Saturday, March 7, 2026

DT26008 Storing Data for Posterity V01 070326

 The faster digital tech developed, the more data we lost — until now

Making Science with Tom Whipple: a new podcast

Tom Whipple

About 38,000 years ago, in a German cave, our ancestors carved symbols in a mammoth tusk. Today, in every indentation, you can still see our shared humanity, stretching over millennia. You can also see something else. Encoded on the tusk, archaeologists argued recently, is data that those ancestors wanted to preserve. It is proto-writing.

Twenty-five years ago, while a student, I also had data to preserve. I took some digital photographs and downloaded them to my desktop.

Three years later, I transferred them to my laptop. When that died, I removed the hard drive, bought a fiddly device to read it, and moved them to my new laptop. Ten years ago, tired of the laptop relay, I moved them to a home cloud server.

Then, two years ago, the cloud company emailed me: they no longer supported the hardware. Next time I tried to access it, I couldn’t. The mammoth tusk survived glaciation, civilisational collapse, Black Death and two world wars. My photos? Done in by a software update.

It’s no great loss to posterity. What it is, though, is an illustration of what’s known as the digital dark ages — a gap in the historical record coinciding with the arrival of the computer. You see the digital dark ages in the floppy disks in landfill. In Wikipedia citations that lead to 404 errors. In the great digital cataclysms — the internet equivalent of biblical floods. Myspace? One day the profiles just disappeared.

GeoCities, that once-thriving community of amateur websites and garish gifs? Dashed by the Silicon Valley gods in an afternoon.

You see it, most ironically, in the Domesday Book. In 1986, the BBC marked the 900th anniversary by making a modern version. It conducted a new census of the nation, taking down thoughts, feelings and occupations, all recorded using a technology 11thcentury scribes would have marvelled at. A technology that was, a few years later, defunct: the LaserDisc. A few years after that, no computer could read it. The original Domesday Book persisted, unchanged, on vellum. The modern one barely lasted a generation.

These days, if we really want to store data, we use magnetic tape. It works. But it’s fragile. Every ten years we have to copy it to a new tape. There are documents in the National Archives, held under the 30-year rule, that will be copied three times before anyone can read them.

Preservation, still, requires a tenuous civilisational thread, reliant on electricity, air conditioning and librarians. We are, still, in the digital dark ages.

Last month there was a bit of light.

Computer scientists unveiled Project Silica. In a Microsoft lab in Cambridge there are blocks of glass, each like a chunky microscope slide.

Hold them up, as I did, and they shimmer — imperfections catch the light, like notches in ivory. They have been carved not with flint but lasers.

Each stores seven terabytes of data.

That’s 10,000 CDs or (roughly) 600 billion mammoth tusks. Each can last 10,000 years, without requiring civilisation (and electricity) to last.

Maybe they will become the archive standard. Maybe it will be something else. But these blocks are a sign that we are emerging from the digital dark ages. Let’s just hope that, unlike the cave people, we also leave the instructions on how to read them.

No comments:

Post a Comment