That said, a big country implementing a new DNS root on a national scale
may not have that problem. The telecom world is already full of systems
that don't cross national borders. In the US case, think of all the cell
phones that have international dialing turned off by default, and all the 800 numbers whose owners probably aren't at all bothered by their inability to receive calls from other countries.
The fact is that most Chinese people want to access the same Internet resources as most Americans. Namely, those resources that exist in their own country in their own language. So if someone offers a root zone that contains everything in the ICANN zone plus additional zones that give access to resources for a specific language group, i.e. Chinese-speakers, then it doesn't seem farfetched for all Chinese-speaking countries to use that extended root zone. And it also does not seem farfetched for American ISPs who market access services to the Chinese speaking community in the USA to also use that extended root zone.
A system that would limit my ability to talk to people in other countries doesn't sound very appealing to me.
Every public root experiment that I have seen has always operated as a superset of the ICANN root zone. In the past they often have not had good ways to deal with TLD collision but this may well have changed. Certainly, the xn-- TLDs seem rather unlikely to collide with ICANN TLDs. I think that the marketing people are going to win this one. There is no marketable benefit to the ICANN root zone but there are clear advantages for countries using non-Latin alphabets to switch to a root zone that allows for their own language to be used in domain names. Turkey was recently mentioned and that is also a country that uses a non-Latin alphabet. --Michael Dillon