Additional Unwanted New TLDs threatens Balkanization of the Domain Name System

[George Kirikos <gkirikos@xxxxxxxxx>]

ICANN should think very hard before adding any new TLDs. Indeed, the
status quo is perfectly fine for most. There is no pressing need for
any new TLDs whatsoever (except perhaps for IDN TLDs that alias
existing TLDs, and which resolve to existing domain name registrants'
names in bundles, like DNAME, etc.).

The current situation with domain name abusers is akin to the arcade
game "Whac-A-Mole"

http://en.wikipedia.org/wiki/Whac-A-Mole
http://whacamole.com/

However, this "game" is more insidious, because just when you think
you've won, the arcade owner adds a few moles that you need to whack
(new TLDs), or speeds up the moles so it's harder to see or catch them
(WHOIS privacy, domain tasting, domain kiting, fake WHOIS). Indeed,
these "moles" are no longer made of hard plastic. They're crafty,
adaptive, organized and well-funded -- it's no longer a single
miscreant working alone in their basement.

The next "level" to their game, in my opinion, will be to actually
become new TLD operators (some have already become their own
registrars) and thereby squeeze TM holders and existing registrants of
as much cash as they can, via self-serving startup processes that
maximize their own benefits. Phishing, spam and other abuses will also
grow. If one takes a look at the dot-asia or dot-mobi experiences and
was concerned, think about what will happen when professional abusers
start applying for their own gTLDs by the hundreds or even by the
thousands. The "moles" are now the arcade owners!

We've already seen a glimpse of what happens when the "moles are in
charge" through the proposed .biz/info/org contracts, where the
registry operators tried to introduce .tv-style  differential/tiered
pricing for domains (and might still do so one day, given the amended
language in the final contracts still left the door open, depending on
results of studies by so-called "professional experts"). Even so-called
"public interest" registries, like dot-org, demonstrated that even
"doves" are not to be trusted when they're put in charge, as the
temptations are just too great to do self-serving acts which harm the
public overall, to transform into "moles" or "weasels", even if one
began years ago with the best intentions as "doves."

Thus, ultimately it's in our best interests to remove those temptations
through higher hurdles for abusers, and mechanisms to ensure that new
TLDs do in fact benefit the public overall. 

ICANN risks balkanization of the internet through deployment of
unwanted TLDs, as just because a TLD happens to be in ICANN's root
doesn't compel everyone to resolve it. As mass influxes of new unwanted
TLDs become more and more like unwanted spam, I can foresee browser
makers, search engines, or ISPs making differentiations between
"classic" TLDs like .com, and these new TLDs (just like network tv
stations like ABC, NBC, and CBS, on "basic" cable, versus specialty
cable channels where availability depend on which cable system you
subscribe to). If ICANN is going to create new problems by allowing new
unwanted TLDs, the market will end up countering it effectively using
the same mechanisms that have been seen in the anti-spam war (and
anti-phishing war). The internet was designed to be resilient, to
"route around damage." When SiteFinder existed, many ISPs reacted by
demanding patched versions of the Bind DNS software to "route around"
that "damage" (and indeed that might have been a compelling reason for
ICANN to move to stop the abuse by VeriSign):

http://www.wired.com/science/discoveries/news/2003/09/60473

What happens when ICANN itself is "damaged", no longer a thoughtful and
trusted custodian or steward, or that new TLDs are "damaged" from the
get-go? ISPs can react by refusing to resolve a new fangled .KOM or
.NEW or .CASH or .BLACKMAIL or whatever abusive new TLD applicants come
out with to squeeze the rest of us. Search engines can simply fail to
crawl those new TLDs entirely, so that they don't appear in search
results. Browser makers too can disable those new TLDs by default, and
require changes from standard settings to re-enable them. Many mail
servers already routinely "blackhole" all mail from certain ccTLDs, so
blackholing entire new TLDs is not without precedent. Perhaps new TLDs
will need to pay fees to others (ISPs, browser makers, search engines,
etc.) for inclusion, a "pay to play" model, just like cable television,
etc. Perhaps TM owners will escalate the game to block entire TLDs via
ISPs and class-action procedures.

This risk of balkanization should be in the mind of policy makers at
all times, as I think at heart we all want good domains to be
universally resolvable, if they are adding to the internet. However, no
doubt should exist that if ICANN start introducing and approving
unwanted TLDs, market forces will potentially counter that harshly by
treating them like spam and phishing is currently treated, namely to
blackhole them. For every unwanted action, there is inevitably a
reaction. 

Potential new gTLD applicants (and registrants) need to be aware of
this risk of being sold second-class goods that could be orphaned,
unless we have very high standards and "raise the bar" on what will
make an acceptable new TLD. The natural balkanization process will act
to bring the new TLD origination process into a steady-state
equilibrium eventually, by raising the entry costs above what ICANN
would otherwise charge for entry, by pushing back the "externalities"
that would otherwise exist back onto the prospective registry operator
through higher costs of universal resolvability. However, why suffer
that great turmoil by doing the wrong thing now? Being more thoughtful
and careful will ensure that negative effects can be monitored over
time, gradually. Throwing the door wide open in a laissez-faire manner
will create chaos, and increase the abuse that is already happening.

Businesses, consumers, and users of the internet are facing
unprecedented cybersquatting and other abuses:

http://www.theregister.co.uk/2007/08/29/biggest_companies_typosquatted/

and until those problems are solved, new TLDs need to be added very
slowly, if at all, before the level of chaos leads directly to
balkanization, and entire TLDs are blocked, with universal
resolvability left as a memory of how the internet "used to be."

ICANN and potential registry operators need to tread carefully, lest
the ISPs, search engines, browser makers and others route around any
abuse through more forceful means, that ultimately cause balkanization.
If part of ICANN's mission is to ensure universal
resolvability of its domains, perhaps now is the time to show better
judgement.

Sincerely,

George Kirikos
http://www.kirikos.com/