Week 3: Digital Forensics, Materiality, Fixity & is-ness
The inner workings of a computer have always been a bit of a mystery to me. I grew up when most people didn’t have a computer in the house, let alone several, and the Internet had never been heard of. While in school, I actually remember taking a computer class in which we were taught how Apple’s Graphical User Interface worked. Over the years, thankfully, I progressed beyond operating the mouse, and learned how to use a variety of software applications. But the inner workings of the computer, were still a bit vague – just visions of bits adding up to bytes and kilobytes and megabytes, and so on.
Kirschenbaum, in his book Mechanisms: New Media and the Forensic Imagination, recounts a story from his youth when he stopped saving files to a 5 ¼-inch disk and began saving them to the computer itself. The storage was hidden away, behind the hard plastic case of the computer. He explains that architecturally the personal computer didn’t really change, but the “psychological impact” of saving information to the computer, instead of a removable floppy disk, cannot be ignored. He explains that no longer labeling a disk or having one to carry home after class just felt different. The physical aspect of the digital work was taken out of his hands and was sitting concealed inside the computer. In other words, what happens in the storage device stays in the storage device – and if you’re like me, the details of it all weren’t something I necessarily needed to know.
According to Kirschenbaum, the storage mechanism of a computer is both a product and a process, and despite the work being created behind the scenes and hidden away as 1s and 0s, it is indeed physical in nature. It has a materiality to it. He goes on to describe in great detail the history of computing, disk arrays, tracks, sectors, read/write heads and hashing.
All of these hidden storage and transmission actions point to a very structured process that must consist of rules and standards in order for it all to work. However, Kirschenbaum refers to a forensic materiality, which “ . . . rests upon the principle of individualization, the idea that no two things in the physical world are ever exactly alike” (p. 10). He explains further that a computer is able to create an illusion of immaterial behavior, in which data can be identified without ambiguity, sent without anything being lost, and copied with no variance occurring. This illusion hides the fact that errors can occur anywhere along the line.
These errors, whether it is a function of copying over old data, as in the case of the Jonathan Larson collection described by Reside, or intentional tampering, which occurred with the Mystery_House.dsk game in Kirschenbaum’s book, could pass us by, completely unnoticed. But, through the use of a hex editor, these hidden artifacts come to light and provide us additional forensic evidence and new insights. Reside’s article for instance points out the ability to see Larson’s creative process after the hex editor finds deleted text.
These pieces of forensic evidence that get tucked away should make us question what in fact we might be copying – it’s not always what it seems. So, as a digital archivist, you have to ask, what versions do you keep? Or do you save all of them? Which version is the “authentic” or “authoritative” one? Or is that an impossible choice to make as a digital archivist?
I think the point about the illusion of immateriality is especially important when answering – or attempting to – about making the “impossible” choice. Knowing that it is an illusion, that those bits have a physical form and that with a microscope powerful enough, they can be seen; with the proper software and education, they can be read and understood. The question of authority in art has more layers than I know what to do with; it’s not just drafts, but, as you mention versions. Art likes series and multiples and I get stuck in the same questions. Drafts are easy; versions and series less so – but that drifts back into artistic intent, doesn’t it? If the Work is meant to be a certain way, is it our responsibility to know that going in? Or is it the artist’s responsibility to make that known to us? Isn’t that the point of the artist’s statement? But if there is a single item that was displayed out of a group of similar objects, or one that was sold, wouldn’t that be the authoritative version? Or, if not, under what circumstances?
At the end of the day, when the budget is getting tight, I would answer the question in two ways: why are we saving this? And why did the creator? If the answers to these two questions aren’t obvious to trained experts, then that takes a lot of the impossibility out of the choice because despite Moore’s law, storage space is still finite and expensive, it might seem infinite and relatively inexpensive, but neither of those things is true. So we make choices, just like with any other special collection.
I think one other question that might be added to yours is this: how can we preserve the work today so that it will be of the most use to future researchers? By answering your first two questions we can determine exactly what needs to be saved, but by answering this third question we can determine the best way of how to do it.
I think Brittany’s added question provides a nice balance and is perhaps most practical way of looking at the overall situation. The things that we may value in a given artwork will change with time and we can never really know the artist’s intentions because we can’t read minds. This is a constant issue in archiving both analog and digital materials; for instance, you write the front matter of a finding aid in as neutral a voice as possible, but you also may throw things away or impose an order if the materials arrive in a jumbled state (though you will note this in the finding aid, of course). It seems that the most honest and neutral actions we can take are to develop things like the NDSA levels while also seeking ways to incorporate more voices (crowd-sourced projects, folksonomies, etc.). The utopian Open Museum may be a long way off, so it is essential for institutions of various sizes (especially those in the underfunded long tail of archival institutions) to understand how best to preserve for future researchers.
Yes, I think all of your questions are valid. As you point out Catherine, storage space might seem infinite and inexpensive, but in the world of day-to-day business operations, it really isn’t. So, you have to balance the reality of your work environment, answer those standard appraisal questions, and make hard choices. And occasionally, yes, it might end up being wrong, but that’s where documenting why a certain choice was made becomes so important.
That’s an excellent point, Kerry.
Honestly, that’s not all that far off from traditional archiving practice. You CAN’T save everything. That’s even true in larger institutions, but especially in smaller ones- some materials will take priority over others and receive more care. Archivists make (and have made) educated decisions about what to prioritize all the time. Mistakes are made in traditional archiving to that effect just as they are in digital archiving. The point is that we do the best we can and keep moving forward.
Agreed… and in addition to the practical issue of cost and storage space, there’s also the question of whether we would want to preserve everything even if we could. The process of selecting materials is arguably important in making the materials you do keep accessible and usable. (If we simply kept everything, would our users be able to usefully access and locate the materials they need from our collections?) While (as Kirschenbaum notes) Google argues that users should simply archive every email and use their search function to find things later, in practice I still think it’s useful to delete insignificant or spammy emails, so my later search results won’t be overwhelmed with mailing list cruft. Likewise, if we want our archives to be accessible and useful to our users, there’s arguably value in making an appraisal and selection regardless of storage capacity.
While I appreciate the Kirschenbaum used hex editor software as an illustrative example of detecting fragments of other works, and I think it does a good job of showing how recoverable digital elements can be, I was disappointed that it never got directly compared to analog examples. Yes, Kirschenbaum references the concept of palimpsets from time to time, but while there is only one Mona Lisa, we can only see the final layers of pigment (and the wear on them over time). We can’t see all thirty layers. Vagaries of code let us see earlier drafts of RENT in the same way that spectral imaging can let us see items painted over in classical works due to the mineral content of various paints used. The finalized text of Frankenstein may not change, but any serious student of the work would be extremely interested in earlier drafts. Archives that deal with late authors’ works often have to deal with the question of ‘destroy an unpublished draft, which is what the author would have done, or accession it?’ I think it’s less a question of which ‘versions’ are ‘authentic,’ and more about ‘which ones provide new information.’
Great discussion of some of the points on the underlaying material and forensic mature of digital media. This is going to be one of the foundational points that we can keep coming back to over the course of the semester. Despite how complex issues might become in terms of pinning down what is significant and making sure that we can provide access to it in the future there, when it comes to digital objects, there is always some set of markings inscribed on a physical medium that we can look at making exact copies of. It is a nice thing to be able to anchor to as one starts to make decisions about what to collect and preserve.
Yes, and the physical medium can provide so many visual cues that we take for granted. There was an interesting point made in the Kirschenbaum book about the Michael Joyce files that went to the Harry Ransom Center, in which some of the documents were nearly impossible to decipher whether Joyce had written them. He did not have his name on certain files and there was no indication of who may have written them. Staff had to have several archivists looking at these, multiple times, to decide whether they were written in Joyce’s style, and even to figure out what the files were. Digital objects lack those physical aspects, like handwriting, letterhead, ink color, etc., that could provide that needed context. So, an archivist might discover this bonus forensic information buried in a digital object, but the flip side of that is lacking even the most basic identifying clues.