Wednesday, July 18, 2012

Internet revisionism

At a campaign stop last week in Virginia, the president said

The Internet didn’t get invented on its own. Government research created the Internet so that all the companies could make money off the Internet.
He also made some other controversial remarks on the role of government in a capitalist economy, but I just want to fact check these two sentences.

Yes, nothing "gets invented on its own." It takes people to do that, whether individuals or quasi-permanent groupings working in an organization who (under US law) are people too (and have been since Roman days). And it’s also true that a complex systems architecture — of which the Internet is the most complex — requires coordination of the distributed efforts of a large number of people.

But who did the inventing? To me, “government research” implies “government researchers” when it was actually university and corporate researchers. Yes, much of this research was government-funded research, including most of the research in the 1960s and 1970s.

The definitive first-hand account of the creation of the Internet was published in 1997 in the leading US computing journal. It said
  • An MIT professor, J. C. R. Licklider, conceived of a "galactic network” in 1962 [although other accounts say he was then a vice president of BBN, an MIT spinoff company]. He then went to ARPA, the DoD’s advanced research funding agency
  • An MIT graduate student, Len Kleinrock, wrote a paper about packet switching in 1961
  • An MIT [Lincoln Labs] researcher, Lawrence Roberts, set up the wide area network (over a dialup line) in 1965 and then in 1966 went to ARPA and proposed the ARPANET. [Wikipedia says in 1971 he went on to found Telnet in 1971, a packet switching common carrier].
  • The switches that made the ARPANET possible were built by BBN, and Kleinrock (then at UCLA) got the first one in 1969.
  • The Network Working Group [a small group of university and nonprofit software engineers] developed a spec for the first host-to-host communication standard in 1970
  • In 1972, Ray Tomlinson of BBN wrote the first e-mail program.
  • (D)ARPA funded a spec for TCP, which was implemented by Berkeley in Unix
  • Connecting more than one machine at a given site was made possible by Ethernet, invented by Xerox PARC [and then sold as hardware by companies like 3Com.]
  • The domain name system was created at USC.
  • In 1985, the NSF created a second network, NSFNET, which would serve universities, not just the DoD and DoD contractors — and then defunded it in 1995, forcing it find its own way to self-finance.
  • Leadership of the Internet passed to self-organizing [formal or informal] nonprofit entities, including the Internet Engineering Task Force, the Internet Society, the Internet Architecture Board, the Internet Engineering Steering Group and the World-Wide Web Consortium. [Starting in the 1990s, most of the resources for these entities were provided by corporations and universities using their own funds]
So yes, it took a village to create the Internet, and the ball would not have been started rolling without ARPA’s sustained funding over many years. But the actual work of designing and building the Internet was not done by the government, but for the government by smart people that it picked.

Then there is the question of what happened after the ARPA-designed data pipes were in place. An independent account by David Mowery and Tim Simcoe of Berkeley wrote in 2002:
Adoption of the Internet in the US was encouraged by antitrust and regulatory policies that weakened the market power of established telecommunications firms and aided the emergence of a domestic ISP (Internet Service Provider) industry. The large size of the US domestic market, as well as American firms’ large investments in desktop computing and computer networks, created the conditions for rapid diffusion of the Internet following the introduction of the WWW. “Network effects” created by the scale of the US market and the predominance of English language content also contributed to rapid US standardization and diffusion.

During the late 1990s, the Internet entered a third phase of growth characterized by the development of commercial content and business applications. This phase followed the completion of a long process of infrastructure privatization and a dramatic surge in Internet use associated with the introduction of the WWW. Commercial interest and activity were fueled by the availability of capital from the US venture capital (VC) industry, as well as the strong performance of the US economy.
Mowery would certainly be the first to argue for the importance of government-funded research, but in the end, the Internet would have been little more than a research curiosity (or an internal network for a few universities or DoD sites) without the private investment necessary to grow it into what we have today.

As campaign hyperbole goes, this probably rates only one or two Pinocchios — certainly not a whopper on par with Al Gore claiming he invented the Internet. But I’d hate to think that young people, listening to soundbites, took away from this campaign claim that an omniscient and omnipotent Federal government is how we got the Internet and how we will get similar innovations in the future.

No comments: