Briefly, though, I think his list of objectives misses two important ones:
More controversially, perhaps, I find one of this objectives to be a non-requirement to any serious effort:
The economic model implied by such a goal is neither clear, nor particularly desirable. If I am going to consume resources on other people's servers, some sort of payment seems appropriate. In fact, this could be a criticism leveled at today's DNS infrastructure--- an entry in the .com database costs the same whether it is looked up once per year or a hundred times a second.
It might not be *feasible* to charge for name ownership in a distributed infrastructure. Or the payment might take the form of a quid pro quo arrangement to host other people's entries. But to assume from the outset that names are free seems potentially disastrous.