The DNS was not designed for that.
Well, there's not a lot of difference between DNS as a residential application for a residence area of a territorial jurisdiction with, and without, an iso3166 code point.
Just last night I spoke to a top bureaucrat at the city's IT agency who indicated that they are looking to Verisign, NuStar, and other prospective contractors for policy guidance. "Here's how you sell more names, faster, and cheaper. Sign here."
Writing CORE's .nyc application was a learning experience, if not for DOITT, then for myself, and for CORE.
There's huge pressure to sell names willy-nilly, both city budget shortfalls and 90% of the industry panting for new TLDs and pointing to historic faults as evidence that a thoughtfully developed .nyc TLD is impossible. We don't believe so and it would be good to have the strong support of NARLO and ALAC and NCUC and other entities inclined toward the public interest in matters Internet to add support to our effort.
As this list discussion illustrates, the desire for support from outside of New York may be difficult to fulfill. A similar lack of external interest in a standing problem was brought to my attention in the Summer and early Fall of 2001. I suggested that local utility in the form of solving the standing problem was more important than the absence of external interest, and in November 2001 the standing problem was locally solved. The standing problem was the absence of Han Script labels in the IANA root. The solution was to enter into service a constellation of servers serving the IANA root extended with the presence of Han Script labels. Initially zero internet users were benefited, as is common in the introduction of new infrastructure. Today in excess of 400 million internet users are daily beneficiaries, and after a decade of demonstration, the IANA root now contains an initial set of Han Script labels. Eric