Within days of the Sept. 11 attacks, the federal Agency for Toxic Substances and Disease Registry rushed to pull a suddenly sensitive report from its Web site titled "Industrial Chemicals and Terrorism." The agency eliminated all traces of the document and its description of sources for home-brew nerve gases and improvised explosives.
But on the World Wide Web, almost nothing truly dies.
Indeed, the thorny report currently lives on at several locations, including the site for the Oklahoma City National Memorial Institute for the Prevention of Terrorism, a UC Santa Cruz graduate student's Web site and the databanks of the Internet Archive, a nonprofit venture that has electronically stored an estimated 10 billion Web pages in an effort to preserve the Web's history. The Toxic Substances and Disease Registry is one of several agencies -- public and private -- facing this problem. Contrary to concerns about too much censorship in the wake of the Sept. 11 terrorist attacks, the reality is that some agencies are having a hard time censoring anything that was once published on the Internet.
"The Internet is not like a faucet you can turn off and on. It's like a leaky faucet that keeps dripping long after it's turned off," said Gary Bass, executive director of OMB Watch, an organization that strives to cut back on government secrecy.
Still scattered across the electronic ether are a host of "erased" documents, including maps of nuclear reactors, pictures of secret spy satellite facilities and a description of a NASA space propulsion project.
In many cases, agencies had no idea that their erased documents are still available for anyone with a Web browser and Internet link. Detailed maps from the Energy Department's International Nuclear Safety Center, for example, are still retrievable through the Internet Archive.
"I have never heard of the archive," said Jeff Binder, director of the center. "Maybe our guys in cybersecurity have."
In the electronic battle against terrorism, the Web has become as porous a landscape as the real battlefronts surrounding Kunduz or Kandahar in Afghanistan.
That's largely due to a kind of Xerox effect on the Web, where pages and even entire digital sites can be easily copied with a few mouse clicks.
Copies of supposedly eradicated reports and documents can be found using common search engines and the Internet Archive's whimsically named Wayback Machine. The "Industrial Chemicals and Terrorism" report can be found in a matter of minutes, even by novices.
Until the Sept. 11 attacks, the porousness of the Web was actually a feature celebrated both in and out of government as a way of providing instant global distribution of information.
Anti-secrecy group now pulling pages
For Steven Aftergood, director of the project on government secrecy for the Federation of American Scientists, the Internet has been a primary tool in the organization's efforts to battle what it considers misuses of government secrecy. It collects and disseminates information on nuclear weapons, the "Star Wars" antiballistic missile initiative and other projects.
Indeed, the Washington-based group was created after World War II by scientists from the supersecretive Manhattan Project worried that the government was concealing the dangers of building a nuclear arsenal.
But since Sept. 11, Aftergood has found himself in the awkward position of following the government's lead in protecting sensitive information. So far, he has removed about 200 pages from the federation's site, mostly concerning intelligence and nuclear weapons facilities.
Or so he thought.
One of the sections he removed from the site showed pictures of an unidentified government agency that analyzes spy satellite data. Included in the cache was a closeup of a sign pinpointing the location of the building and the opening in the security fence used to admit cars.
"I thought the pictures showed where someone could maybe drive a bomb inside," Aftergood said.
What Aftergood did not know was that in the months that the photographs were on his computer servers, a succession of small programs, known as "bots," had scoured the site, indexed its contents and copied them to the Internet Archive.
From today's newspaper