at once. "Too young," Paik replied. "Still too young."⁸

Though ostensibly talking about Brakhage and Paik, this anecdote neatly sums up how Frampton's interest in video's generative powers, especially in creating abstract imagery, was tempered by the limited realities of the technology. By 1974, Frampton had at least three interactions with video, each with its own particular dissatisfactions. In 1971, while visiting the State University of New York (SUNY) Binghamton, Frampton was urged by two students to shoot something on videotape. He recorded his perambulations around campus onto the already obsolete Sony CV half-inch video format. In his often retold reaction to the experience of the tape's instant playback, essentially his larger response to video technology, Frampton invoked Freud's Civilization and Its Discontents in comparing the cultivating nature of filmmaking's "delayed gratification" and the "barbaric" "immediate gratification" of video.⁹ However, he liked what he shot on video enough to have filmmaker Bill Brand kinescope it to film. After slightly reframing the image, and instructing that the film be shown at 16 fps, Frampton titled the work Travelling Matte (1971).

In March 1972, Frampton experimented with the Paik-Abe analog video synthesizer, which was a processor and generator of video imagery created to Paik's specifications. Frampton created two hours of footage, which he exhibited on November 2, 1974, at Anthology Film Archives as part of its From Film to Video series. Shown alongside both the film and video versions of Travelling Matte, Frampton called the new footage Memoranda for a Dream of Magellan. In mentioning his plan to adapt the unfinished video into the third section of his never completed Magellan film cycle, he recounted the difficulty he had using the Paik-Abe video synthesizer. After complaining that everything was labeled in Japanese, except for one control reading only "red," Frampton opined that the device had "been built very much to the need of the Paik aesthetic, which is not the Frampton aesthetic. As a result I spent at least half of my time outwitting, if one can put it that way, Nam June Paik on his own machine."¹⁰

Frampton had a similarly unsatisfying experience in 1972 or 1973 when he spent time at the WNET studios with Bill Etra, then the New York public television station's video artist in residence. The plan was to offer Frampton free access to use the station's vast array of video processing equipment. As Etra recalls, Frampton was too much of a novice to the world of video, much less the highly technical equipment in a broadcast studio, for the experiment to bear creative fruit. The results of this work are unknown.¹¹

Frampton's TREE, for all its visual sparseness, was his assertion that computer technology actualized the failed potentials of video technology. Paik's video synthesizer couldn't re-create the mimetic powers of photochemical cameras. Though keenly aware of the technical limitations of its present state, Frampton predicted that the computer eventually would. In an act of playful hubris, by placing the DAL's name on the flowerpot, Frampton situated himself at the center of efforts to create the new digital imaging technologies.

Frampton stresses the preceding points in an unpublished lecture from 1982 called "Future of Film Technology." He states,

And we stand, I think, on the very eve of the moment when that difficulty [of using video technology] vanishes. In a moment when it is possible to have the same kind of access to video information absolutely as it is to film information, that is to say, the level of the frame. And to manipulate that information, finally, with the freedom that is not available either in pure video technology or in film at a level that commits you to no physical material whatsoever until you're satisfied that you wish to commit yourself to it. That technology is, of course, that of the digital computer.¹²

In this concept of frame manipulation, Frampton was assuring his audience that the aesthetic concerns of modernist film were safe in the world of digital imagery. Frampton mentions, without citation, that Sergei Eisenstein, the Russian filmmaker who was one of his main influences, was interested in applying his advanced editing techniques to television and video mainly as those media were more widely viewed than film. According to Frampton, Eisenstein "at least tentatively was able to perform a disassociation between montage and the medium with which it happened to have been invented."¹³ So, for all of Frampton's expressed distaste for the barbaric immediacy of video, he was clearly excited by its potential to reach new audiences. Frampton was secure in the fact that he could transplant his artistic concerns from film into the successor of video, the digital computer.

Providing proof for Lev Manovich's observation that quotidian digital "technologies overtook" the aesthetic practices of twentieth-century art, in another interview, Frampton connects the process by which a computer scans a photograph with the extended narrative techniques of modernist authors like James Joyce.¹⁴ After detailing the inadequacies of analyzing a photo at its constituent level due to the lack of a diagnostic tool comparable to grammar for language, Frampton describes the impossibility of a human examining the photo at the scale of its grain. To do so would require millions of tiny, precise mental actions to note the location and grayscale of every grain. However, as Frampton points out, this sort of complex mathematical analysis is essentially the reason the computer was invented. "Well, that turns out to be a very, very handy way to describe images from the point of view of a computer, because computers don't get bored."¹⁵

For Frampton, such an analysis explodes the mimetic indexicality of a photographic image into a highly unrepresentative form mirroring the abstracted narrative of his Magellan cycle and its artistic antecedents such as Joyce. As Frampton notes, "If you did it of course [use a computer to analyze an image], it would have the description of a single object that was probably the size of Finnegans Wake, and there would be in no time at all—it wouldn't even be transpiring in twenty-four hours—it would be interminable, you see. One would, of course, lose—one would lose the picture somewhere in there, yes?"¹⁶ The individual grain contains the larger narrative while abstracting the original object beyond identification. Frampton applies this concept to his own work: "Magellan, includes, even, the portions of Magellan that have been made and released—are perverse, or oblique narratives. I mean it's possible after all to have a narrative account of a very brief period of time for instance—all right—and to expand that brief period of time until it's unrecognizable and too fragmented to rearrange it and so forth."¹⁷

This transformation of the analytical nature of the computer into a highly aestheticized narrative form mirrors Frampton's desire to merge science and art. According to Helene Houston, a student at the DAL and an important member of the software team that created the DAL's sound program DEMON, "we began to work on programs, serious programs that were all aimed at creating an air-tight, intellectually rigorous operating system that artists could use and computer scientists could admire. After all, the goal was nothing short of the marriage or re-marriage, as it were, of Science and Art."¹⁸ In an interview with Scott MacDonald, Frampton discussed the pointless divide between scientists and artists. Artists erroneously imagine scientists as emotionless, while scientists see the arts as shrouded in chaos. According to Frampton, "none of these things is true. In the sciences in particular, and in the queen of the sciences—mathematics— and, indeed, in that celestial, clumsily named intellectual entity, computer science, which has already made mathematics a kind of subset of its own interests, nothing is quite as rampant as aestheticization."¹⁹ Mathematical proofs and computer programs must not merely function, they must function elegantly, thereby moving the analytical sciences into the creative realm of the arts.

Frampton saw two important reasons for the involvement of artists in shaping the direction of computer technology. One was related to the means by which artists could use the computer as a tool for creative expression, like the paintbrush or film splicer but updated to the digital world. The second was related to larger social concerns. With regard to the former, computer scientists designed preexisting digital technology for the demands of their field and not for artistic expression. Until the invention in the mid-1970s of microcomputers such as the Altair 8800 and Processor Technology's Sol-20, non-computer scientists had extremely limited access to computers. Before that time, filmmakers who incorporated computer graphics into their work were reliant on the technical expertise of programmers working with extremely complicated and expensive computers in the labs of large corporations or research universities.

Frampton's attempt at amalgamating the artist and programmer into the same person represented a significant transformation from the work of filmmakers such as John Whitney, who had worked closely with Jack Citron of IBM on films like Permutations (1966), and Stan Vanderbeek, whose series of computer-based films, Poem Field 1-8 (1964-67), were created in collaboration with Ken Knowlton of Bell Labs. According to Frampton, "most existing computer programming languages were designed, from economic necessity, to accommodate the scientific and business communities: the result has been a computing environment uncongenial to users in other fields."²⁰

Frampton was purposefully working to correct that imbalance. The DAL was producing digital tools for artists programmed by artists. The added barrier of the computer engineer between the artist and his creative expression in the digital world was removed. Instead of merely repurposing preexisting computer technologies for artistic works, the DAL tools were wholly designed for that purpose. Though slow in speed and functionality compared to later computer imaging tools, DAL software was inculcated with the high-modernist sensibility championed by Frampton. It is unlikely that such a theoretical concern would drive the design of computer graphics programs coming out around the same time from Silicon Valley.

With regard to his second reason for integrating artists' concerns into digital technology, Frampton expressed a concern for the social impact of computing that went beyond the fact that scientists ignored the needs of artists when designing programs. Perhaps he was being overly alarmist about the implications of the digital transformation of society, but he expressed serious, if often vaguely defined, warnings about the new technology that tempered his more utopian views on the computer. In a 1979 interview, Frampton saw the computer as "now in the process, I think, very clearly of becoming something as ubiquitous, although much more problematical in our culture, as the T.V. set." Earlier in the same interview, he expressed disdain for television by refuting the more idealistic predictions for the medium— that multiple channels and interactivity would in some way negate the essential, unchanging state of viewing television, which was that of an atomizing isolation.²¹

More pernicious effects would result from an unchecked adoption of this hyperrational technology. Frampton, in a conjecture worthy of his intellectual forebear Jorge Luis Borges, imagines a computer program that determines the winner of a chess game based on which player gets to move first. Such a program would obviate any reason to play chess. A vital and creative expression of human thought would be rendered pointless. The organizing metaphor of chess, as an abstract distillation of warfare to a personal game of strategy, becomes moot in the face of the all-knowing predictive logic of the computer. It is here that Frampton reveals the role of the artist in society's digital turn:

When it begins to look as though the structure is completed, and the system is self-justifying and self-explaining entirely, and then implies that you are losing the game, and at that point, of course, you change the rules. Immediately, you see. And that was Mike Snow's remark: "I set up a game and attempt to play it, and if I seem to be losing, I change the rules."²²

For Frampton, the artist represented a humanizing counterbalance to the all-regulating impulse of the computer. To offset computer programs that eliminated spontaneity in the cause of economic domination, the DAL worked to create programs devoted to a willfully generative creativity. Ideally, the artist as programmer would wean the computer scientist from the worst impulses of digital technology's unyielding logic. Computer science would give the artist a grammar to analyze the image, albeit one prone to abstraction.

The Digital Arts Lab's Software and Pedagogy

How exactly Frampton began his programming career is unknown. However, it has a strong commonality with his larger aesthetic practice that derived its theoretical concerns from a practical understanding of a medium. Obvious examples in his artistic career include his employment at a photo lab, his numerous lectures on the future of film technology, and, more metaphorically, his aphorism regarding the surgeon who can't teach students that don't "enjoy cutting flesh and bone."²³ While acquiring the knowledge and credentials to instruct students in the bone-cutting techniques of computer art making, Frampton appears to have taught himself how to program based on intensive readings of instruction manuals on computer languages.²⁴

Though still only beginning to learn how to program, by January 1977, Frampton and fellow SUNY Buffalo CMS professor Woody Vasulka were in the midst of intensive plans for establishing a new curriculum devoted to computer technology and the arts. By that date, they had decided on a number of options that would set the technological and coding course of the DAL, including owning their own microcomputers instead of time-sharing on more powerful technologies housed elsewhere, separating out the frame buffer from the computer to maximize limited memory, and coding in the assembly language to maximize limited speed. At the end of March 1977, Frampton and Vasulka presented their plans for the DAL to the faculty at the CMS, including director Gerald O'Grady and instructors Thom Andersen, James Blue, Tony Conrad, and Paul Sharits. For the 1977-78 school year, Frampton and Vasulka had devised two classes: Seminar in the Image: Theory, and Workshop in the Image: Practice.

The proposed descriptions downplayed the classes' intensive concentration on computers. Digital technology is the fifth medium listed after photography, film, video, and electronic music. Both specifically state in emphasized text that "No prior training in computer science or higher mathematics will be required."²⁵ Ken Rowe, a student in the DAL from 1978 to 1979, reported that the class descriptions gave him the impression that they would be more traditional media arts classes. Rowe, who had little previous computer experience, was quite surprised when the focus was entirely on designing the hardware and software for the DAL's computer system.²⁶ The classes themselves were conducted in a nontraditional manner befitting the CMS's status as an innovative media program. Frampton split his time between Buffalo and his home in Eaton. When he was in Buffalo, class met from the early morning until late at night; when he was in Eaton, the students were free to work on their own projects. Students came into the class with drastically different skill levels in computer programming, from none at all to surpassing Frampton's. Those with little experience in computers contributed less and were not as involved in the actual coding of the programs.²⁷

After more than a year's effort in setting up the computer system, Frampton and the students at the DAL began work on DEMON because digitizing and altering sound was easier to achieve than the substantially more complex manipulation of video.²⁸ DEMON was created over the 1978-79 school year, though no printouts of the programs exist from those dates.²⁹ According to Keith Sanborn and Ken Rowe, who were students at the DAL that year, the program was mainly coded by Frampton and student Helene Houston. Other students contributed smaller portions of code based on their programming skill. This collaborative practice represented a significantly different mode of production for Frampton than the lone experimental filmmaker toiling away by himself.³⁰ The work Frampton created in this phase is therefore closer to that of the auteur working in Hollywood as part of a larger group effort, in that the authorship of the programs is shared among the DAL participants.

Frampton's description of the program conveys a sense of its functionality. DEMON was "an interpretive microcomputer language for audio-frequency data editing and control."³¹ In a flyer for a 1981 live performance of the DAL's capabilities, DEMON is described as "a language capable of editing and modifying audio data and of sound manipulation."³² Frampton goes into more detail on the guiding principles behind the program in a successful 1982 grant proposal to the National Endowment for the Arts, though here calling it DAEMON. The program "facilitates recording and changing of sound information, and implements the operations of a sound recordist and editor using such simple instruments as a variable-speed tape recorder, microphone, razor blade, and tape splicer."³³ Frampton is forced to resort to a metaphor employing analog technology, but the quotation imparts an understanding of how the DAL hoped to manipulate sound. The difference is that with DEMON, operations that would take a tape editor hours of strenuous labor could be achieved in real time thanks to the speed of the computer.³⁴

An examination of the existing printouts in the collection at AFA allows for a more precise understanding of what the program could actually do. From a description of DEMON's command structure, it appears that the program manipulated a sound sample mainly by altering the data in the frame buffer. The commands allowed for an addition or subtraction to the buffer, a repeating of the sample, adding a variety of noise effects, adjustments to the rate of playback, recording a new sample to the buffer from the analog-to-digital converter, and reversing the sample.³⁵ Setting the limits of the buffer edited the sample. Since the buffer could store only one to two seconds of sound sample, DEMON lent itself to a loop-based form, where the sample being repeated was constantly expanded and contracted, sped up and slowed down.

DEMON in operation can be heard on the only confirmed recording made from the software. Taped on May 10, 1979, by student Ken Rowe, the recording documents a successful operation of the DEMON program after a year of coding trial and error. Rowe recalls making the recording almost as an afterthought. As such, the recording documents not finished works but experiments in the capacity and limitations of DEMON. The six tracks run a short audio sample through two different macros, or lists of commands, to manipulate the sound sample. Owing to the jovial party atmosphere that was a by-product of the night's successful operation of the program, the assembled group—Frampton, four students, and a dog—recorded themselves saying a mixture of academic and adolescent phrases: "titillated cataleptic diegesis," "oh, they wear no pants," "shazam, it's tenure mouse," "what the fuck is going on here," "happy birthday," and a growl by the dog.

Scan of code for the program DEMON3.DOC. Hollis Frampton Collection, Anthology Film Archives.

In the recordings, DEMON chops up the samples in the buffer, isolating, repeating, accelerating, and elongating snippets of the sample into abstraction.³⁶ As they are experiments, they should not be seen as aesthetic examples of what Frampton hoped to create with the DAL's tools. Even in such an inchoate form, they exhibit Frampton's goal to use the computer to analyze an image or sound down to its constituent level in a way that drastically expands the original object out to an oblique new form. Additionally, these recordings have a clear connection to the editing strategies in Frampton's 1971 film Critical Mass, where a young couple's quarrel is chopped up and repeated at the level of the phrase, word, or nonverbal utterance. This resemblance is greatest in the film's first two minutes, when the sound track plays out to a black screen highlighting the aural composition and prepares the audience for the reiterative visual techniques to come soon after—not unlike the DAL's designing DEMON before IMAGO, with the technologies of the first informing the second.

DEMON was an acronym for Data Editor and Monitor, which is a surprisingly anodyne name for a program that digitally processed sound. Though this is a conjecture, it seems likely that the larger meaning of the term was retrofitted to the acronym DEMON. Though no evidence exists to prove it, it seems highly probable that Frampton was aware of the computer science field's use of the term demon or daemon to signify a program that performs a function in the background automatically based on its initial instructions, without requiring the constant inputting of commands. This connotation of the term seems to have originated with the programming team at MIT in the early 1960s. Apparently the MIT programmers appropriated the term demon from nineteenth-century physicist James Maxwell's thought experiment in which, explaining the second law of thermodynamics, he imagined a demon shuffling atoms back and forth to resist entropy.

Frampton was quite aware of Maxwell and his fabulous molecule-herding spirit. In 1968, Frampton released a four-minute film, Maxwell's Demon, that interspersed footage from a Canadian Air Force training film with one-second blasts that cycled through the six additive and subtractive colors. According to an interview with Scott MacDonald from the late 1970s, Frampton composed the film as a "six-bit binary number." Apparently, there were to be sixty-four different variations of the one-second color sections, though Frampton admitted that he erred in production and only shot sixty-three.³⁷ In the interview, Frampton—who by that time had begun work as a computer programmer—may have been retroactively applying his current project of applying digital sciences to the art of image making to a film he made ten years earlier.

Presumably, one of Frampton's goals with programs like DEMON and IMAGO, though only hinted at in his revisionist application of an algorithmic compositional style to his earlier films, was to industrialize the process of editing and arranging complexly iterative serial-based works such as Maxwell's Demon and Critical Mass. Despite his reluctance to accept the term structural when it was used to classify his films, there is clearly a common theoretical ground between the more analytical and recursive form of structural filmmaking and the basic operating state of the computer. How Frampton would have actually used the computer to create art was unfortunately never realized. That the organizing algorithm behind Maxwell's Demon was not fully completed points to Frampton's willful subversion of a dictating system in the digital realm.

The origin behind the name of the DAL's image processing program, IMAGO, is not as immediately connected to Frampton's larger visual practice as is the name DEMON. Frampton's adoption of the term could be explained simply by the fact that it is the Latin word for image. Considering that Frampton spoke often of an image-machine that would counter the machine of language, in a very real sense, IMAGO, as a computer program that generated and processed graphics and video signals, is a literal realization of that concept.³⁸ In IMAGO, the image on the computer monitor could be manipulated by the precise control of language. For Frampton, the digital not only bridged the gulf between the arts and sciences but represented a dialectical merging of the word and the image in a way unachievable in the photochemical arts.

Regardless of the denotation of its name, IMAGO was built on the DAL's work on DEMON. Frampton wrote to Bill Etra on July 31, 1979, "This [the letter] isn't longer, or funnier, because I'm up to my ass in the compiler [IMAGO]. . . . It's for sound now [DEMON], but will do color-frame grammars for Paul Sharits-like image structures in September."³⁹ Frampton was clearly moving immediately from their success with DEMON into what became IMAGO. This explains the lack of artworks created with DEMON, as the team quickly turned its attention to adapting the program into one that could manipulate images, instead of exploring the artistic possibilities of DEMON. Furthermore, there is the brief but tantalizing description of what the image program was being designed for: "color-frame grammars for Paul Sharits-like image structures." Whether the program was actually being created to emulate the color flicker films of his colleague in the CMS or whether it was a useful shorthand descriptor is unknown. However, it provides a conceptual baseline on which to place the intentions behind IMAGO.

By 1981, work had progressed enough that Frampton and his students were able to define the "preliminary documentation for IMAGO, a video raster graphics language configured as an interpreter."⁴⁰ In this initial conception of IMAGO, only utilitarian issues are addressed. There is no mention of what imagery the program would create. At this early stage, Frampton and the DAL were still concerned with the fundamentals of digital images: delineating the size of the screen, notating how the software writes an image to the screen, setting the pixel pit depth, and designing the available hue and grayscale. It cannot be emphasized enough that the DAL had to create the parameters of the on-screen pixel resolution and program the computer to recognize that area. The generation and processing of computer graphics and video images would have to come after the visual environment was created.

Design of the IMAGO system. Hollis Frampton Collection, Anthology Film Archives.

As this documentation implies, IMAGO should be thought of as an entire system that included the original DAL software, a variety of inputs to generate and manipulate the imagery, a computer operating on a specific commercially available chip, a custom-built frame buffer, and a cathode ray tube monitor. The interaction of the hardware is detailed in a May 1, 1981, document that includes a three-page schema illustrating the connections between the inputs, computer processor, frame buffer, and monitor.⁴¹ There are five inputs on the signal path, beep, relay, joystick, light pen, and random number generator, with the obvious input device of the keyboard missing from the blueprint. From the computer, the signal path led into analog-to-digital converters and the four frame buffers.

These components were all combined on a single board, dubbed the GOLEM. The DAL's frame buffer was designed by Robert Coggeshall, the DAL's technical manager and Frampton's recruited expert in hardware. Work on its construction was begun in June 1981 and completed sometime in summer 1982.⁴² Unfortunately, the frame buffer is no longer extant. Numerous attempts to track it down at SUNY Buffalo over the years by Coggeshall, Woody Vasulka, and Frampton scholar Bruce Jenkins have turned up nothing. The lack of the DAL's frame buffer greatly complicates any future attempts to preserve IMAGO. The buffer's functions could be replicated with current software, but its precise and idiosyncratic interaction with the IMAGO software might present difficulties that would complicate any attempts at emulating the behavior of the original.

As with DEMON, no known artworks created with IMAGO are known to exist. Even less documentation of IMAGO being tested exists than for DEMON, offering an incomplete assessment of the performance of the program. Only one tantalizing fragment is known to exist. In his talk "Hollis Frampton's Algorithmic Aesthetic," Keith Sanborn played a brief snippet of video processed through IMAGO. The black-and-white image shows a spinning ring on a tabletop. Though the sound seems unaffected, the video appears to jump around and become stuck and then unstuck on certain frames, freezing the image in staccato bursts.⁴³ As with the existing recording of DEMON, drawing a comparison between the scant output of IMAGO and Frampton's completed works in film and video is based on an extremely limited data set. However, the repetition and transformation of isolated moments of time in the only extant video documentation from IMAGO evoke similar strategies found in other works by Frampton such as Artificial Light (1969), Zorn's Lemma (1970), and Critical Mass (1971). In discussing the latter film, Bill Simon describes this technique as a "stutter" that visualizes a nonlinear conception of temporality.⁴⁴ This continuity of artistic practice from the analog to the digital realm is indicative of Frampton's confidence in the computer's ability to support the aesthetic aims of experimental film and other high modernist art forms.

There exists no visual documentation of the generative computer graphic capabilities of IMAGO. However, an examination of IMAGO's 171 commands gives a sense of what the program was designed to do, even if these commands don't convey specific visual qualities. IMAGO could write pixels to specific locations on the screen, invert the image along an axis, designate where the raster display starts in the image, draw a line between two points, redefine the display to a smaller size than the normal 256 × 256 pixel area, alternate between a flat or round (objects leaving one side appear on the other) display, set the color of the background, transfer images off of a disk into a buffer, randomly create curves and polygons, precisely determine the size and location of a geometric shape, draw vector lines with a joystick, rotate shapes around a center point, and more. A number of commands synthesized two different images from separate buffers. Others manipulated pixels directly by changing their color, increasing and decreasing their contrast, and expanding each pixel to a larger size.

How Frampton would have employed these generative graphics can only be

Scan of code printout for the program IMAGO.DOC.Hollis Frampton Collection, Anthology Film Archives.

surmised based on tantalizing clues. DAL student Keith Sanborn recalls that Frampton expressed an interest in using the computer to perform live visual improvisations to break him out of the traditional mold of the filmmaker exhibiting premade works.⁴⁵ In a 1978 video by Steina and Woody Vasulka, Funtime at the Vasulkas, Frampton himself states, "I am waiting for the first live digital performance. The first digital quartet. I'd like to do it. I'm quite serious. I think it could be spectacular. Just spectacular."

A different form of integrating computer graphics into Frampton's visual production is hinted at in a stack of papers with twelve by twelve graphs that he called "quarter matrices." In a nod to John Cage's aleatory compositions, Frampton filled in the graphs with dots based on rolls of the dice. Evidently, these matrices were to form the basis for imagery to be used in his unrealized Magellan film cycle.⁴⁶ Though nothing more was done with these graphs, they were to be somehow transformed into computer graphics, according to Frampton's notes found on the papers: "Separate lines into 6-bit binary numbers and translate to decimals (0-63)"; "Is there a program to find and mark redundancies in quarter [matrices]?"⁴⁷ Though no such programming is known to have occurred, Frampton here reveals his mode of digital production that would mix an element of chance into the analytical nature of the computer.

Conclusion

Turning these ruins of Frampton's seven-year experiment with computer programming into completed works is fraught with problems. First, Frampton's death in 1984 leaves us with no direct guidance as to his original intent. While the questioning of individuals involved in the creation of an artwork can be crucial even in the most standardized feature film, it is even more important in avant-garde film or with truly experimental new technologies such as IMAGO and DEMON. This fact is compounded by the dearth of documentation of DEMON and IMAGO in operation and the absence of finished artworks created on the systems. As such, any attempt at preserving these works, which, in relation to computer programs, entails some form of functionality, is sorely lacking in any points of reference from which to judge whether the new versions are authentic representations.

Second, there is the issue of obsolescence and the one-of-a-kind nature of the DAL technology. Though the floppy disks are in remarkably hardy shape, the programs would not be functional even if the original mass-produced computers and disk drive were easily available. This is due to the loss of unique hardware such as the GOLEM frame buffer on which IMAGO and DEMON were specifically designed to operate. Unfortunately, the less transformative methods of new media and software preservation, such as utilizing original technologies or emulating the operating system in which the software was created, do not appear to be viable options.⁴⁸ The remaining avenue for preservation is reinterpreting the programs with entirely new computer technology, which runs the risk of significantly altering the original intent.

Third, there is the fact that how we interact with computers and use digital technology to generate imagery and sound has changed drastically in the last thirty years. Before embarking on any preservation of these works, the question of the value of re-creating programs that would seem today limited, slow, and difficult to learn must be asked. Currently there are a large number of easily available software tools that allow a user to create art on the computer. What would be the use for today's artists in re-creating a program that works in a 256 × 256 pixel area and can color in a circle with only twelve colors and, at the limits of its capabilities, flip a frame of video?

Despite these challenges, there is a reason to attempt some form of preserving these programs, however transformative such an act would be. It would not be for the purpose originally intended by Frampton and his students—creating freely available artists' tools that were both a time-saving device and a new venue to continue Frampton's aesthetic impulse into the digital realm. Instead, the goal would be to historicize the world of new media artworks by focusing on this transitional period when filmmakers first began to learn programming or, more precisely, were forced to program if they wanted to use the computer to make art. Ideally, such a project would take place in a pedagogical environment like the DAL and bring together students from media arts production, preservation, and computer programming. Perhaps DEMON's and IMAGO's restricted capabilities as compared to current digital imaging technologies could show artists the role hardware plays in setting the horizons of aesthetic expression, a fact that today might seem less conspicuous in light of our increasingly powerful computers. Examining the DAL software may offer computer programmers a case study in approaching coding from the mind-set of the artist. Being forced to deal with the gap between the original and reinterpretation could reveal the highly interpretative nature of all preservation efforts to conservation or archiving students.

Hollis Frampton and his students at the DAL did not just create programs of artists' tools. In constructing the entire computing environment—software with its unique command language, one-of-a-kind hardware designed to maximize the deficiencies of commercial systems, delimiting the on-screen space, and so on—they created an idiosyncratic blueprint for interacting with the digital world. Their artist-run model was determined by anarchic contingency, analytical abstraction, and the aestheticization of rationality. That their computing system has been reduced from their epic attempt to harness the artistic promise of digital technology to the ruins of obsolete eight-inch floppy disks only emphasizes its unrealized possibilities. Preserving Frampton's software is not about restoring programs designed to help artists make computer art. The goal is the preservation of a lost potential for shaping our digital environment and investigating an early instance of the transformations that filmic practices underwent in the switch from analog to computer.