Enter the labyrinth
The fire which destroyed as many as 20 million artifacts in Brazil's national museum was caused perhaps by the most trivial of accidental circumstances. The Brazilian culture minister's explanation that the blaze was likely caused "by either an electrical short-circuit or a homemade, paper hot-air balloon that may have landed on the roof" was greeted by riots because mischance does not satisfy the human urge to blame someone when things go wrong.
But catastrophes can happen for no discernible reason. Historically natural disasters have vied with human malevolence in the destruction of libraries. The great 1923 Kanto earthquake destroyed the chief Japanese libraries and the British Army burned the original Library of Congress in 1814. Against nature and man libraries cannot stand forever.
The destruction of 20 million Brazilian records, tragic as it might be, pales in comparison to the potential loss of the digital libraries upon which our world depends. The 21st century is generating libraries at an unimaginable rate and they are by no means safe. Just one radio telescope, the Square Kilometre Array (SKA), is expected to produce "up to one exabyte (10 to the 18th bytes) of data per day, roughly the amount handled by the entire Internet in 2000." According to a source quoted by Forbes, the world may be creating 163 zettabytes of data a year by 2025. (A zettabye is one trillion gigabytes.) Much of this information will be used to create the software libraries upon which our daily lives will increasingly depend.
It predicts that by 2025 nearly 20% of the data in the global datasphere will be critical to our daily lives. In eight years' time, an average connected person anywhere in the world will interact with connected devices nearly 4,800 times per day – one interaction every 18 seconds. ... more than a quarter of all data created will be real-time.
Data creation, it says, will shift its drive from entertainment content to productivity-driven and embedded data, as well as non-entertainment images and video such as surveillance and advertising.
Digital data is vulnerable not only to physical mischance but to logical perils as well. Unlike older forms, you can't even touch or feel them. In the past we could rely "on durable physical objects to carry knowledge across space and time ... cuneiforms date back 5,000 years and are still legible [to the unaided human eye] ... By contrast, digital data are ephemeral, easily overwritten, dependent upon hardware and software, decipherable only by machines." A 2002 article in MIT Technology Review noted that "we can’t guarantee that any [digitally produced data] will be usable 100, or 10, or even five years from now" simply because they may be stored on obsolete operating systems and forgotten file formats.
“The layman’s view is that digital information is more secure, when in fact it’s far more ephemeral ... We know how to keep paper intact for hundreds of years. But digital information is all in code. Without access to that code, it’s lost.”