I had the pleasure of speaking with a class taught by Jean-François Blanchette last week. It was such a good experience that I’m getting two posts out of it, in fact. In this one, I want to give some scope to an idea that I mentioned to them in passing. It’s become a standard part of my fundamentals of digital preservation teaching and I think it’s time to give it some air here on the web.
It is a commonplace to say that we can’t think of digital preservation in the same way we do paper preservation, but I contend that the opposite may be true. I find it useful to think about all preservation efforts within a shared theoretical framework and then try to identify the specific technical knowledge required to make that framework sit up and stay forever.
Lately, I’ve been talking about preservation as consisting in the management of four factors. The first is maintaining a recording substrate. In the case of nitrocellulose film, this is hard because the substrate tends towards combustion. In the case of floppy disks, the substrate is Mylar, a permanent-durable polyester that eventually replaced nitrocellulose because Mylar is chemically inert (read: doesn’t catch on fire) and resistant to physical damage.
The second factor is the recording medium. In the case of a black and white film, this is silver salts suspended in gelatin, and once it is properly processed, gelatin silver film is remarkable stable. In fact, it is the recoding media for the microfilm that many consider the benchmark in preservation copying. Floppy disks, by contrast, are coated with a few micrometers of magnetic metals, usually iron oxides. That metal layer is physically fragile and its magnetic signal will degrade over time.
The printed edition of the newspaper is a familiar example of both of these factors performing poorly. The acidic, ligneous paper turns brittle as hydrolysis shortens the cellulose molecules in the substrate, and the pigments in the soy ink smudge and disperse easily.
If one has dealt with the problems of maintaining a substrate and media or — and this is often the case with the media used in digital recordings — if those problems can’t be mitigated, preservation must act in at least one of two other arenas. The first is the transport (a word I don’t really like for this, but keep using), the means by which recorded information is rendered to its recipient. For the newspaper, this is simple enough: a little bit of light and an eyeball are all that is required. For the nitrocellulose film there’s quite a lot more light in the projector, a lens to make things eye-legible from a distance, and a flat, reflective surface before the eyeballs get involved.
Still, film or paper, this is essentially the same process, just amplified or magnified. That’s a part of the reasons that it’s fair to say that analog media are good preservation bets. There’s very little complexity in the transport. On the other hand, mold, rats, fire, water, acid decay, storage space, and the list goes on.
The floppy disk requires a much more complex transport, with a disk drive (a motor, spindle, magnetic head, logic board, and wiring) and a computer (CPU, circuit boards, RAM, etc.), and a monitor, printer, network connection, and/or various other peripheral ports.
Digital recordings are paired with complex transports, to be sure, but I’m not willing to say that this problem is unique to the last few decades. The transport for a Bach Cantata is quite complex, too, but we have been building organs in cathedral for centuries.
Finally, preservation depends on a language, a system for interpretation. In digital preservation, there is an interesting level of recursion in this factor, as data is interpreted by software that is interpreted on software that is interpreted by, one suspects, an elephant that is interpreted by a tortoise. But, a few prodigies aside, most interpreters of Bach have learned to read music. Language obsolescence can affect analog collections, too.
I often talk about Linear B and the Rosetta Stone to make this connection. Translation is a less substantial preservation problem than decipherment. In the world of digital recording, the term for moving from one language to another is migration, and it is far less fraught than translating or deciphering ancient languages. The core formats that libraries and archives employ for digital preservation of text, sound, images, and structured data have been in use and stable for decades and their contents are so well-specified that a migration from current to future formats can be carried out without any loss of data.
There are formats that are not so stable, though. Video is a prominent example and many industry specific applications rely on proprietary combinations of software and file format. This is why precision in thinking is important. If we are to effectively care for these materials, we must be clear on where the problem lies. One remedy in these instances is the development of a common language, which either facilitates migration or obviates the need for it altogether. The other remedy is emulation, which is the creation of a language for modelling transports. ( I am pretty sure I wrote that correctly, said Alice to the Rabbit.)
Emulation is effective, even essential, for projects like video games, where the interaction with the original transport is essential to the experience. It’s also at work in lots of analog environments, but often unremarked. Going back to Bach, if you’ve ever encountered the historic performance movements in classical music or theater, you have a good analog analog to digital emulation.
All of this brings us around to the idea of curation. Here again, the distinctions between artifactual and digital preservation are the specifics of who is involved. Preservation always occurs in dialog with the users of the information. Pick your medium – Armenian devotional scrolls, literary texts, motion pictures, datasets, videogames, or clinical trials – you need a mix of technical experts and subject specialists in order to preserve something. I think the role of a preservation administrator may be to broker all of those connections, and help each group have a meaningful conversation.