Longtime readers of this site will be loath to discover another tedious book review of a soft technical book, but perhaps there are a few new readers who have some life left yet. Read on, unknowing vanguard of the newly alienated!

Jack Goldsmith and Tim Wu, both law professors, wrote Who Controls the Internet: Illusions of a Borderless World (Oxford University Press, 226 pp., $28.00) to present their answer to the question of “whether the technological changes of the last decade . . . have had a lasting effect on how nations, and their peoples, govern themselves” (p. 180). They make a compelling case that the optimistic hope that the internet would erase national boundaries has been replaced by a reality of local control leveraged through governmental pressure on intermediaries, at least in the case of large multinational companies.

Goldsmith and Wu cite the changing attitude of Yahoo’s Jerry Yang between mid-2000 and the fall of 2005. In 2000, Yang was defiant toward French regulations that prohibit the sale of Nazi goods: “Asking us to filter access to our sites according to the nationality of web surfers is very naive” (p. 6). Five years later, Yahoo, in compliance with Chinese law, collaborated with the Chinese government to identify a dissident journalist, Shi Tao, through his Yahoo email account. Said Yang, “To be doing business in China, or anywhere else in the world, we have to comply with local law” (p. 10). The information that prompted Chinese authorities to jail Tao, like the Nazi paraphrenalia offered for auction in 2000, was found on one of Yahoo’s servers in the United States. The situations are not exactly analogous, but Yang’s change in attitude is what’s important.

The enforcement pattern illustrated by Goldsmith and Wu is one in which local authorities pressure local intermediaries, generally ISPs, to control the content flowing over their wires. Search engines are also targeted, with Google.fr and Google.de cited as examples. Additionally, large multinational companies doing business over the internet need local law enforcement on their side. The capture of criminals is enacted as some local police officer pointing a gun at a would-be h4X0r and asking him or her to come quietly (as normal people are asleep at this hour). This is the major point that Goldsmith and Wu get right.
I’m not convinced by their thesis relating to free speech laws racing to match the lowest common denominator– I think that this is exactly what is occurring, but they disagree. In describing a case involving Australian libel laws and a wealthy Australian named Joseph Gutnick, Goldsmith and Wu suggest that wsj.com can choose not do business in Australia if they don’t want to be subject to Australian laws. “Compliance with Australian libel laws . . . is a cost of doing business in Australia” (p. 157).

Goldsmith and Wu assume that publishing on the internet works like publishing newspapers. In traditional publishing, newspapers read in Australia are printed and distributed in Australia; the printers and distributors are work in Australia. In internet publishing, documents are stored on a central server, and copies are transmitted to whichever computers request them. The server is maintained, at least in part, by people who work in the country where the server is hosted. Where a website does business is determined only by the consumers whose interests lead them to request the site using their web browser.

I could imagine a restricted case of the argument, in which we limit ourselves to the consideration of only subscription sites where local caches are used. If wsj.com maintained a server in Australia and charged subscribers to access it, that would be a compelling argument, but this does not describe the vast majority of the traffic on the internet. Most sites do not require subscription, and while many high-bandwidth publishers do maintain caches around the globe, most domains reside on only one server in one country.

Is it really nothing new that multi-national corporations need to either limit their speech to everyone or pay for communications filtering (in the form of IP geolocation)? In the case where people pay for permission to see content, as with the bulk of wsj.com, there is a business relationship formed. But consider a Z-list blogger such as myself who might decide to stain his soul by displaying Google ads for high-value keywords (exchange-traded funds, for example). When Goldsmith’s Australian cousin who likes reading the Z-list finds this review on Google and feels that I have besmirched his family name by disagreeing with his lawyer cousin, do I need to invest in geolocation filtering? Or will Google.com.au just do it for me? Preemptively, perhaps?

As one might expect from Oxford University Press, the mechanics of the book (rectangular pages, clear type) are well executed, save one typo in the last paragraph on page 160: missing period (or spurious capitalization) between “costly” and “But.” The cover, with its ghostly green pixel trails and monospaced subtitle, uses the imagery of the The Matrix, or was it the cover to Stephen Levy’s Crypto, published in 2000? Goldsmith and Wu’s book is printed on acid-free paper, but not on recycled paper. Crashing the Gate, by Jerome Armstrong and Markos Moulitsas Zuniga, which was published just a few days earlier, is of similar cost ($0.128/page vs. $0.124/page for Goldsmith and Wu), and target market, but it’s printed on “100 percent post-consumer-waste recycled” paper. Perhaps Oxford was concentrating its efforts on the editing, rather than the physical production of the book.

Unfortunately, the editing of the book could be better. We can be certain of this because an “adapted” subset of the book appeared in the January/February 2006 issue of Legal Affairs, and it was better. The improvement I noticed first was that the book’s strange assertion that “Apple came in Icelandic” (p. 50), which mistakes the company for its product, appeared as “Apple software came in Icelandic” in Legal Affairs (p. 42).

In general, the prose in Legal Affairs flows more smoothly and tells a more precise story. Consider these two sentences from the book: “Net geo-identification services are still relatively new but are starting to have their effects on e-commerce. Online fraud, and in particular online identity theft, has been a big challenge for e-commerce, causing firms and consumers to lose billions of dollars each year” (p. 61).

The corresponding sentences in Legal Affairs read: “Internet geo-identification services are still nascent, but they are starting to have effects on e-commerce. Online identity theft in the U.S. causes firms and consumers to lose billions of dollars each year.”

That’s a decrease of more than 20% in number of words for the same content and a 50% reduction in usage of the word “e-commerce.” Note also the substitution of “nascent” for “relatively new” and the omission of “their” as a descriptor of “effects” in the first sentence. It’s unfortunate that the more permanent version (the book) couldn’t be the better version.

Who Controls the Internet nails the most important points of support for its thesis. I would have liked to see more information about the rise of ICANN and the history of the organizations that preceded it. The central argument of the book convinced me that the idea that the decentralized architecture of the internet would ensure decentralized control of the internet was wrong; as Goldsmith and Wu claim, we are at “the beginning of a technological version of the cold war” (p. 184).