National Archives in WIRED on-line report

["James J. O'Donnell" <jod@georgetown.edu>]

POSTS|national archives

As a young historian visiting the National Archives more than four decades
ago, Allen Weinstein met an employee named Mr. Taylor who seemed to know
the whereabouts of every document - from the Declaration of Independence
to the latest Bureau of Mines report - in the entire block-long
neoclassical complex. Mr. Taylor was still working there last February
when Weinstein was sworn in as the ninth archivist of the United States.

Weinstein has a lot to say about the 84-year-old civil servant when I meet
him in his vast office, furnished mostly in the generic
colonial-federalist style favored by embassies and bed-and-breakfasts. It
doesn't matter that I've come to talk about the new Electronic Records
Archives project, which Lockheed-Martin will build at a cost of $308
million over the next six years. "What are your hopes for ERA?" I ask the
nation's archivist. "What are your concerns?"

"I worry about losing Mr. Taylor," he mumbles, his voice barely audible.
Weinstein is a lean man with sparse white hair, a round face drawn into a
pout, and a laminated ID hanging over his necktie. He clings with both
hands to the padded arms of his chair. "I worry about losing Mr. Taylor's
expertise and the expertise of people like him. They've been living with
history."

Several rooms away, one of Weinstein's deputies, Robert Chadduck, is
trying to figure out how to live with history of the digital variety. As
research director of ERA, he has set up his computer laboratory in the
only space he could get: the conference room of the archives' National
Historical Publications and Records Commission. On the shelves surrounding
his three screens are hard=ADcover reprints of writings by Frederick
Douglass and John Marshall. He doesn't appear to notice the books as he
paces the room, waiting for an assistant named Mark to call up some data
he wants me to see. "ERA is unprecedented for the National Archives,"
Chadduck says, and his voice almost breaks with pride as he informs me
that "it's identified in the president's budget as a major systems
acquisition."

On the monitors behind Chadduck, I watch Mark mutely arrange the 
images. There's a topographical map of Hawaii, a fly-through simulation of 
the
Great Lakes, and a virtual reality model of a NASA space platform. Each 
is stored on disk or tape in a different location: the National Archives'
College Park storage facility, the University of Maryland, the San Diego
Supercomputer Center. "The challenge is to create a transcontinental,
persistent archive," one that can be accessed from anywhere, at any time,
Chadduck says. But his sample fails to convey the vastness of the
estimated 347 petabytes of data preserved in the archives in thousands of
digital formats.

The National Archives has been receiving electronic materials since 1970,
but plans for long-term preservation of it all didn't begin until 1998.
And the government has only started to take it seriously in the past three
years. "Isn't that a bit late?" I ask Chadduck. "I won't presume to speak
for the White House," he replies testily. He directs my attention
elsewhere, handing me a small poster cluttered with more logos than a
stock car. There are insignias for the Library of Congress, the National
Science Foundation, the Department of State, and the National Nuclear 
Security Administration, among others, each indicating a research
partnership that the National Archives has recently established. These
entities have been grappling with digital preservation for a while, and
Chadduck hopes to benefit from their efforts. "We don't want to 
reinvent the wheel," he says.

So ERA will be a modular system, relying as much as possible on
technologies (many of them open source) developed elsewhere. For
example, one anticipated module will be responsible for determining what
kind of software was used to create an incoming document. Another will
translate it into a usable format. Others will handle distribution, 
backup, and searchability. The modules can be replaced or added as
technology advances - there would never be a need to reengineer the entire
system. Another boon, at least from a bureaucratic standpoint, is that
nobody has to define the limits of what the system will actually do.

Of course, no matter how the system evolves between now and 2011, one
module it won't encompass is Mr. Taylor. While Lockheed's design prototype
emphasizes intuitive access for users ranging from amateur genealogists to
career paper pushers, no software on the market today or in the future is
likely to have the veteran archivist's idiosyncratic expertise, his
intuitive grasp of the collection's contents.

Mr. Taylor is elusive these days. He hides in the stacks whenever
Weinstein shuffles by. "He's afraid that the archivist is trying to retire
him," explains Miriam Kleiman, another Taylor protégé. 
"Retire him?"
Weinstein counters, enunciating his words for the first time since I've
met him. "I want him to work 40 more hours a week."

He sighs. Weinstein has just been given the largest appropriation in
National Archives history, for a system that will be the envy of every
library around the world, but I see that it doesn't satisfy him in the
least. "What happens next, I'm not sure, just not sure," he mutters.

The archivist of the United States looks down at the floor. What he really 
needs, no technology can provide.

- Jonathon Keats
---2071850956-1910065217-1136602098=:27615--