Monday, July 23, 2012
Posted by Robert David Graham (@ErrataRob)
I believe in evolution. The Internet is the product of thousands of people spanning industry, universities, and yes, even government. In order to prove their point, the supporters of Intelligent Design magnify the contributions of Government, and trivialize the contributions of everyone else.
The Darwin in this evolution debate is Gordon Moore, who back in the 1965 noticed that the number of transistors on a chip doubled roughly every 2 years. “Moore’s Law” has become our generation’s “Natural Selection”. Moore’s Law predicts that the “mainframe” computers of the time would eventually fit within your pocket. The iPhone indeed has as much compute power, storage, and network bandwidth as all the world’s computers combined when Moore first formulated his law.
That we would all be interconnected via a world-wide network is likewise a natural outgrowth of Moore’s Law. Had there been no Internet or World Wide Web, there would be something else very similar in its place.
Moore’s Law says that the computer evolves through thousands of small inventions, and is not the product of a few big inventions, nor is it the product of a Master Plan. Yes, some of those small inventions were by Government, which the Creationists seize upon in order to “prove” the Internet was created by Government.
Because the TCP/IP Internet has been the only network for 20 years, we have this perception that it appeared fully formed out of the void, that it was invented in a single step. The reality is that it evolved slowly over time, step by tiny step. During the Cold War, our government spewed research dollars indiscriminately at any project that might have military applications, which means that some of these dollars would inevitably touch TCP/IP. Many more government dollars also went to alternatives – because they didn’t know what they were doing. That government contributed to the version of the Internet that eventually won doesn’t mean that they deserve credit for it, or even that it was their plan.
The most famous government contribution is the specification for “TCP/IP” funded by the military by university researchers. TCP/IP is the “protocol” that runs the Internet. However, there is very little that is new or innovative in TCP/IP. It’s largely a clone of existing packet-switching technology that was mostly developed by the computer industry. What made TCP/IP different than competing solutions was the idea of “end-to-end” – but that, too, was copied from others. The inventors of the “end-to-end” idea (Saltzer, Reed, and Clark) deserve every much the fame as those who copied the idea (Bob Kahn, Vint Cerf).
I’m not saying Kahn and Cerf don’t deserve a lot of credit. They do. TCP/IP is an exceptionally well-written protocol compared to all the others they copied ideas from. I’m just trying to point out how much of TCP/IP relied upon the inventions of others.
We have this Lamarkian view that the Internet was designed this way on purpose. The reality is that that there were many competing versions of an Internet. Techniques that proved their worth survived and were copied by others. Techniques that work less well withered and died. When TCP/IP was created in the early 1980s, it had to compete against many alternatives from companies like IBM, DEC, and Xerox. It won the competition largely because it was the latest protocol to be invented, and had the benefit of learning from those that preceded it, to copy the best ideas.
Humorously, the TCP/IP Internet also had to compete against something known as “OSI”, which was the government-designed alternative. Yes, the government had a Master Plan, but that plan didn’t include TCP/IP. Even into the early 1990s when it became obvious that TCP/IP had defeated all competing protocols, government regulations still mandated that all government computers support the OSI protocols.
But this discussion of TCP/IP and competing protocols is a distraction. Protocols aren’t technology themselves, but simply control the technology. The technology itself that transports the data is the underlying physical network, namely the fiber optic network that crisscrosses the world. If there is any one big invention that deserves credit for the Internet, it is the invention of fiber optics in a Owens-Corning laboratory in the 1950s. If anybody deserves credit for building the Internet, it’s the trillion dollar investment by Wall Street that laid those cables across continents and under oceans.
I'm not trying to argue one side of this political debate such much as point out that it's only a political debate (not technical). It's a tautology: you can't use the idea that the "government created the Internet" to conclude the worth of government actions, because it's the focus on only worthy government actions that you use to prove the premise. It's a slight of hand ignoring the enormous contribution by non-government researchers. It's a circular argument, like claiming that the "Bible is the infallible word of God -- because it says so in the Bible".
In a recent speech claiming that nobody but government builds things, President Obama justified his claim by saying “The Internet didn’t get invented on its own. Government research created the Internet”. Well, yes, if you believe that only government could've created the Internet, then it naturally proves your assertion that only government could've created the Internet.