Dream Futures, Archival Nightmares
THE FUTURIAN #6
Historians are keen to point out that the medieval “Dark Ages” refers more to a relative lack of historical records rather than a period of cultural decline. This was an age where reading and writing were specialist skills and tools for automated data storage and reproduction were basically non-existent. Fast forward to today and the rate at which we produce media is accelerating. In many ways its been accelerating for decades. By some estimates, the rate of publishing exceeded a normal reader’s capacity to keep pace sometime in the sixteenth century. In more recent years the Internet, cloud platforms and ubiquitous smart devices have enabled exponential growth in not just written media but photo and video media too. But in our ever-more digitized world data can be stored, shared, transmitted and backed up remotely with ease; so the prospect of losing data to a new dark age is laughable right? Right?
Well, lets start with the easy stuff, technologies that have survived since the medieval period, books.
Books are dying. Not in the melodramatic “people don’t read long-format fiction any more” kind of sense, but in the literal bindings disintegrating and pages falling to pieces sense; and not just a few books. One of the biggest developments in publishing was the invention of the paperback. Paperbacks enable much cheaper book production than traditional hardbacks and, with that, drastically reduced the cost of publishing and opened up vast new markets for mass-production fiction and non-fiction. Crucially, this made publishing under-served genres suddenly much more viable and the twentieth century saw a boom in so-called “pulp” fiction. This was literally because many of these books were printed on low-quality recycled paper stock, and other cost cutting measures that, increasingly, are making these books rarities. While well-made hardbacks from three centuries ago are still going strong, obscure genre fiction published in living memory is at risk of disappearing, potentially forever.
There are, of course, archivists working to preserve older works. Project Gutenberg and Librivox are both extensive passion projects seeking to digitize copyrighted material once it falls into the public domain, but the wait between publishing and something falling into the public realm is a long one, and one that remains the topic of fierce debate (Mickey Mouse, the works of H.P. Lovecraft and even the song “Happy Birthday” all come with extraordinary stories of the legal efforts taken to keep them out of the public domain). In the gulf in between low-end paperbacks are vulnerable to physical deterioration and outright loss.
Once something is digitized that’s only the start of the problem. Computers have been an important part of day-to-day life for less than half a century. In that time computing technology and digital storage formats have changed drastically. This presents a huge challenge for anyone seeking to digitize a document in a way that they want to be reasonably sure will be accessible in the future. For example, jpegs are one of if not the most common image format in use today. Their ubiquity means that, in all likelihood, devices will continue to be backwards compatible with jpeg files for a long time to come. But there’s a problem; jpegs are a pretty poor choice for actually preserving data; they use a lossy compression process that keeps files small but, over time and successive save cycles, leads to image deterioration and compression artefacts. Lossless compression formats are available, but they are broadly more obscure and the guessing game of which will be supported in the future is tricky.
The prospect of preserving data is also increasingly complicated by question marks over who actually owns it. Or rather who possesses and stores data and who can make use of it. Most terms of service include at least some allowance for companies to use your data either explicitly or implicitly and they are not necessarily under an obligation to preserve it in the highest fidelity possible. For example, Google photos offers a two-tier system. You can store photographs in their original resolution if you pay per gigabyte of storage. Alternatively, you can store an unlimited number of photographs if you limit the size/resolution to Google’s preferred constraints. Facebook (Meta) takes an even more aggressive approach to handling photographs with layers of processing on uploads that can raise very real questions as to whether the final file on their server that gets sent to browsers is really the same as the photograph you uploaded.
It’s perhaps also worth asking what happens if you want to get rid of your data. There can be any number of extremely valid reasons to want to delete things because of privacy, trauma, or, in the case of some creatives, simply being too embarrassed of your early work to ever risk letting another human see it. Various pieces of data protection legislation have been enacted to grant at least nominal control over users’ data. It’s nominally relatively straightforward to delete social media accounts and the data associated with them, but often companies will have already got what they want from you.
An ever growing part of Big Tech’s operations is data harvesting, processing and analysing as training sets for the myriad of advertising, image processing and other algorithms that form the heart of many modern enterprises. This means that the question of what counts as your data and their data gets very muddy very quickly, with your videos, photos, browsing history, and shitposts quickly turned into training sets for machine learning. You may delete the originals, but even aggressive purging of personal data might not stop it from living on in the lessons you can teach a company’s software.
And right now, that might be about as close as data can come to immortality.
© Charlie Garrud 2022