ICANN is a body that has resolutely refused to try to figure out how many new TLDs might safely be created -- most likely because the answer is "a lot" and this isn't an answer that furthers the interests of either the IP lobby or the incumbent registries. Manheim and Solum's suggestion that a technical coordination body that doesn't even have a Chief Technologist really needs a Chief Economist most of all is certainly refreshing.

Of course the the issue of how to allocate new TLDs is only the second hardest issue in new TLD policy -- the hardest is how many new TLDs to have per year. (And the third hardest might be what to do with excess profits created by an auction....). Judge for yourself if Manheim and Solum can charm ICANN out of the trap it has driven itself into (and they are very charming!), or if they are just too optimistic:

Closely related to the question of which gTLDs should be added is the issue of how many should be authorized at any time. Here ICANN must play an important role. ... ICANN should make an informed judgment, based on technical and public policy factors, on how fast to expand the domain name space. But their decisions must be transparent, so as to avoid any suspicion that stakeholders are manipulating the auction process so as to maintain scarcity or protect incumbents. ... ICANN needs to articulate ... objective criteria for any decision reached on how many gTLDs to add.

In this respect, "proof-of-concept" is a prudent policy. We think ICANN can legitimately limit the first round of gTLD auctions to a technically and administratively manageable number. ... There were 44 serious applicants for new gTLDs in December, 2000, each anteing up the $50,000 application fee. Several more applicants have emerged since then. We believe this number -- roughly 50 -- provides a suitable lodestar figure for ICANN to consider.

There's lots more good and interesting stuff where that came from. It's a long article, but well worth reading.

One of the most important features of the architecture of the Internet is the Domain Name System (DNS), which is administered by the Internet Corporation for Assigned Names and Numbers (ICANN). Logically, the DNS is organized into Top Level Domains (such as .com), Second Level Domains (such as amazon.com), and third, fourth, and higher level domains (such as www.amazon.com). The physical infrastructure of the DNS consists of name servers, including the Root Server System, which provides the information that directs name queries for each Top Level Domain to the appropriate server. ICANN is responsible for the allocation of the root and the creation or reallocation of Top Level Domains.

The Root Server System and associated name space are scarce resources in the economic sense. The root servers have a finite capacity and expansion of the system is costly. The name space is scarce, because each string (or set of characters) can only be allocated to one Registry (or operator of a Top Level Domain). In addition, name service is not a public good in the economic sense, because it is possible to exclude strings from the DNS and because the allocation of a string to one firm results in the inability of other firms to use that name string. From the economic perspective, therefore, the question arises: what is the most efficient method for allocating the root resource?

There are only five basic options available for allocation of the root. (1) a static root, equivalent to a decision to waste the currently unallocated capacity; (2) public interest hearings (or "beauty contests"); (3) lotteries; (4) a queuing mechanism; or (5) an auction. The fundamental economic question about the Domain Name System is which of these provides the most efficient mechanism for allocating the root resource?

This resource allocation problem is analogous to problems raised in the telecommunications sector, where the Federal Communications Commission has a long history of attempting to allocate broadcast spectrum and the telephone number space. This experience reveals that a case-by-case allocation on the basis of ad hoc judgments about the public interest is doomed to failure, and that auctions (as opposed to lotteries or queues) provide the best mechanism for insuring that such public-trust resources find their highest and best use. Based on the telecommunications experience, the best method for ICANN to allocate new Top Level Domains would be to conduct an auction. Many auction designs are possible. One proposal is to auction a fixed number of new Top Level Domain slots each year. This proposal would both expand the root resource at a reasonable pace and insure that the slots went to their highest and best use. Public interest Top Level Domains could be allocated by another mechanism such as a lottery and their costs to ICANN could be subsidized by the proceeds of the auction.