User Confusion, Scarcity Among Issues Raised by One-Character Internationalized TLDs

The draft JIG report sets out several policy issues related to single-character IDN TLDs. One is possible confusion with single-character ASCII (Latin) TLDs, which ICANN reserved for future expansion of the name space. While there are clear differences between ASCII and IDN TLDs, introducing single-character IDN TLDs could conflict with the reserved names, the group said. The potential problem could be handled by rejecting a single-character IDN that’s visually similar to any possible one-character ASCII TLD, by allowing only certain scripts to apply for single-character IDN TLDs, or by a combination of both, the JIG said. A policy could specify, for example, that single-character TLDs based on Latin, Greek and Cyrillic scripts are intrinsically confusable with ASCII and won’t be approved without a special exception, it said.

Another question is whether single-character IDN generic TLDs should be considered “premium real estate” because shorter domains are more desirable, and subjected to higher application fees or special resolution procedures for multiple claims, the JIG said. A third issue is whether, due to the relatively smaller pool of possible names, special allocation methods are needed. Under that reasoning, setting different policies for single-character TLDs would mean doing the same for two-character, three-character and other TLDs because each pool size is different, another flawed idea, it said.

A fourth issue is whether a one-character IDN TLD will make it easier for users to mistype addresses. User confusion has already been addressed by policies covering IDN gTLDs and IDN country-code TLDs, and the argument that one-character IDN TLDs will create more havoc is wrong, the group said. It could equally be argued that the possibility of user error is higher for multi-character IDN TLDs, and the assumption that the smaller number of possible single-character IDN Tlds will lead to more typos is “purely speculative,” it said.

Another question is how to distinguish between single-character generic (g) and country-code (cc) TLDs, the JIG said. Before TLDs in non-Roman script were introduced, country-codes were distinguishable by having two letters (.us, .uk), generic three (.com, .gov), it said. The distinction may, however, be irrelevant now because limiting internationalized ccTLDs to two letters may no longer be possible, it said. A final issue is whether new policies are needed to make single-character IDN TLDs usable, given the fact that some applications and databases, such as spam filters and registration systems, don’t recognize TLDs of less than two characters, the JIG said.

Asian Countries First Up?

One-character TLDs are very different across languages, said James Seng, an IDN liaison to the ICANN At-Large Advisory Committee. In Chinese, particularly, where every single character is a “word” with a meaning, “forbidding single TLDs means rejecting nearly 100,000 words,” he told us. Most of the arguments against single-character TLDs don’t apply to Chinese, he said.

In a May ALAC statement on IDN issues, Seng urged ICANN to “act global, be local” in its handling of IDN issues. Internationalized domains matter to every Internet user because they allow names in their native languages, he said. ICANN should “be more sensitive [to] the needs of different language groups” and be willing to adopt different policies to address them, he said. ALAC wants further studies on one-character IDN TLDs, he said.

"I think you will definitely get [China/Japan/Korea] single-character IDN TLDs, both country-code and generic,” said Avri Doria, an ICANN participant and adjunct professor at Lulea University of Technology in Sweden. Some think this is enough -- except for .A., .E and .I -- since these languages are the only ones with single-character words, she said. Latin and ASCII IDNs will “keep dangling for a long while, except for the few that are grand-fathered in and which serve as proof of there being no technical problems,” she said.

Scarcity is a key issue for Latin single-character TLDs, said Christopher Ambler, who described himself as a long-time domain pioneer. There are only 26 single Latin characters, he said. “How will they be given out? Auction? Lottery? Consulting the Flying Spaghetti Monster for divine delegation?"

But former ICANN director Karl Auerbach said single-character, as opposed to single-word, TLDs shouldn’t be allowed. ICANN “has shown itself unable to deal with the issues of creating more TLDs in general,” he said. “The land rush for the 26 single ascii character ones would stress ICANN even more,” said Auerbach, now chief technology officer at InterWorking Labs.

Besides, the domain name system itself will likely fade away, and ICANN’s importance with it, Auerbach said. The DNS won’t disappear entirely but will be submerged under newer naming and addressing regimes, he said. Domain names will become “internal tokens,” not unlike Internet Protocol addresses now, he said. The trend has already begun, as people increasingly turn to speed dial directories, search engines, and social site names in lieu of domains, he said. Cloud computing, where resources split and fuse, will require name and attribute-based lookup systems that are far more dynamic and flexible than the DNS, he said. Comments on the JIG draft are due Sept. 9 -- jig-initial-report@icann.org.