January 20, 2004

Misunderstanding the Internet

I have rarely come across a more tragic knot of misunderstanding regarding the Internet than the article by Adam Thierer at Cato, analyzing Howard Dean's Principles for an Internet Policy, David Weinberger's analysis of the Cato article, Lawrence Lessig's wrongheaded commentary on same, and Smart Mob's meta analysis of it all.

Every single one of them is wrong.

Let's start with what the Internet is. It is not a commons. It is not in the public domain, it is not a set of protocols, it is contracts. The physical computers are owned, but they are not the Internet. You can run the protocols but not be on the Internet. You can create your own network, but it is not the Internet. What is the Internet is the agreement you have to carry the traffic across your network from any computer on the network to any other computer on any other network that shares the same agreements.

The Internet is a remarkably efficient piece of legal contract work, informally enforced without very much state action at all. You break the rules and people generally just disconnect.

To say no one owns the Internet as Howard Dean does in his article is as true, and as fatuous, as saying no one owns the market. Furthermore, Howard Dean seems to have just come out against IPv6 in his point #4. The QoS bit would seem to be an endangered feature in Howard Dean's administration.

Markets are owned. They are creatures of private ownership, contracts, and settled rules. Government can help by enforcing evenhanded rules and insisting on truthfulness but that's about all. The Internet can use the same evenhanded rules and insistence on truthfulness. Legally recognizing digital signatures, for example, is a legitimate government function that enormously aids the progress of the Internet. At this stage in the game, it's a little difficult to tell where a candidate will jump. Fleshing out policy papers is usually a summer phenomenon but I'm not particularly cheerful at the way Dean's principles are written up. Number four is somewhat troubling, number one shows that Howard Dean doesn't understand the beast that has propelled him to prominence.

Thierry is, sad to say, not much better. He has the classic disdain for commons that every true libertarian has but sadly seems to lack the technical knowledge to correct Dean. He flails wildly and misunderstands what open access regulations are about. They are attempts at fixing the distortions caused by government granted monopolies and, as such, should be looked at as halfway houses to true infrastructure competition, not a dose of socialism in a robust capitalist market. Then he compounds his error by randomly shooting his mouth off about movements he obviously understands little about. The problem is that cyber economics has odd characteristics. There is a tremendous amount of unused capacity in our machines and in our networks. This drives people to make rational capitalist, free market decisions that would look utterly insane in the physical world. In on-line social forums, people actively encourage free riders for sane reasons. The word of mouth advertising value far exceeds the cost of their free riding but both are too small to bother measuring most of the time so it comes out looking and sounding like socialism when in reality it's a special case of capitalism that uses very specialized shorthand.

Weinberg is not much better. He does a semi-creditable job of tearing into Thierry but he is both in error in dismissing private property (which is what most of the Internet Infrastructure is) models and in characterizing the Internet as a set of protocols. I can run these protocols on my own network without being related to the Internet in any way. Once I have subscribed to the principles of carrying traffic and internetworking the specific collective called the Internet, I have joined the Internet even though I may not run TCP/IP on my own network but will encapsulate it/translate it into whatever I'm actually running as a network protocol.

It's this sort of fuzziness that makes non-technical free market theorists worry. They've never had peering agreements explained to them properly. They don't understand how "free" traffic flow is just a monitored settlement mechanism that just charges when there are persistent traffic imbalances between networks that exceed the cost of collections. It's as socialist as check clearinghouses.

Larry Lessig's contribution to this snafu is even more disappointing as not only does he have a law degree but he has made something of a national name for himself in the field. The protocols do not define the Internet. You do not have a degenerate case of an Internet when you have only one network running TCP/IP protocols. You have no internet because you are not inter (between) networking at all. You have to have at least two networks for the smallest case of an internetwork and considerably more before you can seriously label your collection of nets an internet and not get chuckles. The term Internet, by common usage, is only capitalized for the biggest worldwide internet. If almost everybody on the Internet voluntarily dropped off one day, changed their protocol stacks on all their equipment and joined a new set of networks with completely different protocols, but the same principle of carrying traffic, the few remaining networks on the old system would no longer be the Internet. The new network would justifiably claim that title. The appendix network would be an internet still, but not the Internet.

If there was venom in Thierry's assault on Lessig, Lessig returns the favor in classic snippy net style. Dollar signs instead of s's to imply a commercial sell out to the man! is just old. In place of education and clearing the air, he makes it much more likely that he'll have a persistent fight on his hands.

Finally, Howard, at Smart Mobs, ties the whole stinking mess together on the blog and adds his own small contribution to the mess, capitalist baiting. The Internet is not a commons. If it were, there would be no peering arrangements. You wouldn't be able to drop people from the net for abusing net policy. You would have a very unhealthy, unstabled, unuseful structure groaning under the weight of even more abuse than the present structure.

Posted by TMLutas at January 20, 2004 04:59 PM