Whoops, there goes another CD-ROM

By: Laura Tangley
February 16, 1998 ‑ U.S. News and World Report

The parchment has yellowed and the ink is badly faded, but with a bit of effort one can still make out the words of the Declaration of Independence and the Constitution, penned more than two centuries ago. Both are painstakingly preserved by the National Archives in Washington, D.C., not merely as historic curiosities but, in the words of an official sign, as testimony "to the accountability of a government that lays itself open, through its records, to the scrutiny of present and future generations."

Future generations will be fortunate, however, they get a chance to view the records of the current Congress, or to look at some 8 million presidential files due to arrive at the National Archives soon after President Clinton leaves office. Most of the documents will be in the form of computer disks, CD‑ROMs, and magnetic tapes. And these modern record keepers, archivists and librarians warn, are turning out to be far less durable in many cases than simple parchment.

Tests by the National Media Lab show that op‑quality VHS tapes stored at room temperature preserve data dependably for just a decade. Average‑quality CD‑ROMs become unreliable ‑ some can be read, some can't ‑ after five years. And even when tapes and disks remain intact, the hardware and software needed to read them may no longer be available.

This is a formidable threat, considering that by the year 2000 about three quarters of all federal transactions will take place electronically. Records pertaining to health and human survival‑studies of disease transmission, for example, or the location of toxic‑waste sites ‑ are of particular concern. The danger extends to the nation's cultural legacy: Virtually all new music, animated art, and early drafts of literature and academic works are created and stored in computers. If such accomplishments are lost, says Deanna Marcum, president of the Council on Library and Information Resources, "we leave an incomplete legacy to future generations."

Part of the problem is that tapes and disks, unlike paper, often do not show degradation until it's too late. Occasionally tapes become so brittle that the magnetic coating actually separates from its backing. More often, however, the signs of damage are subtle: Routine exposure to everyday magnetic fields will rearrange some of the tape's magnetized iron particles. When a machine plays the degraded tapes, these alterations make the tapes unreadable, resulting in missing data. A few years ago, for example, scientists from NASA's Jet Propulsion Laboratory tried to read some of the magnetic tapes that contained the results of the 1976 Viking mission to Mars ‑ tapes that had been carefully stored and appeared to be in good shape. Even more troubling than the fragility of storage media, however, is how rapidly computer hardware and software become obsolete. Unlike paper documents and traditional audio and video recordings ‑ which as analog media present a continuous, start‑to‑finish record of information ‑ computers store data digitally. By breaking information into electronic or magnetic strings of 1s and 0s, digital technology has made it possible to store and access enormous volumes of data using very little space. But without the programs and equipment used to encode it, digital information makes no sense.

Time bomb. Any computer user who has tried to find software to translate WordPerfect 4.0 is already familiar with the problem. In government an industry, the difficulties are magnified. Dectape and UNIVAC drives, which for years recorded massive amounts of government data, have vanished, as have software programs like FORTAN II. Often, archivists don't discover that data are inaccessible until someone requests the information. Donald Waters, director of the Digital Library Federation, calls the problem a time bomb whose full impact will register only in the future.

Instances of lost data already have begun to surface. Waters himself faced the problem a few years ago when, as associate librarian for Yale University Library, he developed a project to transfer 2,000 books from microfilm to optical disk. Mid‑way through the project, the software he was using became obsolete, and the disks no longer could be easily read. At the New York State Archives, Margaret Hedstrom, now at the University of Michigan, tried unsuccessfully to read magnetic tapes, recorded in the 1960's which mapped land use throughout the state. On a larger scale, satellite photos of the Brazilian Amazon taken during the 1970s ‑ data critical to establishing deforestation trends both regionally and globally ‑ are also trapped on indecipherable magnetic tapes.

There's no quick fix in sight. But Marcum says that librarians and archivists must now start thinking about preservation as soon as new knowledge is generated ‑ deciding what's important to save, putting the information into as common and standard a format as possible, and carefully recording what machinery and software were used to encode the data. This week, the Getty Conservation and Information Institutes will convene a meeting in Los Angeles to discuss these issues. Some archivists look to technology as a long-term solution. At Rand, for example, senior computer scientist Jeff Rothenberg is designing new kinds of "emulation software," programs that instruct new computers to behave like older ones so they can decipher obsolete digital data.

Ironically, some of the latest ideas sound more like the past than the future. Norsam Technologies, for instance, is promoting HD‑Rosetta, a system that "permanently and safely" stores historical documents ‑ but only if they are converted from digital back to analog data. Another company, Cobblestone Software, actually uses paper to print out complex patterns of dots and dashes representing computer files. Called PaperDisk, the product is designed to resist the damage that heat, cold, and magnetism inflict on magnetic and optical media. The company says it invention should last for hundred of years: about as long as old‑fashioned, top‑quality paper.