@riddim has described how a DNS could be implemented, so it is feasible based on the current design. Based on that anyone can set this up and apps can adopt it or not.
Pointer (secret key derived from name, count exhausted which makes it read only) --points-statically—> pointer (owned by the name owner so he can change where the name points at) --points-to-website—>
(in this scenario names would be non-transferable…
Because the initial owner of the pointers knows their secret keys)
I really hope we get there - petnames are perfectly fine imho and gives the power to the end user instead of allowing people to reserve names just because they happen to be first
And I’m all for that. Anyone should be able to “start” their own, like email domains (or just domain names in general)
If there was one default network DNS then everyone would have to run to squat a million of the common word names from day one, and then try to re-sell them or whatever.
It would be nice to see something like
ant://dave.matrix
where “matrix” is someones chosen DNS fork and “dave” is a users site address.
Except now that I think about it all that means is it shifts the squatting race to the DNS system namespace.
I know (at least years ago last I tried it), in the I2P project anyone can run a DNS, but then there is never a common truth on what name points to what address.
Just remember the NRS (not domains) - Name Resolution System will be an Application level app.
Autonomi does not have network domains like the current internet was built on which required a system to easily access resources on that domain of the internet space. Thus DNS.
Autonomi is a flat system without anyone owning sections of the network space. So its a Naming system without the need for centralised control like DNS
Another is to have the user run and update a NRS for their public files, then a NRS system only needs to point to the user’s NRS list and thus we can build a web of NRS
All sorts of ideas are possible.
I suspect we’ll get something like adblock where you subscribe to a a few trusted lists.
We may want to big the concept of dot separated subdomains too. They have special security meanings with browsers, https, etc.
We don’t want 2 completely different owners of data pretending to be same entity, by virtue of sharing the same parent domain, i.e. each register my.google and nomine.google and browsers treating them as the same from a security perspective.
There are ways to limit the impact of those issues, but we may just want to avoid that can of worms. Especially, if we want to re-use existing software.
Caveat: some software may require a fully qualified domain name, but they should accept a simple hostname too (with no dots!).
If anyone can pick any (pet) name, there can be no control over who re-uses the higher level domain names.
So, I could save traktion.cooldomain and you could register neo.cooldomain, but there may be no relationship between the sites/apps/data they point to.
This could be abused intentionally or accidentally.