From: Kathy Kleiman [mailto:[log in to unmask]]

Because a domain name is not a Top Level Domain.
[Milton L Mueller] Sorry, Kathy, you're just wrong here. .FOO is a domain name. FOO.FOO is a domain name. FOO.FOO.FOO is a domain name. The DNS is recursive, is recursive, is recursive, is recursive...(repeat 64 times)

If we follow your reasoning, there's no reason to perform extensive Technical, Operational and Financial Showings/Review of New Registries.  We don't examine registrants, so why should we examine New gTLD Registries?

[Milton L Mueller] BINGO! You've just unmasked ICANN.

Fact is, there is _no_ reason to perform those showings other than the very fervent and politically potent desire of trademark interests, law enforcement, incumbent operators and governments to gain regulatory leverage over the domain name users and market. And I strongly suspect that you know that as well as I - but only when you are performing in your normal role as domain name rights advocate and not as lawyer for a client. ;-)

"Extensive technical, operational and financial showings" are regulatory mechanisms that reflect layer 9 policy demands. As layer 9 phenomena, they could be applied to SLDs as well as TLDs. Yes, the scary thing about your argument, Kathy, is that we COULD examine SLD registrants as intensively as we examine TLD registrants if the political demand exists. Indeed, some countries already do so (we will be visiting one in April). When you think up rationalizations for those regulations at one level, you are providing fodder for their application at any level.

Even in the most narrow technical sense, there is no difference. A badly operated TLD is no more or less impactful on "Internet stability" than a badly operated SLD. You may try to reply, "but wait, if .ORG goes down then millions of people are affected" but I will quickly reply "but wait, if FACEBOOK.COM goes down...." You get the picture. What matters is how many people use it, not what level of the DNS it is at.

If you are basing your entire argument against closed generics on the "special status" of one level of the DNS, you've lost the argument.