DNS articles Brad Templeton Home Brad Ideas (My Blog) ClariNet
|
Why generic TLDs are a bad idea
To investigate the question of what generic TLDs are a bad idea, it's best to ask the opposite question -- why do people want them so badly? A generic TLD has an inherent meaning. It has a value outside of that created by the system. The generic TLD "com" -- abbreviation for commercial, holds tremendous value, and almost all the disputes in the domain system arise from this value. Because a generic TLD has inherent value, it is intrinsically better than other TLDs for some purpose. If you have generic TLDs, then all the TLDs are not on a level playing field. ".com" becomes the best place to register a commercial site. ".museum" declares itself clearly the best, even only good place to get a domain for a museum. A generic TLD is a monopoly on a generic term when it comes to internet naming. That this is a bad idea is no surprise to most, and because of it, the popular early fix for this was to go to a "many registrars, one registry" system where no registrar had a monopoly on such registrations, though the registry did. However, this simply solved the problem at one level. If you have a generic TLD like ".museum" it usually means that some party will get to register the generic term "art.museum." In the multiple-registrar system, that is almost always the first party to register. To many, first-come, first-served seems fair, but from many views it is not. Why should some party get ownership of this ordinary English language phrase in the internet naming space just because they were first? Trademark law, which has dealt with naming problems for centuries, came to the conclusion ages ago that it's not fair or wise to assign ownership rights to generic terms to the first party to claim them. In fact, nobody can get ownership rights on them. That law only gives ownership rights in non-generics, to the first party to use them, in the sphere where they use them. Thus all the seemingly positive reasons you might want a generic create an unfair monopoly. You crave a generic because it lets you own something clear and obvious. That implies that nobody else can use that obvious name. Non-generics, because they contain a component without an inherent meaning, are unbounded. The ability of one person to own one places no limit on the ability of others to own equally useful ones. That means a level playing field. A level playing field means that we can assign the TLDs with minimal regulation. We can assign as many as there is demand for, and their owners can operate them any way the market sees fit. This fosters competition and innovation. On the other hand, if somebody owns a TLD like ".museum" it is inherently the best place for a museum to find a name. Competition is severely limited, and there is little need to innovate. Indeed, there is little need, other than that dictated by regulatory oversight, for the TLD owner to respond to the needs of its registrants. If they are a museum, they have little choice but to use the TLD that got ownership of that word. Having multiple registrars allows competition, but only in a very limited sense. Multiple registrars compete almost entirely on price, and end up charging the fee of the monopoly registry plus a dwindling epsilon. They compete a little on service (ie. how easy is it to register and renew) but they have very little room to compete on things like policy. They are all forced to use a first-come-first-served assignment policy, for example. In the end, they are all selling the same thing, with slightly different ways to pay for it. Innovation is very difficult in such a system. To innovate, the registry must cooperate, and thus all the other registrars in most cases. So competitive innovation isn't possible. For example, a TLD that offered "typo protection" so that any registration included registration of all the common typos and homonyms of a name, is not practical in such a system, but would be highly useful to people who like to be able to say domain names and have them heard and typed in reliably. I am interested in what values people who promote generic TLDs feel they hold, and reasoning as to why these are not in fact negative values, because every value they give to the owner excludes others. But generic TLDs are so clear and useful!People actively seek generic TLDs -- either as a TLD manager, or a party wishing to register a 2LD -- for a reason. But is it always a good reason? Certainly drug.store (or drugstore.com) is a clear an memorable name, even one that people might guess for an online drugstore. But of course there are many drugstores but only one gets to own this ordinary English word, which is a dangerous mistake. If we're trying to guess a name, we have a much better solution in search engines. Tools like google do a better and fairer job and finding a site when you know something about its name. User can choose their search engine based on how it ranks pages, and one search engine might rank different drugstores more highly than another for such a search. What about TLDs from trademark categories?One of my first ideas was to either build a TLD (I approached the republic of Turkmenistan to see if I could run the .TM domain) which would be divided up by trademark categories. Others have suggested that trademark categories could be used as TLDs. Thus we could have .computer and Apple Computer company could be the only party allowed to have apple.computer as a domain. Apple.records could go to the label. This seems good at first but what about "binary.computer" and "personal.computer" -- should the first person to claim these generic terms get ownership? Will the .computer registry have power over this or will a multi-registrar system force everybody to do first come first served on the generic terms? In addition, trademark categories are not a simple taxonomy. Trademark law simply looks to see if uses overlap in a way that would be confusing to the public. In particular, two people can own the exact same trademark if they don't overlap in geography. There is a Blue Note Jazz club in Missouri and another one in New York. They coexisted fine until they both wanted the bluenote.com domain, as it turns out. Sometimes the taxonomy is simple, but other times its not, and I recommend against trying to reflect it in a TLD structure. Let different competing TLDs try to come up with their own policies to do appropriate allocations to trademark holders. And of course national TLDs can follow whatever local rules they want. Such meaningful TLDs mean that the only way registrars can compete on an even footing is to use the multiple-registrar/single registry system, which grants monopolies to 2LD holders, and makes innovation and competition on other than price almost impossible among registrars.
With brand-name TLDs, disputes are handled through the courts or the dispute policies of an individual TLD. TLDs will compete on how good their dispute policies are. There will be no requirements to comply with globally set dispute policies other than about generic use. Trademark disputes are handled in trademark courts, where they should be. Trademark holders would sue infringers and get them ordered to fix any infringement. The TLDs should not be involved -- and ideally the courts will accept that they are no more involved than a business directory like Dunn and Bradstreet would be involved in a trademark dispute against a company named in their book. |