Founded in 1998 as an international, nonprofit, private-public partnership, ICANN manages the internet’s Domain Name System (DNS), including both generic (gTLD) and country code (ccTLD) Top-Level Domain names and the numeric address infrastructure. The original seven gTLDs (.com, .edu, .gov, .int, .mil, .net, .org) were followed by seven more introduced in 2001 and 2002, including four unsponsored (.biz, .info, .name, .pro) and three sponsored (.aero, .coop, .museum). Sponsored sTLDs are limited to narrow communities for which the sponsor sets the rules generally. Since then several more were added (.asia, .cat, .jobs, .mobi, .travel). Each TLD has its own registry, a master database of all the domain names registered in each, maintained by a registry operator. Internet users set up their domain names through registrar companies which, in turn, feed the technical information to the registry.

Years of research, debate, and development and an estimated $10 million went into the changes now being introduced by ICANN. Most of the effort leading to the approved changes came from the Generic Names Supporting Organization (GNSO; http://gnso.icann.org) and its Aug. 8, 2007, report, "Introduction of New Generic Top-Level Domains" (http://gnso.icann.org/issues/new-gtlds/pdp-dec05-fr-parta-08aug07.htm#_Toc43798015). The new process will let users self-select domain names appropriate to their communities or markets, such as brand names or areas of interest. Already, three cities—Paris, Berlin, and New York—have indicated a desire to establish city TLDs, each in conjunction with groups of companies.

The main change, according to Jason Keenan, ICANN’s media advisor, is that the process of domain name creation at ICANN has become "permanent. The previous two rounds gave us a lot of information. We did a lot of work between 2003 and 2009. It isn’t like product development and marketing for a public company. Our community has to know what should happen. We have to share our thoughts, with wide-ranging discussion of issues and ideas. Now we have a more solid-state process."

Once begun, the first application round will involve a limited application period. Keenan estimated that the window of opportunity would involve only a number of weeks or months. Any established entity from anywhere in the world can apply. The first filter in the evaluation process is sizeable, however. Applicants must submit a nonrefundable application fee, which Keenan estimated to fall between $100,000 and $500,000. "The fee has yet to be determined," says Keenan. "It is simply designed on a cost recovery basis only, so that ICANN covers the cost of the processing and the approval process. We’re not making a profit."

I asked if not-for-profits or poorer organizations could get any discount. (See my Searcher editorial for February 2001, "Dot-Lib," and Wallace Koehler’s article, "Dot-Lib for Libraries—Can It Happen? Ask ICANN," in the April 2001 issue.) Keenan responded, "As the process stands, there is no discount in application fees. However, there is a nod to communities in the string contention element of the evaluation process. Applicants determined to be community applicants will be given the option to select comparative evaluation, as opposed to auction, to avoid entering a bidding war with applicants who may have more resources."

The final implementation plan will have to address a number of challenges and problems. Trademark protection springs to mind. According to the current announcement, "Trademarks will not be automatically reserved. But there will be an objection-based mechanism for trademark owners." Details of how this will work still have to be set, but Keenan indicated that the process would probably put the burden of identifying a problem on the trademark owner. ICANN would assist to some extent through open documentation. Keenan says, "After all of the applications have been received, a comprehensive list of all applications and applicants, along with other limited data will be released, likely through our interface (called TLD Application System (TAS) right now). However, the posting of the applications will be accompanied by a communications campaign that will target the people who have been contacted on a regular basis by ICANN’s global partnership’s team. The campaign will aim for as wide a distribution as possible, rather than specific to a certain trademark owner or community."

The other major issue affecting domain names that arose at the Paris ICANN meeting involved Internationalized Domain Names (IDNs), i.e., domain names available in character sets other than ASCII. The conference approved opening a proposal to support and fast track IDNs for public comment and expects a report at the next annual meeting in November. This would mean expedited approval processes for TLDs written in non-Latin scripts, such as Asian and Arabic. If approved the fast track procedure would be put in place and a long-term policy developed. Peter Dengate Thrush, ICANN chair, stated, "This is going to be very important for the future of the Internet in Asia, the Middle East, Eastern Europe and Russia." Keenan indicated that they planned to have the new IDNs procedures in place when the next application round opened. "Technical testing is still going on. We’re using Unicode but not all its characters are available, some for technical reasons, some because they’re technically unstable. But we will post a list when they become available."

But Will It Work?

Who would know better than the internet pioneer and former chair of ICANN? Cerf, vice president and CIE (chief internet evangelist) of Google, retired as chair of ICANN in 2007, however, he states, "pressure for expansion of the TLD space was present during most of my tenure as chairman." Cerf does expect a challenging process lies ahead of ICANN. He provided the following comments:

The consequences of an essentially open ‘call’ for new TLDs, including internationalized domain names for both generic and country-code TLDs, are in some ways not easily guessed. One could imagine the re-appearance of ‘.xxx,’ for example, or ‘.web’ and their histories may have to be factored into their treatment. Conflicting but equally valid trademarks, famous marks, socially controversial proposals, fairness across a spectrum of proposers, etc.

However, it seems rather important to point out that the creation of TLD implies at least that the proposer is committing to operation of the TLD, dealing with registrants, expectation of a sound and sustainable business model, coping with complaints and even litigation, maintaining high reliability resolution, dealing with fraudulent attempts to take over domains, all the things you read about in connection with the current list of TLDs will have to be dealt with by the operators of these new ones. Proliferation of TLDs will also add work for ICANN in oversight and compliance, so any expansion model also has to deal with costs to ICANN. The ability to keep up with the oversight work, contractual negotiations, and so on also put some constraints on the scale at which ICANN with its present processes can function.

Extending the TLD space will also pose challenges to any piece of software that does not have an automatic way to discover new, valid TLDs. We have seen problems with the introduction of domain names longer than three Latin characters for example. Things get even more complex as we introduce TLDs in UNICODE using non-Latin scripts that are also encoded in punycode format. The different ways in which the same TLD may appear add complexity to processing.

It is an open question whether these new TLDs will prove popular. Trademark owners may feel compelled to register in all of them (at an unknown cost since I am not sure whether the pricing for registration in these possible new TLDs will be uniform or open to arbitrary choice by the operators). I think this is all very hard to predict. The secondary market in domain names, domaining to generate revenues from advertising, reselling and many other practices all emerged as domain names acquired monetary value well beyond their early manifestation as simple mile markers in the internet naming space.

It seems likely that many of these issues and more have been debated within the Board, in ICANN meetings, in various of the supporting organizations associated with ICANN and elsewhere. So I am not likely raising anything new. As to browsers, they will all have to be prepared to recognize and process domain names using these new TLDs. To the extent that it is helpful to recognize which are valid TLDs and which are not (to avoid trying to crawl invalid ones), one needs an easily obtained list of valid TLDs that can be automatically refreshed as new ones are authorized.

This is not an attempt [by ICANN] to raise money but to respond to repeated claims that competition demands that more TLDs be permitted. I am not personally convinced of the utility and fear many of the side-effects I suggested in earlier comments, but I do understand the likely feeling of the board that some response to this demand is needed. Balancing all the very diverse views on this question is a very hard proposition.

If domains proliferate, multinational enterprises could find themselves registering their identities over and over. Wallace Koehler, director/professor in the MLIS Program at Valdosta State University, observed, "The intellectual property people at major corporations must be crawling the walls. They could end up having to buy hundreds of domain names."

ICANN already puts registry operators under formal agreements. Keenan explained, "gTLD operators will be required to enter into an agreement with ICANN and will be expected to abide by the terms set forth in that agreement, just as existing gTLD registries must operate. As for compliance, a specific schedule or set of criteria to test on has not been determined at this stage."

One problem the new domains will not have, according to Keenan, Koehler, and Cerf—Google will find them. Koehler thought that some other web search engines might have some problems, e.g., Yahoo!’s chain indexing system. However, Cerf considered Google should have no difficulty. He did say, "If there are a large number of pages in which a particular search term is cited and if most references to those pages are to domain names not associated with a new TLD, it is possible that the pages with the new TLD will not rank as highly as others. However, one can tailor Google searches in a variety of ways to focus on particular TLDs among many other focusing features. For example you can say in the Google search bar: tahiti site:.travel to search only web sites ending in .travel." (See www.google.com/support/bin/static.py?page=searchguides.html&ctx=advanced&hl=en for some detailed information about tailored Google searches.)

The introduction of new gTLDs in the past does not seem to have changed the world or its web significantly. ICANN posts regular statistics on usage from all its registry operators (http://icann.org/tlds/monthly-reports). You can download PDF files from each operator by month. VeriSign, registry operator for the ancient dot-com and dot-net, measures its domain names in tens of millions; its nearest competitor, NeuStar with dot-biz, measures fewer than 2 million.

On the other hand, the expansion of domains means an expansion of registry operators. Things could take interesting turns. For example, who would have expected the third least populated country in the world and fourth smallest country in land size (the island nation of Tuvalu) to have a vigorous ccTLD? But Ted Turner’s TNT superstation promotes its vigorous domain service every time it shows its station screen credit—TNT@TV. If cities start proclaiming their domains on every form and on every city bus … who knows what the future will hold? Dr. Paul Twomey, president and CEO of ICANN, saluted the change: "The potential here is huge. It represents a whole new way for people to express themselves on the Net. It’s a massive increase in the ‘real estate’ of the internet." Some observers predict a land rush.