>I suppose you could have an even less centralised system to resemble Tor hidden services. This transcript of Assange and Schmidt's meeting discusses this at a certain point, just grep for 'hash'[1]. That way the domain name itself is proof of its authenticity.
Using a hash as the name can be useful, especially where the name is only being read by a machine, e.g. you post a link and somebody can just click it, or it's part of your app which you're only using instead of an IP address in case the IP address changes. Or you can put it in a QR code or use NFC on mobile devices etc. The trouble is that it causes the name to be full of encoded data and humans can't remember it. You would still like some way of using memorable names for instances where someone is going to have to type the thing.
>Couldn't you also have something akin to bitcoin's blockchain, where the identity of a domain is agreed upon by the majority of the creators of the blockchain? I don't know what the equivalent of mining would be though. Assange also suggests how domains could be made hard to make, so that they can be 'mined', creating scarcity so that "some arsehole" doesn't "register every short name themselves"[sic].
I was thinking about something like that, it seems like the trouble is how do you calibrate the amount of work to be done. If you make it massive (like $200,000 worth of CPU time on Amazon) then you're excluding a lot of the people you wouldn't want to exclude, or causing them to waste a lot of money. But anything significantly less formidable just isn't going to solve the problem -- at $200 you can still imagine a slew of jackasses registering all the short names. Especially when they're evildoers who are using a botnet and don't actually have to pay anything for the computing resources. And that also doesn't solve the problem of scammers getting ".bank" or ".irs" or something, where the point isn't that they're getting too many names, it's that they're getting unreasonably misleading names.
Using a hash as the name can be useful, especially where the name is only being read by a machine, e.g. you post a link and somebody can just click it, or it's part of your app which you're only using instead of an IP address in case the IP address changes. Or you can put it in a QR code or use NFC on mobile devices etc. The trouble is that it causes the name to be full of encoded data and humans can't remember it. You would still like some way of using memorable names for instances where someone is going to have to type the thing.
>Couldn't you also have something akin to bitcoin's blockchain, where the identity of a domain is agreed upon by the majority of the creators of the blockchain? I don't know what the equivalent of mining would be though. Assange also suggests how domains could be made hard to make, so that they can be 'mined', creating scarcity so that "some arsehole" doesn't "register every short name themselves"[sic].
I was thinking about something like that, it seems like the trouble is how do you calibrate the amount of work to be done. If you make it massive (like $200,000 worth of CPU time on Amazon) then you're excluding a lot of the people you wouldn't want to exclude, or causing them to waste a lot of money. But anything significantly less formidable just isn't going to solve the problem -- at $200 you can still imagine a slew of jackasses registering all the short names. Especially when they're evildoers who are using a botnet and don't actually have to pay anything for the computing resources. And that also doesn't solve the problem of scammers getting ".bank" or ".irs" or something, where the point isn't that they're getting too many names, it's that they're getting unreasonably misleading names.