I disagree with your assertion that multiple DNS roots would necessarily mean “that identifiers would no longer be universal.”

First of all - DNS names are not universal - they never have been.  They lack several forms of stability - most notably they lack temporal stability, i.e. their meaning changes over time (often very quickly when a name registration lapses and is picked up by another.)  But they also lack geographic stability and client stability (i.e. a name may mean different things depending where a DNS query is issued or who issues it.)

Second, users, when given the power to chose (a power that may be delegated to the user’s ISP), will tend to shun things that surprise them, most particularly, any root that offers top level domains that are significantly conflicted.

We allow anyone to start building automobiles - but woe will be the fate of anyone who tries to sell their vehicles under the name of “Honda” or “Ford” - not only would there be lawsuits on the basis of trademark and misrepresentation, but consumers would simply chose, at internet speed, not to buy these things.

In other words, there are strong forces of self-interest that work to heal name conflicts.

The fear you express sounds much like the fear of a person who wants a planned economy (a la the USSR in 1950) versus an open system in which anyone is free to compete in the marketplace (and perhaps loose their shirt in the process), subject only to reasonable regulations (such as health and safety.)

The further argument that a “split” root would necessarily lead to a “disaster” is an argument that is not only incorrect but also an argument that if believed could lead to the conclusion that the internet should eschew languages other than English and scripts other than ASCII.

Karl: Thanks for your comment. The two examples you provide of temporal changes and changes via geography/client are both under control of the owner of the domain name. To that extent, there’s no problem under the existing single root (i.e. the single root is pointing ultimately to the nameservers of the domain registrant, who can then return any results they desire). 1-800 telephone numbers allow those types of temporal changes and changes via caller origin too, so there’s nothing “new” per se—the single owner of the 1-800 number has control over the destination of the call.

However, when some third party starts injecting their own results, or there are multiple claimants to the same (supposedly unique) identifier, then that’s when all hell breaks loose, diminishing the value of the DNS. Back to the telephone example, if there were multiple telephone “roots”, 1-800-FLOWERS could have its calling results hijacked (by a competing flower company, etc.) and that would diminish the value to the true owner of the identifier. Email would get misdirected, web pages would get misdirected, etc. Imagine what the world would be like if there were multiple US Postal Services, and “100 Main Street, New York, NY” meant something different depending on which of the postal services you used!

The example you give of “trademarks” being a solution to the problem is hardly a good one. First, trademark enforcement is extremely expensive. Solutions that are very expensive are hardly desirable. Secondly, of the 200 million or so domains that are registered, only a small fraction are associated with registered trademarks, and in many/most cases, it would be impossible to secure registered trademarks (because the domains are descriptive, generic, etc.).

I am both a techie and an attorney.  With my latter hat on your lack of faith in the legal system makes me uncomfortable.

My discomfort is elevated as a result of my participation over the years in several internet related start-ups - I react strongly when I perceive arbitrary, incumbent protective, or contrived barriers to invention and innovation.

To my ear what you are suggesting sounds like a proposition to abolish our system of free enterprise and open competition and replacing it with something reminiscent of the central planning of the old Soviet Union.

There is no doubt that our legal systems are expensive and slow - which are well known problems (and hardly new ones either - Dickens in “Bleak House” in the 1840’s made a similar complaint.)

We (as a society) do need to remediate those problems, not just in the context of the domain name disputes, but for all other classes of disputes that might be recognized under civil laws.  You are quite right in recognizing that in many regards under our current legal procedures justice is only for the wealthy.

In solving that problem ICANN’s UDRP and proposed new TLD process is hardly a model of the goal to which we ought to aspire, and hardly a solution that ought to drive us to abandon free enterprise and open competition.

ICANN’s UDRP is a tool that is available only to those who own trademarks; those who base their rights on other grounds are left without a remedy under the Lex-ICANNia.  And we’ve seen the systemic bias that exists in favor of the trademark owner plaintiff under ICANN’s system.

And ICANN’s new TLD process is very clearly designed not to promote but rather to impede those who wish to compete in the domain name marketplace.  The internet would never have come to be had it had to jump through the kinds of hoops that ICANN is trying to impose on domain name entrepreneurs.

Do we really want to adopt the the medieval guild as the model for innovation (or rather, non-innovation) in the domain name spaces on the internet?

One solution - competing, consistent roots - is rather more in keeping with the kind of open, competitive system, one in which users can innovate at the edges, that was the genius of the end-to-end principle that allowed the internet of today to develop.

The debate ought to be about how we define “consistency” and what mechanisms we create in order to drive consistency.

My answer is to use existing trademark law - and expand it to encompass other foundations for rights in names - in conjunction with expedited, but fair, speedy, and inexpensive, procedures.  That, seems, to me anyway, as far less dangerous than abandoning our free enterprise system on the internet.

(I have not even begun to wonder out-loud about what legal excuses could exist to justify ICANN immunity from the various laws regarding restraint of trade that could apply to the kinds of incumbent protective constraints on marketplace entry, TLD product specification, price setting and splitting, and other similar matters that form part of today’s ICANN regulatory structure.)

We’re going around in circles, because in order to get a “consistent” system, there ultimately needs to be just one authority at the very top, one basic parameter that we all agree upon (you even used the word “we”).

Instead of reinventing the wheel, I point readers to RFC 2826.

One of my favorite maxims is “Locus ab auctoritate est infirmissimus” (“The argument from authority is the weakest.”) from Thomas Aquinas.  I find it amusing to cite a Saint of the Catholic Church, and thus one of its highest authorities, on the proposition that authority ought to be questioned.

One can engage in rather a long philosophical debate about sources of ultimate authority.

But in that debate one will rarely find the kind of claim that ICANN claims - that it represents a subject matter based sovereignty that obtains its authority “just because”.

Such a claim is particularly specious given ICANN’s rejection of the kind of constitutional constraints and popular democratic accountability that we have come to expect underneath a modern sovereignty.

Rather, ICANN is a private forum of industrial actors - with a teensy tiny, and ultimately nearly vacuous, nod to the public - in which prices are established (registry fees, ICANN tithe), product specifications defined (such as 1 to 10 year terms in one year increments), product sales terms agreed upon (UDRP, Whois), and potential vendors admitted or excluded - all the hallmarks of both a guild and a combination (or conspiracy) in restraint of trade.

My claim is that we do not need that kind of guild-like central planning and regulatory body over domain names.

Rather we can use a much more competitive system in which much of the pressure to channel behavior into productive channels occurs through competition than by regulation.

For that to occur we do not need ICANN regulation.  (There are jobs that an ICANN could usefully do, but we are not discussing those here.)

Regarding your citation of the IAB statement - that statement is widely viewed as being little more than an exercise in turf-protection.  It is interesting to compare that statement with the purported techno claims made by AT&T;and the FCC in the old Hush-A-Phone case (1954 I believe).  In the Hush-a-Phone case AT&T;, backed by the FCC, argued that a totally passive plastic hand clamped to a phone mouthpiece would destroy the internet of the era, the public switched telephone network.  The court found that despite the backing of the IAB of the era, the FCC, that those claims were laughable.  The result was the first crack in the AT&T;monopoly, leading to Carterfone and MCI and eventually to the breakup of AT&T;and thus to our modern more open world of communications.  Over the years I’ve seen a lot of pretense claims that this or that technological option would end the world as we know it.  I don’t find the IAB statement to be any more credible than were the claims of AT&T;in Hush-a-Phone.