David Clark ’66, a senior scientist at the MIT Computer Science and Artificial Intelligence Laboratory, spoke Thursday afternoon on some social aspects of how the Internet has developed. He has been instrumental in shaping the Internet and its underlying protocols since the seventies, and during the eighties served as its chief protocol architect.
Clark’s wife, also a Swarthmore alum, is a classicist. (He was an engineering major- “a mixed marriage,” he joked). He said that she convinced him not to use any slides during his talks because “Cicero never even heard of PowerPoint.”
Clark stressed how technical decisions made during the development of the Internet have affected whole industries today, and how social and other non-technical factors have changed the development of the Internet.
Early on in development, he said, the problem was how to get the system to work at all; in the eighties, it became how to get it to scale. Then, in the early nineties, he realized that it was no longer under the technologists’ control.
“First the economists came to us,” he said, “and said ‘you really screwed this up,’” because the Internet didn’t have built-in support for monetary transactions. (The idea of commercializing the Internet was “irritating” to the original developers.) Then the lawyers came, complaining of the lack of accountability, and the political scientists complained about the Internet’s agnosticism toward national sovereignty. He said “I knew I had really made it when a philosopher-ethicist came and complained,” about privacy concerns.
Clark discussed how the current state of the Internet Service Provider industry stems out of design decisions in the underlying protocols; the role that ISPs play is an area of “loose coupling” in the specification, and so it is a natural place for firms to establish themselves. Design decisions made in the seventies and eighties have determined the structure of a multi-billion industry today, something that Clark said no one thought about at the time.
He also described a “quality of service” protocol, which would allow applications that need high-speed access (like VoIP applications, such as Skype, and games) to perform better. The necessary technology was developed in the early nineties, and many multinational corporations with their own telecom structure use it today to great success. But because of “net-neutrality” concerns, the protocol has not been and probably will not be implemented on the Internet at large, out of fears that providers would use it to force consumers and content providers to pay more.
Another social issue affecting the internet is security. Some people favor strong privacy measures, while some “man in a trench coat who’s concerned with security, but the kind that has ‘national’ in front of it” would object that such measures make legal wiretaps impossible.
Sovereignty also poses problems for the Internet. As examples of how this can come up, Clark cited the 2000 case where Yahoo!’s auction service allowed the sale of Nazi memorabilia, which is illegal in France but protected in the United States, and the 2007 block of YouTube in Thailand over a video that insulted Thailand’s king, an illegal act in that country. Google and its unpopular decision to comply with Chinese law in censoring search results in that country, as well as its January announcement it would shut down its China operations after a cyberattack supposedly sponsored by the Chinese government, also came up.
The structure of the Internet, in many ways, embodies Western values and Western law, Clark said. How much are we comfortable with imposing that on other parts of the world with different values and different laws — even when we don’t like some of those values or some of those laws all that much?
Perhaps the day-to-day most frustrating effect not predicted by the developers of the Internet was the issue of spam. Early on, he said it just didn’t occur to anyone that spam would be a problem. “We should have talked to a social anthropologist,” he said, “or maybe just your mother, because she’d have known that some people would behave like that.”
Clark’s talk focused on how social and legal factors interact with technical aspects of the Internet. He ended his talk with another question on that topic: if we ever fight a cyberwar, should the computers wear uniforms? Computer attacks on, say, a hospital system, would be considered a war crime under the Geneva Convention, after all.