[gnso-vi-feb10] Reflections on types and axii of policy development

[Eric Brunner-Williams <ebw@xxxxxxxxxxxxxxxxxxxx>]

Reflecting on some of the proposals it seems useful to distinguish
between those which seek exception from Recommendation 19 for some
type of application, and those that seek exception from Recommendation
19 generally.

Starting from the Board's Resolution #5, which we presume means
"zero", while we wait for Counsel to respond to requests for
clarification, there seem to be two major axis of proposals

1. those which retain Recommendation 19 and advocate some number
greater than "zero" for registry ownership or control, with some
limited exception from Recommendation 19

and

2. those which do not retain Recommendation 19

I suspect that few contributors to this working group are interested
in exploring policy development which does not retain Recommendation
19 for the "standard" application type. It would be nice to poll on
this, but we don't yet have polls.

There are some contributors to this working group who are interested
in exploring policy development which does not retain Recommendation
19 for the "community-based" application type.

Various rationals have been offered for exception to Recommendation
19, though none of the advocates have, so far as I know, offered to
explain why, starting with a marketing budget of exactly $2,000, and
though being profitable from the second month of operation, and having
the genuinely surprising, no, staggering page-per-domain google
statistics(21M for 40k sites, today), .cat has failed the Catalan
linguistic and cultural communities, or could be significantly
improved by exception to Recommendation 19.

Which leaves the possibility that the actual locus of interest in
policy development which does not retain Recommendation 19 is in the
"single registrant" type, which is defined by the absence of its
definition, other than the creative reading of the "Reserved Names"
section (2.6 in specification 5) and "discretionary" language
concerning the reservation or blocking of additional character strings
within the TLD (kudos! to RT for imaginative issue advocacy).

Overlooking the issue of timing, that is, the claim by some advocates
that a policy be developed first for the "single registrant" type, or
that no policy be developed, and all the other undeveloped policy
areas for some "single registrant" type(s), it seems to me to be
useful to discuss scale.

There is the well-known concern that lots of "single registrant" type
applications will fill or flood the evaluation process (Staff has made
that observation, as have others) or the IANA root rate of change
budget (I've made that observation, as have others, and to be fair,
Staff and the Root Server Advisory Committee have asserted that each
can sustain a rate of 3.5 contracts and delegations per day, an order
of magnitude greater than I thought possible, so kudos! to RT again
for being on the more correct side of the scaling issue).

The assumption that there is a potential for a large number of
applications of a "single registrant" type suggests that there are at
least two distinct sub-types of a "single registrant" type.

Those which are quite small in terms of zone file entries, and
essentially no more than any brand manager's marketing requirements,
and those which are quite large in terms of zone file entries,
essentially the accounts receivable of a large retail services
corporation.

It seems unlikely that the pool of potential subscriber-based
applicants, commercial and non-commercial, is as vast as the pool of
potential brand-management applicants, as the former has the predicate
condition or investment, of creating the member or customer
relationships, an activity which usually takes significant investment
in time, care and capital. The latter is a modest cost to a brand
manager promoting a global or regional brand.

For which of these is policy development truly necessary?

There is a model for very few entries in a zone with no delegations

{abuse,marketing,sales}@revlon

There is a model with thousands, even tens of thousands or entries in
a zone with some delegation

{branches-or-affiliates}@{ibm-or-kentucky-fried-chicken}

There is a model with millions of entries in one or more zones with
lots of structure.

{accounts}@{honking-big-isp}

There is a model with hundreds of millions of entries in zones with
lots and lots of structure, and probably damn few people.

{objects}@{we-haven't-opened-that-box-yet-pandora}

As the real cost of registration is measured in pennies, the case for
avoiding the unnecessary cost of registrars seems to only be
compelling where the value of the domain registered is not a lot more
than that. Essentially, we're looking at the issue of reducing the
cost of something inherently worthless. Where the value of a domain is
greater than some pennies, all policy development does is either move
margins from the registrars to the registry, making the registry more
profitable. This will have the effect of making viable some small
number of registries which would be non-viable without this transfer
of margins.

A poll would be useful, from it we could learn whether the no
restrictions positions are primarily an artifact of the widely held
belief that brand managers are about to create a bonanza, or are an
attempt to make mass numbers of marginally desired objects
sufficiently profitable to support incidental usage business models.

Eric