This is not an issue of trading off closed doors for quickness or efficiency. There is not a credible claim that ICANN has been either quick or efficient.

The root of the problem, I believe, was ICANN's original "volunteer" lawyer, Joe Sims. He wrote the bylaws of ICANN and structured it so that it would be, as completely as possible, unaccountable to anybody. I was involved in these discussions while Jon Postel [the person responsible for assigning numerical Internet addresses for much of the Internet's existence] was alive; we were all afraid that any small organization set up to end the NSI [Network Solutions] monopoly [on domain names] would be sued out of existence by NSI. We didn't want NSI to succeed at that.

But there's a difference between accountability and transparency -- and a difference between being immune from suit by a subcontractor, and a trustee of a global resource being immune from scrutiny by the users of that resource. EFF's comments on the initial draft NSI charter and bylaws said, "There's no transparency, no accountability, and no protection of civil rights," and we made specific wording suggestions as to how they could fix that. They ignored those comments, and formed ICANN anyway. The culture of secrecy and unaccountability has festered in there ever since. They hold "open public meetings" where the public is free to shovel its comments into a dumpster. But then they ignore the comments and do what they want in closed-door meetings.

	Jesus Christ Superstar: in 'Masterpiece' presented by Lexus

You spent two years working on domain name issues at CORE, a not-for-profit domain registry, and you've argued that even then, corporate interests trumped technical needs. In your letter, you also mention people "pulling strings" behind the scenes. Who would these people be? Do you have any theories?

The strings that were pulled before and during the Clinton administration's "Green Paper" and "White Paper" process, that ultimately resulted in the creation of NewCo, also known as ICANN, were pulled by SAIC. SAIC is a very interesting for-profit company with a multibillion-dollar annual revenue, most of which comes from classified contracts with the U.S. military. What's even more interesting about SAIC is that there is no external control on it: It is "employee-owned," i.e., there are no outside stockholders. If you leave the company, you have to sell your shares in it. SAIC's board of directors reads like a who's who of the military-industrial complex (former secretaries of defense, spy-agency heads, etc.). When you read about the government wasting billions on "homeland security," guess who gets it. SAIC's home page features their new brochure on "SAIC -- Securing the Homeland."

Somebody at SAIC noticed that a tiny company had gotten the temporary monopoly to run the domain name system, and was being paid a few million dollars by the government, over a few years, to do all the work. In March 1995, SAIC acquired this company (Network Solutions) for $3 million, from its founder, who had won the bid because his five- or 10-person company was "minority owned." (He later complained bitterly that they'd screwed him.)

Within the next six months, somebody inside the U.S. government suddenly decided that Network Solutions (the new SAIC subsidiary) could charge every domain name holder $50 per year, extracting hundreds of millions of dollars from Internet users. That policy was instituted despite the best efforts of the Internet community to stop it. That's one string that was pulled. Who exactly pulled it? Sounds like a job for an investigative reporter.

I helped to design and build the infrastructure for CORE to become a domain name registry. It cost us less than 25 cents per year per name to run. Even if you added the likely legal bills from NSI suing us, it amounted to less than $2 per year for each domain name. NSI is still charging $6 per year, and doing it in much higher volumes, where it should actually cost them less than 1 cent per year to do the work.

NSI, wholly owned by SAIC, then filed for a public offering. You can read their prospectus. It said, effectively, SAIC is going to keep 90 percent of the shares, and SAIC's shares each get 10 votes compared to the shares we'll sell to the public. They sold off 10 percent of the company for $54 million, during the early part of the Internet stock craze. That 10 percent owned by the public had only 1 percent of the voting power, i.e., even though it was a "public company," SAIC could do exactly what it wanted with the company, whether it was in the stockholders' interest or not. Indeed, the prospectus was blatant enough to say that it would immediately transfer significant money (I think it was $20 million) to its parent company, SAIC, from the proceeds of going public!

The next challenge was maintaining the SAIC monopoly on selling domain names. Clearly when you're selling something for more than 50 times its cost, you need a way to keep out competitors, or the price will drop. The NSI contract with the government was coming up for renewal, and since it had only been a small contract, it had been competitively bid for a fixed period, and was intended to be competitively rebid after that period. Also, lots of angry people were complaining about how rotten the NSI domain name registration policies and prices were. There were two challenges: keeping the government contract despite high prices and lousy service; and preventing other top-level domains from offering reasonable prices and good service. This is where the second string was pulled.

Clinton's White House delegated the job of sorting out this mess to Ira Magaziner. He held meetings with all the stakeholders, generally telling each of them what it wanted to hear, in order to extract more information and cooperation from each one. I also talked to all the stakeholders, and it was clear that the general consensus favored competition and sane policies, though you sometimes had to cut through a thick fog of private interests pulling for their own advantage.

But somehow in the actual government decisions, what always ended up happening was that NSI would keep its monopoly. The rationale was "stability of the Internet" (a hoax, since hundreds or thousands of people in addition to NSI were capable of doing the technical work involved), "avoiding trademark squatting" (NSI had already set its domain-dispute policy to "anybody with a trademark trumps anybody without one," so they were working in league with the trademark lobby) or "preventing chaos" (another word for competition). Somehow that statesmanlike guy Ira Magaziner managed to screw the public, screw competition and reward the billionaire spook contractor who had the monopoly contract.

But this isn't the end of the story. SAIC was particularly smart. They knew the monopoly wouldn't last forever, so they milked it and then got rid of it to a sucker. They dribbled out more NSI stock into the market as the stock price rebounded after the monopoly was secured. Then in March 2000, they sold the entire company to VeriSign for $21 billion in stock (yes, that's billion). SAIC almost certainly had enough sense to sell off at least a good chunk of this stock before it dropped through the floor. (Try reading SAIC's SEC reports for 2000 and early 2001, maybe you can tell.) Thus SAIC made billions of dollars out of spending $3 million and pulling a few strings to keep the monopoly alive.

There were other strings pulled at other times (e.g., when Vint Cerf had just joined the ICANN board, the board voted to convert NSI's short-term contract to run the "com" [top level domain name] into a permanent contract), but those first two SAIC tugs were the key ones.

The Free Software Project
Read Andrew Leonard's book-in-progress on Linux and open source -- and post your comments.