The Culture’s Immortal, but Not the Disk
For storing large amounts of data, nothing is more efficient or dependable than a computer--but only for a while.
That’s the realization computer scientists and archivists are coming to now that so many documents--ranging from U.S. census data to images of great works of art--are being stored in the digital code of 1s and 0s. It turns out that magnetic tapes, CD-ROMs and other storage media have shorter life spans than originally thought, and the machines and software needed to read them become obsolete even faster. Ironically, today’s high-tech storage methods are far less durable than ancient traditions of carving words in stone.
One of the first people to recognize and define the problem was Jeff Rothenberg, a senior computer scientist at the Rand Corp. in Santa Monica. Last week, the Getty Conservation Institute and the Getty Information Institute hosted a two-day conference at the Getty Center in Brentwood to call attention to the issue. Afterward, The Cutting Edge talked to Rothenberg about the scope of the problem and prospects for a solution.
Q: How did this problem sneak up on us?
A: For the most part, librarians haven’t digitized things in order to preserve them. Their goal is to make things that were formerly available only in dusty old rooms available over computer networks. But there is an implicit, fuzzy notion that if you digitize these materials, the preservation would take care of itself.
*
Q: But isn’t digital information supposed to be indestructible?
A: Digital information has the unique and wonderful quality that you can reproduce and distribute it infinitely without any loss of quality. So people got this feeling that once they’ve got it in digital, they’ve got it forever because they can make perfect copies.
*
Q: But that’s not true?
A: Not really. There’s a marvelous infinite longevity of digital information in the abstract, but when you come down to how is it actually recorded and stored, you suddenly hit this wall.
*
Q: What goes wrong?
A: The most obvious problem is that the media doesn’t last forever physically. We’ve moved from stone to papyrus to rag paper to wood paper to tape, and each has a shorter life span.
Then there’s the media obsolescence problem--how long will you have a machine into which you can stick a disk? It’s hard to find a computer that can read a 5 1/4-inch floppy disk anymore, let alone an 8-inch floppy. The 3 1/2-inch disk will probably go away pretty soon too.
Then you have to consider that the format in which the data are written changes. You really have to have the original software to read it, and of course that’s all saved on the same media, so it’s a recursive problem.
*
Q: So were scientists negligent in letting this happen?
A: The problem with computer scientists is that we are in the business of inventing the future. One of the reasons we’ve been so successful is that we’ve had the luxury of abandoning the past.
*
Q: What about the archivists?
A: For them, things move really slowly. They haven’t had to deal with sudden phenomena like this. But we are now in a position of recognizing that it’s a potential problem and we had better get our act together.
*
Q: What can we do about it now?
A: There are a couple of possible solutions. One thing people have proposed is instead of storing files in a proprietary format, we pick a standard that everyone uses. If we had a universal, for-all-time standard that could encode any kind of file we could possibly imagine, that would be good. But we’re 20 or 50 or 100 years away from that.
In the meantime, we can rely on standards, and as they evolve we can translate things from one standard to the next. But translations always lose something.
*
Q: What else can be done?
A: If you don’t want to rely on translations, then you have to run the original software. That means you also have to save the programs that run them, plus the operating-system environment and the platform, since they will all become obsolete.
You have to have a plan like Noah’s Ark, where the Smithsonian saves two of every machine that’s ever been built. But that’s just not workable in the long run, either.
*
Q: What do you think should be done?
A: The key is to find a way to run obsolete software. If you can’t run the original computers, what are your other options? My proposal is that along with each document, you save all the old software and you also save a description of the hardware platform so that you’ll be able to create a program that can emulate that software in the future.
*
Q: So you’re betting that will be an easy problem to solve in the future?
A: The one thing this approach relies on is that in the future, computers will be even faster and cheaper and more reliable than they are today.
*
Q: Why not just print out all of the computer programs onto paper?
A: You could print the bits, but you couldn’t print the documents because of hypertext. Plus, once you print it and it’s no longer in a machine-readable form, when it gets put back into a machine at some point, you’ll introduce errors.
*
Q: Why not just let all of this data go?
A: We don’t want to lose data like, “Where did we bury the plutonium?” But there are other aspects that are important--the fabric and history of our culture.
*
Q: Should we save everything?
A: We’ll only save a small portion of the data we generate no matter how heroic our efforts. It’s not a good idea to save everything, and probably most things deserve to be lost. Archivists only save things that are considered to be of enduring value.
But because of the short life cycle of digital information, it’s hard to figure out what will have enduring value. It’s relatively easy to decide when a document is 100 years old whether it’s of enduring value, but it’s pretty hard to decide when it’s only 5 years old, which is when we’re going to have to decide if we want to have a shot at saving it.
*
Q: Are some documents already irrevocably lost?
A: Yes, for some things it’s too late. It’s hard to say how much, though. The volume of stuff that’s lost may not be that huge yet, but it’s growing.
*
Q: How soon can we start putting a solution in place?
A: If we started doing something serious in this area right now, we could get at least a stopgap measure going in five years. This problem has been likened to the environmental movement in the 1970s.
In the beginning nobody wanted to listen, and it took 10 years before there was any real momentum because the problem was so overwhelming. This isn’t as bad because this problem can actually be solved if we put our minds to it.
*
Times staff writer Karen Kaplan can be reached via e-mail at karen.kaplan@latimes.com
