Re: [wg-c] New gTLD's

[Eric Brunner <brunner@world.std.com>]

Ian,

The last late-joiner, a 1L from Harvard, offered a lack of domain-specific
knowledge as the basis for objective analysis of the position paper series.
I won't tell you that being a practitioner doesn't provide domain-specific
knowledge, only that the domain is only one of several that are required.

You milage may vary in the problem statements, degrees of difficulty, 
or choice of sodas in what follows.

The hardest bits in the WG-C problem set are not primarily policy, but
investment-blind mechanism.

1a.	How does data propagate from registrant to registry, and is it
	replicated, and if so, how is consistency obtained, and who has
	subsequent read-access to the registry, and how is registry
	recovery accomplished?

The question posed as read/write access issues can be expressed as in the
legal jargon just as easily as in the computer science jargon. One not so
helpful consequence of this ease of re-expression is keeping consistency
in two or more jargons.

1b.	Is registry data property? If so, whos? Who may have access
	to it and how is that access mediated by its holder? How is
	it transfered and recovered in the event of failure?
 
The next hardest bits in the WG-C problem set are technology-blind
investment mechanism.

2.	What is the actual cost to create and service a registry record,
	how does cost-recovery affect the capitalization of a registry
	operator, and what other direct and derived valorizations are
	available to the operator and its economic partner entities.

3.	Are registry operators and registry registrars distinct legal
	entities? 

4.	What is the legal structure for registry operators?

The not-very hard problems are (your numbering):

1.	How many?

Add incrementally until a constraint is reached. Constraints range from
ICANN Constituency defined limitations to ICANN DNS Root Server System
Advisory Committee defined limitations.

2.	How fast?

See the answer to #5, above. Modify for arbitrary delay upon direct or
indirect demand by NSI. The modification is sufficiently compelling that
in fact no one knows what the skeleton of a new ICANN-operator agreement
looks like. As someone remarked recently, "no one has even seen the
whiskers of that rat yet".

3.	New registries?

It is a given that the set of new registry delegations will not initially
be made soley to NSI. It may get there, or to something equivalent, but not
ab initio.

4.	Which names? 

No one much cares beyond the original IAHC set. The English phonebook is
a candidate, some think it intuitive. Another school is let the operator
create a brand (semantic content), using semantic-free strings. After all,
.COM could have just as well ment where the commies hang their DNS hats.
Try and keep search engine coverage in mind, Altavista is running under
20% on "the web", and a phone book most of whose pages are blank, and the
rest stale, is not the first reference tool to reach for.

Would you mind answering a few questions? Little ones about abstract
clients and price points for non-litigated and litigated conduct arising
from the DNS playpen, as well as abstract payments to abstract inexistant
new domains. Thanks in advance.

Cheers,
Eric