Re: [gnso-vi-feb10] VI in practice, Rights Protection Mechanisms, Registry Agreements and Variants of TLDs

[Eric Brunner-Williams <ebw@xxxxxxxxxxxxxxxxxxxx>]

Just in passing, one classification scheme mentioned in a related
context is the size of the applicant and holder pools.

For the pre-ICANN gTLDs we have around 10.

For the post-ICANN gTLDs we have anther around 10.

For the pre-ICANN ccTLDs we have around 200.

For cities with a population greater than 1M we have some low multiple
of the aggregate of everything currently in the root.

For linguistic and cultural (sub-national and trans-national) with a
population likely to achieve 10k registrations we have some low
multiple of the aggregate of everything currently in the root.

For IDN equivalencies to cities and linguistic national, sub-national,
and trans-national with the union of the demographic limits above,
several multiples of the union of the cities and linguistic and
cultural numbers, or the sum of two or more low multiple of the
aggregate of everything currently in the root.

When we get to SR instance estimation things get a bit odd. We have
the capitalization required rather than a demographic minima, which is
less than $1M, perhaps only a quarter of that. Unfortunately, the
universe of marks is simply not bounded, as new marks are created, and
expire, much faster than cities or linguistic and cultural identities.

In sum, at some point we will have served all the cities, all the
populations not served by the 3166 model, in Latin, Arabic, CJK,
Hindi, ... At no point will we ever be able to say we've run out of
marks. A root many multiples of its current size excites little
concern, other than the rate of change concern. A growth that is
unbounded does excite some concern.

So while you may be right that only a few idiots may run with scissors
while operating a SR, it is an assumption that the number of SR
applicants will ever be below any workload threshold for ICANN's
application evaluation resource, and that the SR model and ICANN's
years long application planning process assuming disjoint,
independent, registrant populations and generally available services
are not fundamentally challenged by problems of scale.

I've cc'd Bertrand as the taxa-by-scale goes to the discussion of
categories.

Eric