Digital information occupies a weird functional space in the information spectrum, turning a linear spectrum into a Möbius strip. Unlike earlier information storage media, digital information is easily copied and redistributed (a characteristic that has spawned a whole, massive new arm of intellectual property law). But unlike more physically tangible information storage media such as print or magnetic recordings (or socially shared memory in oral cultures, for that matter) digital media decay with frightening rapidity—both due to the relatively fast degradation of digital media (compared to earlier analogue media) and changes in digital media formats. I remember 8-inch floppy disks, but I wonder how many drives than can read them still exist?

Jim Barksdale and Francine Bateman at the Washington Post outline some of the challenges facing efforts to archive digital media. (Yeah, I know it seems like digital media should last forever—it's just one's and zeros. But CDs and DVDs are, due to their commodity status as economic products, usually not all that stable over time. See this Datacloud entry from a few years ago.)

Current estimates are that in 2006, 161 billion trillion bytes -- 161 exabytes -- of digital data were generated in the world -- equivalent to 12 stacks of books reaching from the Earth to the sun. In just 15 minutes, the world produces an amount of data equal to all the information held at the Library of Congress. While it is unrealistic to think that we will be able to preserve all the data produced solely in digital form, NDIIPP convenes top experts to help decide which at-risk content is most critical and how to go about saving it.

Responsible preservation of our most valued digital data requires answers to key questions: Which data should we keep and how should we keep it? How can we ensure that we can access it in five years, 100 years or 1,000 years? And, who will pay for it?

The importance of developing sensible plans to preserve our digital heritage cannot be minimized. We can't save it all, nor do we want to. It's also critical that we agree on how to save this data. In the next 100 years, we will go through dozens of generations of computers and storage media, and our digital data will need to be transferred from one generation to the next, and by someone we trust to do it.

[via the Association for Recorded Sound Discussion List]