part 2: attempt to summarize some public comments to ``ICANN Economic Framework for analysis of expansion of gTLD names'' report posted in june 2010

[k claffy <kc@xxxxxxxxx>]

ICANN has not yet provided any summary of the public comments
of the previous (all) four economic studies of new gTLDs.
Here is a cursory attempt at summarizing some of the less
subtle positions in the public comments on the July 2010
Economic Framework report.

        ( http://forum.icann.org/lists/economic-framework/ )

More careful public summaries of and responses to the comments
of all four reports are needed.  As noted by others in this
thread, the most recent (December 2010) "economic framework"
report fails to address the majority of previous public comments, 
and the parts of the report that do address such comments,
e.g., case studies of previous TLDs and their impact, make an 
even stronger case against additional generic TLDs than the 
previous reports.  

The vast majority of the public comments reflect an abundance
of distrust of ICANN internal process, pointing out veiled
connections with previous specious study by Compass Lexecon,
lack of expertise and relevant (DNS industry) knowledge, clear
affirmations of the need for more data collection, more
scientific study, importance of trademark and consumer protection.
Notably, those in favor of moving forward with gTLDs typically
omitted their affiliations, which might have revealed potential
conflicts of interests.  Those opposed were generally more
transparent about the (typically commercial) interests funding
the comments.

+ AT&T

Finds initial studies of gTLD issues inconclusive. More data
and information is needed. Fully supports the recommendations of
the paper that ICANN gather more comprehensive data about new and
existing gTLDs. Better data collection important not only to
understanding gTLDs but also for understanding malicious conduct
and associated consumer costs. Also, framework paper supports the
need for trademark protections and other safeguards to mitigate
consumer confusion. support implementation plan that introduces new
gTLDs in discrete, limited rounds and prioritizes introduction of
IDNs.

+ Time Warner Inc.

Deeply concerned that ICANN appears poised to move forward with the
launch of new gTLDs despite the fact that none of the "overarching
issues" identified by ICANN in early 2009 have been adequately
addressed in draft applicant guidebook v4.0.

Believe ICANN is indifferent to comments submitted by stakeholders
in this process but do concur with the reports recommendations to
collect more data, consider potential confusion to consumer, addtl
studies to try to lessen delegating new gTLDs that have negative net
social benefits, should introduce new gTLDs slowly.

+ Coalition Against Domain Name Abuse (CADNA)

Clearly unhappy with ICANN process, study author selection. Urges
ICANN to slow process down.

+ Michael D. Palage

Request basis for reconciling the default action mechanism
currently in the DAG with the net social benefit concept set forth
in this report.

What is the basis for 500% increase in the base registry fee in the
new registry agreement?

+ Christopher Wilkinson 

Believes that:

Bad timing business-wise world wide for new TLD registries.

Model for DNS registries should be not-for-profit orgs operating
in public interest.

Should be no secondary market for names. Discontinued, abandoned,
renounced names should return to the Registry and thus the pool.

ICANN should target new TLDs that address new markets, economic and
social niches, under-served languages, local and regional identifiers.

Because of lack of portability, creating new TLDs does not improve
conditions for competition in DNS industry. New TLDs increases scope
for choice in initial registration of a new domain, but should not
be confused with improving conditions of competitiveness between
Registries.

+ Richard Tindal <richardtindal@

Generally supports the paper.  Thinks ICANN wisely included:

1. background requirements and checks on applicants;
2. evaluation questions and contract provisions regarding Security, DNSSEC, and 
Abuse Prevention and Mitigation
3. What is Expected of a Registry Operator' obligations; and
4. legal rights protections such as the string rights objection process, URS, 
Trademark Clearinghouse,
Trademark Launch Claims, Sunrise, PDDRP and Thick Whois.

None of these are required of current gTLDs.

There are however flaws with the concept of 'discrete limited
rounds' that have not been adequately examined:

1. There is no way of knowing if one round will produce results
representative of others.  The assumption is that the measured costs
and benefits of one batch of applications will be indicative of
other batches but this seems both illogical and unlikely.

2. There were discrete limited rounds in 2001 and 2003.  There
is no reliable study that shows trademark or consumer confusion
costs to be higher with those TLDs than with COM, NET or ORG.
Indeed, the data that is available tends to show the opposite.

3. There is no fair or predictable way to decide which applications
should be selected for a limited round. Of the various methods
available to do this almost all favor the well funded, technically
sophisticated or politically connected applicant.  A truly innovative
application, with the most potential consumer benefit, might be
delayed years or never eventuate.

4. Finally, and perhaps most importantly, there will be a natural
phasing of TLDs entering the market as the evaluation, objection,
review, testing, launch and adoption cycles of the impending round
are very likely to be spread over a 20 to 30 month period.    This
is the same reason, in my view, why root scaling issues will be
manageable.

+ Jon Nevett 

Recommends moving ahead with discrete, heavily vetted batches, 
can use what we learn to refine process.

+ Antony Van Couvering (Minds + Machines) 

Believes the entire report could have consisted of this one 
paragraph, which contains the entire wisdom of its contents:

"Because business model innovations are difficult to predict,
experience with the development of gTLDs that serve specific
communities is limited, and the community has no experience with
IDNs at the TLD level, it is difficult to describe the expected
effects of new gTLDs with precision."

i.e., the report is mostly a farce but required by that 
Affirmation of Commitments.

Also notes that the report uses lots of imprecise language --
the words "may" and "might" appear 128 times, or roughly twice
per page. Predicts risks and benefits without quantifying any of them.

Rhe report provides something for everyone...

"But the professionalism of the authors shows through: their most
important recommendation is that the new gTLDs will provide data
for -- wait for it -- another study.

"I commend the authors for taking money from ICANN, and for setting themselves 
up for more work later, and for producing a document that looks entirely 
professional, while saying nothing more than ``it depends.''
They were given an dubious task, and performed it to the hilt."

+ Coalition for Online Accountability ( includes
  ASCAP, BMI, ESA, MPAA, RIAA, SIIA, Time Warner, Walt Disney )

Much concern about the process and specific wording of the questions
used for background checks and vetting of registrants as they relate
to infringement of intellectual property rights. (No suprise there)

Recommends new applicants for gTLD should have to disclose policies 
for WHOIS data quality.

Concerns about community objections and the relative sizing up and
evaluation of competing communities.

Still Need a mechanism by which a registrant, though otherwise compliant
based on baseline standards, may still be denied based on potential
threat or detriment to the community. Example given .kids

ICANN should employ recent best practices from registry agreements
for .asia, .mobi, and .post that require searchable whois, and
require compliance with designation of a PoC, investigation of
claims of false info, speedy cancellation/correction of registrations
found to be false, and allow 3rd parties to invoke these procedures.

COA hopes ICANN will follow Economic Framework and evaluate expected
costs and benefits for various types of gTLDs and focus on "Types"
that offer the greatest promise.

Lots of concerns about not enough meat in current docs regarding 
addressing malicious content.

Clearly still not enough rights management for their taste in the DAG v.4

No reason why ".brand" registrants should not be allowed to have
full vertical control and act as their own registry.
 
+ Paul Tattersfield

http://www.gpmgroup.com

There are still serious questions about the nature of what it
means to be a registry, how competition should be introduced,
whether through fixed term with invitations to tender on each
expiry or whether to award registries in perpetuity to a gTLD.

Observes that comments to the earlier Carlton report seem to be 
ignored in this report, requiring redundant comments.

If new gTLDs are launched without price caps, on the principle they
do not enjoy market power and given existing registries do not all
enjoy the same level market power, smaller existing registries which
consider themselves to have less market power than larger incumbents
are likely to apply to have their price caps lifted.  This has the 
potential to introduce massive externalities for existing innocent 
third parties.

"What if Rupert Murdoch purchases a controlling interest in a company
that gets awarded .news?"

+ Olivier MJ Crepin-Leblond (GIH: global consulting/network services firm)

"This document has raised more questions than answers in my mind."

With no conclusion, neither abstract nor clear overall vision emanating
from the prose, it is hard to draw a clear set of conclusions.
Nonetheless, the conclusions which I sketch from the document are:

* The support by the authors for a proposal of two basic categories of
new gTLDs:
   - gTLDs with private benefits only; and
   - gTLDs which bring social benefit;
which is something that ALAC has been emphasizing for a long time.

* Not that many studies on costs and benefits of new gTLDs have been
published so far.

* The authors recommend mathematical modelling of Costs of increased
registration, monitoring, and enforcement of trademarks across multiple
gTLDs. I differ in viewpoint because believe their proposed model is
over-simplified, uses too many assumptions, and risks either wasting
valuable time or will lead to incomplete and flawed results.

* Case studies based on historical data of how a new gTLD introduction
has affected another gTLD are recommended by the authors. I think that
this is long overdue and I am baffled as to why this has not yet been
undertaken in-house.