It’s a basic premise of the digital preservation community that preservation will require ongoing effort over the years. Let an archive lie neglected for twenty or thirty years, and you might as well throw it away. No one will know how to plug in that piece of hardware. If they do, it’ll have stopped working. If it still works, its files will be in some long-forgotten format.
The trouble is, this is an untenable requirement over the long run. Institutions disappear. Wars happen. Governments are replaced. Budgets get cut. Projects get dropped. Organizational interests change. The contents of an archive may be deemed heretical or politically inconvenient. The expectation that over a period of centuries, institutions will actively preserve any given archive is a shaky one.
Information from past centuries has survived not by active maintenance, but by luck and durability. Much of the oldest information we have was carved into stone walls and tablets. It lay forgotten for centuries, till someone went digging for it. There were issues with the data format, to be sure; people worked for decades to figure out hieroglyphics and cuneiform, and no one’s cracked Linear A yet. But at least we have the data.
Preservation of digital data over a comparable time span requires storage with similar longevity. This is a very difficult problem. If it’s hard to figure out writing from three thousand years ago, how will people three thousand years from now make any sense of a 21st century storage device? But we have advantages. Global communication means that information doesn’t stay hidden in one corner of the world, where it can be wiped out. Today’s major languages aren’t likely to be totally forgotten. As long as enough information is passed down through each generation to allow deciphering of our stone tablets, people in future centuries will be able to extract their information.
What we don’t have is the tablets. Our best digital media are intended to last for decades, not centuries. Archivists should be looking into technologies that can really last, that will be standardized so that the knowledge of how to read them stands a good chance of surviving.