I commend the ICANN President for demonstrating some movement on the issue of new gTLDs. In these brief comments, I focus only on Parts 1 and 3 of the plan. Part 1 calls for immediate steps to add three new sponsored TLDs; Part 3 requests the DNSO to choose between a centrally planned, “taxonomic” approach to name space development and an “open entry” approach that relies on private sector initiative.
Any attempt by ICANN to accommodate real demand by legitimate organizations wishing to utilize the DNS name space is to be welcomed. Thus, I strongly support the recommendation that the Board solicit proposals for new sponsored TLDs. I welcome the President’s recognition that “there is little risk….in moving forward with a limited number of sponsored TLDs even to the point of full operation.”
I part company with the President on his decision to limit that number to three, however. That number seems unnecessarily restrictive and is unsupported by any technical or economic analysis. It seems to be contradicted by the President’s own report, which states “it is hard to find anyone who would argue that tens or perhaps even hundreds of new small to medium sized gTLDs could not be safely added,” as well as the statement later on that “there was nothing magic in the number seven for new gTLDs. It could just as well have been three, fourteen, or twenty.” I agree with the President that the number of additions at this point needs to be finite and, given the politics of the situation, somewhat modest. But 10 new sTLDs would be a far more appropriate number at this stage of the game. The smaller the number the more irrational and potentially unfair the choices that must be made, the greater the potential to make wrong choices, and the more the market is distorted by the exclusion of valid applicants. A modest number such as 10 would allow most existing legitimate sTLD proposals to be developed, eliminating the need for apples-vs.-oranges comparisons and bitter political battles.
The plan notes that the President will ask the DNSO for its opinion on whether a taxonomic or open approach will be used for name space additions. My frank opinion is that a moment’s thought will reveal that engagement with that question is a waste of time. The taxonomic approach is not a real option. Over 40 million legacy registrations, more than half of them in one TLD, make it impossible at this stage of the DNS’s development to “rationalize” the name space by promulgating a classificatory TLD scheme. The name space can only evolve through accretion; i.e., by the gradual addition of new TLDs that meet naming needs not adequately met by existing names. Taxonomies and organized naming structures can and do exist within TLDs (as well as in portals, search engines, private keyword spaces, and digital libraries). But the idea of a global DNS taxonomy that divides the entire Internet up into neat little cubbyholes once and for all is an impossibility.
The pre-commercial Internet cannot be used as a counterexample. The ARPA-Internet’s original naming taxonomy was developed for a tiny, closed network of academic researchers and government agencies. The community involved was composed entirely of English-speaking participants in a common computer culture. If you think that a planned taxonomy is “rational,” let me remind you of the fate of .com in Postel’s original taxonomy. As conditions changed, one TLD – dot com – suddenly found itself with 87% of all registrations in the USA and over 75% of all registrations worldwide, leading to a level of flatness and imbalance in the DNS’s hierarchical structure that we are still recovering from. No planner can anticipate the unanticipated; a taxonomic structure – even assuming that it was possible to force users to fit into the categories the planners create for them – will quickly be obsolete in a complex, changing, multicultural world.
I could write (and in fact am writing) a dissertation-sized paper on the numerous absurdities and problems inherent in the concept of a planned DNS name space. But detailed discussion of those issues is best left to the DNSO itself, should it be unfortunate enough to be given this option. For now, I would simply request that common sense prevail and the President avoid wasting the DNSO’s time on an idea that is not feasible.
The name space will evolve through addition, not through “rationalization.” Therefore I urge the President to reconsider this request.
Dr. Milton Mueller
Associate Professor, Syracuse University School Of Information Studies