By Bruce Einhorn
Can Chinese Net surfers see George W. Bush's home page? Jonathan Zittrain wanted to know. The Harvard Law School professor, working with a first-year student, 22-year-old Ben Edelman, recently started researching policies that governments impose to censor the Net, and the pair devised a program that allows them to sit in their Cambridge (Mass.) office and see what China's Web cops are up to.
According to Zittrain, you can easily type a Web-site address into the program and see whether it's accessible in China. Zittrain says he and Edelman were "agog" to discover that www.uscourts.gov, the Web site of America's federal court system, was off-limits. I recently talked to Zittrain and Edelman on the phone, discussing the latest crackdown by China on the Internet, and I asked Zittrain if other U.S. government sites are taboo too. "I'm going to test www.whitehouse.gov right now," he tells me and fades out of the conversation. For a moment or two, Edelman carries on alone, but then Zittrain is back with his answer.
"Whitehouse.gov is accessible," he says. Not only is President Bush's site accessible, but so too is Secretary of State Colin Powell's, since according to the Harvard duo's program, the State Dept.'s Web site isn't banned. But, Zittrain adds, not all U.S. government offices are so fortunate. The Voice of America's site is taboo, Zittrain reports. Academia can be blacklisted too: Stanford University's and New York University's sites are off-limits.
Zittrain and Edelman are at the forefront of this research, and their work is creating waves felt far beyond the banks of the Charles River (see BW, 9/23/02, "The Great Firewall of China"). For years, people have known that Beijing has been censoring the Internet. But the Harvard duo is focusing attention on the issue by identifying systematically the names on China's blacklist.
In early September, they announced the first batch of Web sites that they had determined were banned in China. Besides the predictable pornographic, prodemocracy or religious Web sites, the list included major U.S.-based news outlets, universities and government organizations. And surprisingly, they also found that search engines Google and AltaVista, popular among Chinese Net surfers, had been banned. (Since then, news reports indicate the Chinese have unblocked Google.)
Here are edited excerpts of the rest of the phone conversation I had with Zittrain and Edelman:
Q: Why are you doing this research?
Zittrain:We've both long been interested in tracking filtering on the Net. On all levels -- from parents installing filtering software for their kids, all the way up the line to ISP [Internet service provider] filtering that a government does for a whole country.
Q: How much typically gets filtered?
Zittrain:The lists are never made public. You don't know what's being filtered. The companies that maintain these lists, the ones that sell filtering software, it's like a trade secret [for them].
Q: Who are the customers for this kind of software?
Zittrain:[Customers include] companies that might want their employees to be restricted, parents for their kids, and one government -- Saudi Arabia.
Q: What's the difference between what happens on the individual level --parents looking to screen what sort of stuff their kids see on the Net -- and what organizations or governments do on a mass level?
Zittrain:The companies [that make the filtering programs] produce software for PCs or for servers that works to restrict what servers can do. There's a base set -- a starter pack -- that has a list of sites that have sexually explicit content.
The Internet became mainstream in the mid-1990s. After that, you had some demand created for this filtering software. If we implement a system where parents can shield their kids from porn, that system might be adopted by a country in one fell swoop. That sort of vertical portability is actually happening. In Saudi Arabia, many of the sites they block for sexually explicit content [came from the initial filtering list]. They add their own sites that are politically sensitive.
Q: What has been the response since you went public with your China findings?
Zittrain:The most interesting question right now is the public use of our Web page -- by which they can enter the site and test it against the China firewall. Over 100,000 sites are in the repertoire of what we test. What's downright amazing to us is the number of additional sites the public has discovered that were blocked.
Edelman: We've gotten maybe 60,000 visits in the last week and a half. We have about 10,000 new sites to test. Now we have 180,000 sites to ask about.
Q: And how many are blocked?
Edelman:We need to test three or four times. [But there are] several thousand that we know are blocked.
Q: Any surprises?
Zittrain:One of the first things we were agog at was that www.uscourts.gov was blocked -- and was consistently blocked. You wouldn't expect that the federal judiciary is a hotbed of revolutionary [activity].
Edelman: Chinese students who have Web pages in the U.S. [with antigovernment content], when they post those pages on main university Web servers, the Chinese government seeks to restrict access to those pages. So for mit.edu, the main server is blocked. Caltech.edu. Columbia.edu. Each of those has some Falun Gong or democracy [content]. The expression "cutting off one's nose to spite one's face" might be appropriate here. They appear to be prepared to sacrifice access to all of these universities sites.
Q: What are some of the goals of your research?
Zittrain:What we're most curious about is the extent to which China's Internet filtering policy is an active instrument in the toolbox of the government, of their politics, of their views. If tensions with Taiwan rise at a moment, is that reflected in the set of sites filtered at that moment? Do they set up [the filtering system] and then put it on autopilot? Or is it important to them, [so that] every week we see them tending to this list? To answer that question well, requires good longitudinal data over time to tell what's blocked.
Q: Lately, the censorship seems to have changed, with censors rerouting requests to other approved Web sites, right?
Zittrain:They're giving us a great puzzle. This makes our research much more interesting. The closest I've ever seen [to this before] is being directed to a page that says "sorry we are not going to give you what you want." But I don't think we have seen anybody just redirecting to an approved search engine rather than the one you ask for.
Q:: How does China's filtering differ from that in Saudi Arabia, the other country you've studied?
Edelman:The Saudis use proxy servers to examine the substance of every request. They look at the entirety of the request. In China, they look only at the IP address that is being requested. Because China is willing to accept [so-called] overblocking, the technical details of the task are quite a bit easier. It's not that hard or costly to configure their routers. If they want cutting-edge blockers, they would need to send [traffic] through more sophisticated [equipment] and that is slow and expensive.
China is a much bigger place. It processes a lot of data -- much more than any 100 proxy servers in Saudi Arabia. In China, they would need thousands, tens of thousands, one doesn't even want to guess.
Einhorn covers technology from Hong Kong for BusinessWeek. Follow his weekly Online Asia column, only on BusinessWeek Online
Edited by Beth Belton