Continuing the discussion from RFC - Decentralised Naming System V - No DNS:
Don’t take this as support but one way would be for billboards (etc) to include enough information to include a lookup service and their identity on the service. This isn’t restricted to solving the above problem just for your scheme. It can resolve to native SAFE addresses, so is a stand alone proposal for multiple provider DNS.
So you can imagine multiple services starting up which could be named memorably as:
Someone would register their desired “domain” name with one of these app/services (paying fee) to the service owner.
To support this, the network implements a special network owned address (a nice catchy domain in itself) which is used to host these services and handle the registered names. Let’s call the network thing “name”.
So let’s say you register “apple” with the service called “lookup”. On your napkin, billboard etc you would specify:
Or if you’d used business:
Or if you’d used “spangle”:
Its not ideal, but it maps fairly well to the existing DNS (so we can predict the behaviour) - the service names map quite well to TLDs. So the network could potentially make a substantial charge for the registration or use of one of these names on its “.name” facility.
The URLs are easily understood (if backwards). Perhaps we can somehow ditch the “.name” part and find a way for the network to handle these like real TLDs.
So instead of
You would have:
And obviously someone could setup “com”, “uk” etc so on SAFE we’d have domains that look very like the current TLD system:
But each TLD would be administered by an app builder who paid to register with the network’s built in name service feature. Squatting on TLDs can be handled by making it very costly - as with the current system which is I think is about $200k for a TLD, but they actually auction them so it can be more.
This requires users to register with a commercial company - whoever owns the TLD - through their SAFE App - or a general purpose SAFE App that lists all the TLD services for you to choose from.
It effectively replicates the current system, but providing extra security, because once you have registered the domain, it could be made impossible for it to be changed without your signature. So avoiding many of the security issues of the current DNS.
There are various issues to be considered and decided on, but it seems we could come up with something remarkably similar to the existing DNS but with some of the problems removed.
This was a brain dump. Not thought through, but worthy of consideration I think.