Preservation as Software Studies

My "pecha kucha" presentation at Softwhere 2008.

Steel

Last night I got to hear film director James Cameron speak about his exploration of the wreck of the infamous German battleship DKM Bismarck at the American Society of Naval Engineers (a venue that doesn't draw many English professors). I mention this here because Cameron's talk introduced me to maritime forensics, a fascinating field at the intersection of history, high-tech, and underwater archeology. In the course of his 45-minute talk, Cameron painstakingly detailed the difficulty of determining what actually sank the Bismarck, the shellfire and torpedoes of the British fleet or, as survivors have long claimed, the crew's own scuttling charges. The key witness is "the steel," the entity Cameron repeatedly invoked as a kind of synecdoche for the mute material evidence of the wreck.

What was most striking about the talk was how the brutal chaos and violence of a modern naval engagement could be reduced to the engineering teams' clinical reconstructions of the trajectories of specific shells, fired from a particular battery at a particular moment in the fighting. (Such feats are not uncommon in naval history, however: John Campbell's Jutland: An Analysis of the Fighting offers similar data on hundreds of shell strikes during the largest dreadnought battle of the First World War.) To obtain an accurate picture of the battle damage, however, Cameron's team first had to "subtract" (his word) the damage incurred when the ship plunged three miles to the ocean floor, impacted, and then slid down the side of an undersea mountain. The maritime forensics specialist must therefore become adept at differentiating and interpreting cryptic marks and shapes in the ship's steel, all that is left to testify to unimaginable stresses and trauma.

Bismarck's underwater remains are a grim reminder of that key dictum of forensic science that is also central to Mechanisms: every contact leaves a trace. It is also, however, a reminder of the intimate and essential connection between forensics and humanism, for while there were few things on earth as inhuman as the warship and the regime it served, Cameron compellingly demonstrated the eloquence of the signs we read in those twisted shards of steel when they intersect with survivor's memories to tell the stories that make up history.

Leonardo Review

Jan Baetens (poet and Professor at the Institute for Cultural Studies of the University of Leuven) reviews Mechanisms for Leonardo On-Line, calling the book "a watershed":

In the reflection on new media, this book is undoubtedly a watershed publication. Its basic stance is that electronic writing can only be understood if we accept to consider it a real form of writing, i.e. of material inscriptions on material surfaces, and therefore to leave behind many of the myths that surround digital culture.

What makes Kirschenbaum’s work so thrilling and innovative is, however, not only the demonstration that electronic writing is also a way of writing, even if the computer is a machine meant to withdraw its own material operations from our attention [. . .] At least as important is the humanist viewpoint defended by the author [. . .] In his book, Kirschenbaum uses forensics as a tool to think of electronic writing as a chain of contacts which are never materially lost, while at the same time insisting on the fact that it is much more than just a sequencing of inscriptions on a hard disk (of on other types of surfaces, although the hard disk has now become the dominating form).

Mechanisms, which opens totally new grounds for electronic textual scholarship, will be one of the books that can redefine what it means to be a digerate.

Wow. It's an amazingly generous review, but also very detailed. I'm floored.

Titular

I knew from the moment that I began working on the project that I wanted the book to be called Mechanisms. I was terrified, though, that someone else in the field would use the title first. It seemed like such an obvious choice--"mechanism" is a word that crops up in our conversations all the time. For me it was perfect, capturing the mechanistic side of computation that I was keen to foreground, as well as connoting both product and process, artifact and algorithm.

The subtitle was much more difficult to nail. Initially it was something like "New Media and the Forensics of Digital Inscription." Ugh. One of the original reviewers for the project gently suggested I might like to reconsider. I did like the focus on inscription (as opposed to text or textuality), but let's face it, that's not a construction calculated to grab even the scholarly impulse buyer. I subsequently played with many variations, trying to retain forensics, textuality or inscription, and new media as key concepts. For a long while, the book was subtitled "New Media and Forensic Textuality," and it appears as such in some early citations. Textuality was important to me because the book is so deeply grounded in humanistic traditions of textual scholarship, but in the end I just couldn't make it work. "New Media and Forensic Textuality" was a phrase which smothered; "New Media and the Forensic Imagination" struck me as dynamic and evocative. The book's coda is also called "The Forensic Imagination," so there's some nice structural reinforcement there. In the end I'm quite pleased with how I titled the book, and I should also mention that the MIT Press gave me complete freedom to do so, for which I am very grateful (I know many colleagues publishing elsewhere who have had glib or awkward titles foisted on them by marketing departments).

A Hard Driving Review

Nick Montfort posts a hard driving review of Mechanisms over on Grand Text Auto.

Kirschenbaum’s forensic approach to information storage technologies shows us qualities of the machine that have seldom, if ever, been remarked upon in new media studies.

This is not your father’s book about hard drives. It’s also not the type of book we had in digital media studies five or ten years ago.

Mechanisms is about more than hard drives of course. But my biggest concern was finding a way to discuss new media with the same degree of material specificity we've been accustomed to seeing in other fields, and I'm thrilled that as perceptive and knowledgeable a reader as Nick has found that in the book.

Every Contact Leaves a Trace

This phrase, adapted from pioneering forensic investigator Edmond Locard, is not only one of my chapter titles in Mechanisms, but also one of the central dictums of the book, for I claim this holds true as much (or more) in digital environments as the physical world. So I was intrigued to read an account of software developed to detect digital image tampering in the current issue of the Chronicle of Higher Education (June 6, 2008):

Mr. Farid, of Dartmouth, has developed software tools that can automatically check for image tampering. The software looks for patterns in the digital code underlying an image. When files are opened and altered in Photoshop, for instance, codes are added that Mr. Farid's software can detect. Likewise, when scientists copy and paste parts of images in software programs, their actions leave a digital mark. (A10)

In the book, I described the kind of "marks" Farid's software detects as an artifact of what I termed formal materiality. I wrote:

An image file is typically thought of as consisting of nothing but information about the image itself—the composition of its pixilated bitmap, essentially. In fact, however, the image can carry metadata (documentation as to how it was created, embedded as plain text in the “header” of the file), as well as more colorful freight, such as a steganographic image or a digital watermark. This content will only become visible when the data object is subjected to the appropriate formal processes, which is to say when the appropriate software environment is invoked—anything from the “Show Header” function of an off-the-shelf image viewer to a 128-bit encryption key. (12-13)

The kind of software the described in the Chronicle's article contributes to our understanding of digital objects as mechanisms, that is as artifacts with a recoverable past--as opposed to black boxes or inscrutable blobs of code.

June Open Thread

Comments, questions, rants, raves. You know the drill.