The challenge of archiving digital media
February 24, 2010 | 8:45 am
By Chris Meadows
It is an interesting conundrum that every new data storage medium invented since the stone tablet has had a shorter physical lifetime before it degrades. American Scientist looks at this issue in light of the preservation problems presented by digital media.
Kurt D. Bollacker begins by talking about issues in data migration from his own past (such as the time he painstakingly backed up his hard drive onto 20 floppy disks…then when he needed to restore discovered he could no longer find the backup software), then talks about how our current digital lifestyle makes it harder to keep our data safe.
From an e-book reader’s point of view, this can be a problem: physical libraries will remain as they are for decades, while an e-book library can be wiped out by one hard drive crash. (Though some e-book fans have lost all their paper books in house fires but still had their e-book library safe on their computers, so this is not necessarily always true.)
Bollacker goes into a great amount of scientific detail about how errors can creep in when media is damaged, and the fact that greater data densities in newer digital media means that the same amount of physical damage to a denser medium results in losing considerably more information. He also talks about methods of error checking and correction to lessen this type of damage.
But on the other hand, the best medium in the world is no good to us if we lose the player for it. This problem is happening in the short term with older media formats from the early days of computers—how much more will it affect information archaeologists of our future? Even if they find a DVD that has been perfectly preserved for all that time, will they be able to read it a thousand years from now?
The best solution, Bollacker suggests, is “data promiscuity”: making frequent backups, and transitioning media from old formats to new ones as they become available. But there are alternatives.
The Rosetta Project, which aims to preserve modern language texts into the distant future, has created what amounts to a microminiaturized “stone tablet”, etching microscopic images of 13,000 pages of text and images onto a metal disk that could last as long as 2,000 years.
(Found via Slashdot.)



Previous

SUBSCRIBE TO RSS
Comments:
I’d recommend looking at LOCKSS, which stands for Lots Of Copies Keep Stuff Safe. It’s a network of communicating file server which trade copies of what they have, with a voting mechanism to repair damage copies. As individual nodes get upgraded, the data automatically moves onto the newer storage media.