07.08.05
Jessica Helfand | Essays

New Models for Design Efficiency: Introducing Otto


Palimpsest, 2005. Designers: Elisabeth Prescott and Otto.

As a metaphor for the depersonalization of industry, the 1957 film Desk Set pitted humanity against automation as the central quandary of modern civilization. The film starred Spencer Tracy as an efficiency expert who introduces a mega-computer to modernize a corporate reference library. In the role of the head librarian, Katharine Hepburn's suspicions are confirmed when the computer starts spitting out pink slips along with its overly literal, and frequently inaccurate answers.

Based on the broadway play by William Merchant, Desk Set dramatized the notion of computer efficiency as a harbinger of dystopian doom. The computer, dubbed the "Emerac" (clearly modeled on the Eniac) becomes a kind of silent fall guy, big and impersonal and impervious to the subtleties of life, love and the Dewey Decimal system.

Of course, nearly half a century later, none of us really believe computers are running the show.

Unless the computer's called Otto.



Charles Addams ©1961 The New Yorker

Otto, short for automatic, is the result of a collaboration between a small group of editors, programmers and graphic designers (all of them graduate students at Yale) who designed this year's issue of Palimpsest, the School of Art's literary magazine, essentially by proxy. In spite of its occasional gestures of stylistic clumsiness, this experimental publication suggests fascinating new opportunities for editorial collaboration, and engages the role of computation in an entirely new way.

In an effort to create a system that would, in turn, "design" the magazine, Otto began by assigning a three-digit numerical tag to each submission according to the order in which the editors solicited the work. This became the basis for a strict file structure Otto required for analysis. (All of Otto's analysis was built with Processing, a Java-based programming language developed by Ben Fry and Casey Reas at the MIT Media Lab.) The editors then issued a questionnaire to their contributors, narrowing the submissions using a traditional editorial process (work was either "in" or "out," ) thereby reducing Palimpsest to its material core.

Contributors were then asked to submit descriptions of their work through answers to a series of question sets. For the first eleven multiple-choice questions, icons were paired with each answer. Treatment of the icons, however, was determined by the answer to the twelfth question, which generated a corresponding connection to one of four templates: "clean and pristine," "full and colorful," "illustrated," or "lots of patterns."

Otto then took a pixel from every image and averaged the values of red, green, and blue, dividing each sum by the number of pixels to find its essential color and then converting that number to CMYK values. (The saturation of each of the four inks determined where the submission was placed in the book.) Following its press run, the magazine's signatures were hand-shuffled at the bindery according to a set of rules that preserved the order of the text (as pre-determined by the editors), but randomized the order of the art. Each book, as a result, presents the same material in a different sequence.

In 1961, a Charles Addams cover for The New Yorker showed a small woman sitting in tiny room filled with an Eniac-sized computer: in the dim light, it spat out a tiny white page with a red heart on it. The cover was for Valentine's Day (the date the Eniac was actually introduced in 1946.) Desk Set provided yet another goofy storyline that illuminated the fallibility of the machine along with the triumph of human understanding. But Otto's not so much funny as it is fascinating, and indeed, bizarre: it proves, quite simply, that a designer can cede control and still retain a kind of basic curatorial hold over a piece of work. For now, Otto's maneuvers are mostly operational and service oriented — it's essentially a hierarchical system, much like the Dewey Decimal system — but it is not so difficult to imagine a different level of interpretation taking place. How long until Otto makes its own decisions, develops its own arsenal of aesthetic or, far-fetched as it may seem, conceptual notions?

For now, Otto's pretty literal — not unlike the Emerac, actually — and that's probably a good thing. But at the end of the day, as Otto's designer points out, it's all a question of interpretation. Otto's capacity to interpret content is guided by editorial conceits that frame its capacity for expression: for now, icons guide placement and answers to specific queries drive template choices. But could Otto do more? Could it match census data to Flickr photo pools to generate real-time visual maps? Could it embed a translation engine that enables multi-lingual global communication and bypasses cultural stalemates between nations? Design, in the service of the real world, would do well to consider such critical matters. And if Otto can help, so much the better for us all.

Posted in: Arts + Culture, Graphic Design, Technology



Comments [5]

The seeds for the process that created Palimpsest predate the MIT Media Lab and even Desk Set. Indeed, much of what we would classify as modernism in graphic design practice has to do the imposition of different kinds of impersonal systems upon decisions that designers used to make intuitively. Technology has simply accelerated a process that has been long underway.

All grid-based composition systems, for instance, from Joseph Muller-Brockmann on down (or perhaps even back to, okay, Guttenburg) are motivated by the belief that a set of rules, rather than the designer's eye, can determine the layout of a page. Palimpsest takes that premise and pushes it to new limits, raising the question of what lays beyond.

(Some larger views of more spreads would be great, by the way!)
Michael Bierut
07.10.05
12:44

To get the full effect, all the pages in the book are scrollable here (at a larger scale.)

Jessica Helfand
07.10.05
05:00

I'm glad to see that Otto goes in for the same kind of gratuitous decor that some of us humans do. And in a style that nicely reflects Otto's computer heritage and world view.

A computer that can reliably produce layouts brings us much closer to the great day when we can order magazines from an a la carte menu of articles and illustrations. Everyone can become their own personal editor.
Trent Williams
07.10.05
07:13

Interesting as an experiment, though some of the steps seemed a bit fussy. The randomizing in particular seems extraneous.

To continue Michael's prior art tack, John Maeda attempted to write and then abandoned some automated layout prototyping software for Maeda @ Media over the course of seven months. In his case, it was a concern over time constraints, but I wonder if the results would've been acceptable. His mention of it at the end of the book doesn't actually show any output, that I can tell.

Somewhat related, a few years ago, a preview of the n-Gen Design Machine was released. I can't find the location at the moment, but I remember sharing this, with the reception—as opposed to the reaction, mind—being best described as "cool." Most were irritated by the overall concept, others even less amused by the implication of the modules. California Noir was "inspired by" David Carson, for example. Unacknowledged by the n-Generate site in its current incarnation (Wayback Apr2004), the program was created by Move Design (Flash required), where the curious will find a download link. It's a Shockwave application, so no installation required.
Su
07.10.05
08:25

Hi Jessica,

I was actually wondering about the image you posted at the top of this thread right below the "New Models..." title. It looks like a very interesting architectural design website. Is it a website and if so, what is it's URL?

Thanks
mg
07.25.05
10:57


Jobs | October 15