Asking the right questions starts with looking at the wrong candidates. Is a central-control domain system a good idea? What else is there? What else can one think of? Is trusting anyone good architecture? Is there still hope to implement something else? Are domain names the first kind of e-coin? Are the keepers innovative companies or just rent seekers? Do the customers have any influence or are TLD's an unavoidable monopoly?
Central control over this type of resource is demonstrably a good idea. Domain naming, considering the global scope and collision possibilities, is analogous to the electromagnetic spectrum from a regulatory perspective. If we didn't have the FCC and instead relied on private entities to regulate our radio spectrum, I am almost certain we would be amidst technological anarchy on a daily basis. Even with the FCC as the central authority, we continue to have a number of conflicts that come up. It would seem as technology progresses this space is getting harder and harder to manage. Satellites, self driving cars, next-generation cellular communications...
I hesitate to start involving real-world geopolitics, but perhaps this is yet another compelling use case for a body such as the UN. Perhaps you don't trust one nation (e.g. the USA) to administer the entire domain naming system, but you would almost certainly trust a quorum of (ideally peaceful and democratic) nation states to do so.
Wireless spectrum isn't a great analogy for domain names. Spectrum is extremely scarce in practice, names aren't. There are hundreds of bits of entropy available so we're never going to run out of domain names even in an individual TLD. Even good names aren't particularly scarce because there are so many TLDs.
The real issue here is that once you start using a name, it becomes a dependency. You may not have any preference for example.org over example.us on day zero but you sure do after you've been using one of them for a decade.
So the thing that works really well looks a lot like Namecoin, but not quite. You do want some modest annual maintenance cost per name just to discourage squatting on short/common names and cause people to return names to the available namespace once they're not using them anymore. What you don't want is for anybody to be able to ever cancel an active registration or change the terms or cost once a name is already registered.
You could do that with blockchain like Namecoin, or you could do it through legislation or contracts. Either way, it shouldn't be possible for anybody to cancel or change the terms of a registration for a name that's already registered, and that's the biggest failing of the existing system.
Yes the spectrum is scarce in practice, but in theory it doesn't need to be. The problem in practice is that there are massive chunks of extremely useful spectra locked up in legacy telecommunications stacks that make the spectral efficiency of 4G look like an impossible feat by comparison.
If we had a magic wand and could instantly put everything on a standardized global DSSS/CDMA-style scheme (across any arbitrary frequency ranges), I feel we would have far more to go around.
That doesn't really change the comparison. It would be better if we made more efficient use of the wireless spectrum, but if we did then people would just replace more copper and fiber with wireless. The wireless spectrum would still be in contention as a scarce resource in a very practical sense.
By contrast, people registering every possible domain name is never going to happen. Just the computational resources necessary to do that don't even exist in the world.
What makes a name valuable isn't that names are rare, it's that people know your name. Things point to it. So we don't really need to ration them but we do need to make it so that once a name is yours, it stays yours. Nobody should be able to hold your own name hostage from you.
Quorum / consensus seem to have a problem where players are factionalized or share dissimilar aims.
See: various UN bodies
Particularly with bodies where alignments of players change, or where new players are added. Either you set the threshold too low, and it doesn't represent consensus; or you set it too high, and the body is gridlocked by minority dissent.
Consensus seems to work better when at least all players share a nominally similar interest (read: not hot-button international geopolitics).
I do agree that the actual implementation of this body would be tricky to balance. But, the stakes involved here are pretty high. Virtually everyone has a vested interest in globally-consistent DNS resolution, excepting a few rouge actors. That said, even things that 99% of the world can agree upon (e.g. stop building nuclear weapons so we don't vaporize ourselves) are not being taken to heart by a large number of actors today.
Perhaps we simply look at China & Russia as the worst-case situations, and try to build policies from these perspectives. Would it be that terrible for the rest of the world if Russia decided to break away from the internet and form a domestic internet? Obviously, many businesses and individuals would become impacted by this. But, if no compromise can be reached in a global body, this is seemingly an acceptable outcome for all other actors. We already collaborate with China on dismantling the intent of a free and open internet for their citizens, so I don't think there is any way through this where we don't end up a little dirty in the process.
Maybe what we do is proceed with the globalization of DNS and just build the best goddamn internet we possibly can. If the dissenters do not want to participate initially, they will likely have to re-join the rest of the world once they realize the financial impact of not doing so.
> Virtually everyone has a vested interest in globally-consistent DNS resolution
I'd disagree and instead say that most nation-states prioritize their own aims (e.g. copyright policing, piracy, sanctions) above having a globally-consistent name resolution, where the two come in conflict.
And furthermore, many states prioritize their own economic development above globally-consistent name resolution. E.g. How many English domain names are owned by Liberians?
The trouble with international consensus is that is has to build one something that everyone wants.
... And the closest I could come to that would be a hash-based GUID resolution system. (IOW, something that lacks scarcity)
IMO: WiFi works well enough we should try upping the power limit in a few bands (maybe lower frequency than WiFi.) I’m not sure exclusively licensing bands makes sense with packet radio.
So while something like namecoin might be an fine experiment for a green-field system, the thing about DNS is it's already established so there's way to much fraud that would happen on even a slightly bungled transition. I'd be for the UN too.
Also, if something like trademark law ever gets involved (not saying I hope it does) and the international dispute resolution would be really hard to automate and a UN body would also be useful as the central authority.
Finally, I think it would be good for the world if most people interacted with a UN-administered thing every day. Not really the "teeth" it's missing, but a good step in that direction.
Does anybody know whether `.ngo` and `.ong` -- also managed by PIR -- are part of the package? Hunting around on the PIR website produced no hint either way.
