[At-Large] [Fwd: Comments on the proposed recommendations for the introduction of new gTLDs]

[Vittorio Bertola]

FYI - These are the comments that I just submitted, speaking personally, 
to the GNSO report on the introduction of new gTLDs.

-------- Messaggio Originale  --------
Oggetto: Comments on the proposed recommendations for the introduction 
of new gTLDs
Data: Thu, 30 Aug 2007 18:01:06 +0200
Da: Vittorio Bertola <vb at bertola.eu>
A: gtldfinalreport-2007 at icann.org

The comments that follow represent my own personal view, as an
individual who has been active in ICANN's At Large constituency for the
last several years. They do not represent the ALAC's or anyone else's view.

On principles:

Principles A, C: Both principles clearly frame the issue of the
introduction of new gTLDs in pure business terms: the introduction must
be "timely" and "predictable", its reasons are because there is demand,
and its objectives should be to promote competition and differentiation.
While all these principles can be supported, there is a striking lack of
appreciation for the major cultural implications that in many cases can
be associated with the introduction of a new TLD. For example, the main
aim of the ".cat" proposal was to promote and defend the Catalan
language and culture, something that has nothing to do with making
business out of selling domain names. When allocating the TLD space,
multiple purposes - be them economical, cultural, social... - need to be
accommodated and treated evenly; the allocation system should not
promote some over others. So, these principle should be revised: "fair",
"transparent" and "accountable" should be other characteristics listed
in principle A; the second sentence in principle C should be removed, or
completed with other possible purposes, or (better) replaced by a
recommendation of purpose neutrality for the allocation mechanism,
including non-discriminatory allowance for innovative uses of the DNS.

Principles D, E, F and recommendations 7, 8: The language in itself is
agreeable in line of principle, but is extremely vague; in particular,
technical and operational requirements should not be set so high to
prevent opportunities for access to almost anyone but established
registry operators. They should rather be proportional to the intended
scale and target community of the TLD; this principle of proportionality
should be recognized in the recommendations.

Principle G: This principle is entirely laudable in spirit. However,
"the applicant's freedom of expression rights" might have to be
confronted with other parties' rights of various kinds; a relative
statement of support for free expression as long as it does not hamper
other people's freedoms, rather than an absolute pledge of support for
one right by one of the involved parties, might be more balanced. (See
also the comment on recommendation 9.)

On recommendations (I agree with the ones that I do not comment upon):

Recommendations 7, 8: See above.

Recommendation 9: It is, in my opinion, utopian to think that everything
can be evaluated in a measurable and objective manner: strings have
intrinsic semantic values and semantics are by definition subjective.
Perhaps there should also be recognition that, in the end, not
everything can be measured objectively, and in controversial cases there
might be the need for a case-by-case call. At the same time, it must be
clear that ICANN's role is not to make subjective judgements itself,
especially when they fall into the field of society and politics, but
rather to evaluate what are the different subjective judgements of the
various stakeholders; ICANN should try to promote consensus, and if
consensus cannot be reached, try to evaluate what is the course of
action that could be acceptable for the broadest possible number of
stakeholders.

Recommendation 10: It should be recognized that base contracts might
differ according to the legislation to which the prospective registry is
subject.

Recommendation 19: I disagree with the recommendation. I agree that, if
the registry adopts the registry-registrar model, then the registrars
should be the ICANN-accredited ones. I also agree that the
registry-registrar model should be mandated for commercially oriented
TLDs acting on a global scale. However, there is the need to provide a
certain degree of freedom and to leave applicants the possibility to
propose different registration models, if they think that such different
models better suit the scale, purpose and target community of their
proposed TLD. For example, a TLD might be aimed at members of a given
organization, which might then handle the registry and the registration
process directly, or even give domain names for free to all members. Or,
a TLD might be aimed to a small and geographically concentrated
community, for which interaction with all ICANN accredited registrars
would be overkill, or for which local operators, who are not ICANN
accredited registrars nor have a reason to be so, but have strong ties
with the local community, are better suited to sell the domains. There
must be the chance for the applicant to propose different models and to
get them evaluated according to the general characteristics of the TLD
proposal.

Recommendation 20: The principle is agreeable in general, but the
concept of "implicitly targeted" community is dangerous. In particular,
TLD proposals for strings that have a specific meaning for their target
community should not necessarily be rejected because other groups see
other meanings in those strings, as long as the registration
requirements restrict access to the TLD only to the "explicitly
targeted" community. In general, it is necessary to keep some room for
case-by-case evaluation of semantic conflicts.

On implementation guidelines:

Implementation guideline E: I am not sure whether "application
submission date" means deadline, or opening date. As some matters might
be prioritized on a first-come, first-serve basis, a term of four months
should be ensured before the submission period starts, not just before
it ends - i.e., people who actually need the four months to compile the
application should be able not to come last.

Implementation guidelines M, N, O: See below.

Missing issues:

A basic consideration that seems to be missing is that the TLD space,
while being much bigger than often suggested, is not infinite in size;
each new TLD prevents future proposals for the same string, as well for
similar strings (as per Recommendation 2), up to the point where many or
most reasonable strings could be unavailable. There should be a
principle stating that the liberal introduction of new TLDs (that I
support) should be tempered with the need to preserve a sufficient
availability of strings for future applications; the TLD space should be
allocated with generosity, but also with care.

As a consequence, the recommendations do not seem to address the issue
of whether minimum requirements in terms of scale and relevance should
be posed on the target communities. Would a TLD corresponding to a
company name, and reserved to itself, be acceptable? Or to a family
name? Or to a small village? Or to a city? Without any criteria, there
is a serious chance the the TLD space would be cluttered by vanity
registrations from the wealthy, which might prevent more collectively
useful applications in the future. There is the need to develop some
recommendations to this regard.

In general, while there are provisions for objection by other parties
that deem that the TLD proposal goes against their rights or interests,
there is no clear provision for an evaluation of the global public
interest, to determine if the introduction of the string could be
harmful to no one in particular, but to the Internet in general.

Also, the recommendations do not seem to address sufficiently the issue
of how to ensure fair access to all potential TLD proposers, in
particular to those with less developed technical skills, or less easy
availability of risk capital, or less familiarity with English and with
American legal customs. This is made worse by the fact that first movers
have a clear advantage, so establishing facilitation mechanisms later
might be equivalent to not establishing them at all. Thus,
implementation guidelines M, N, O should be turned from "may" to "must".
Also, further guidelines should be introduced for fairness:
- The application fee should be, at least in part, proportional to the
intended size of the target community and of the TLD (with additional
fees in case of unexpected success). Also, payment facilitations could
be introduced.
- Applications submitted by non-profit entities for social and cultural
purposes (with clear constraints to ensure that this is the case) should
be facilitated with lower fees.
- Applicants should pay a first, lower fee to get an assessment of
whether the proposed string can be accepted, and a second fee, to cover
the costs for negotiation and implementation, only if the assessment is
successful (otherwise, rejected applicants will be paying a share of the
negotiation costs for successful applicants!).

Regards,
-- 
vb.                   Vittorio Bertola - vb [a] bertola.eu   <--------
-------->  finally with a new website at http://bertola.eu/  <--------