Thread

The real problem is not the IP addresses is the DNS resolver.

Even though i think it's a good experiment we are not even able to get rid of ipv4 in favour of ipv6 after 30 years. Even though it has many benefits, backwards compatibility also makes a strong case here.

https://tim-bouma.npub.pro/tag/dnspub/ nostr:nprofile1qqsthdwa5rs42euhnuz5xsrmmssr84hshwes7uj392vpeldj7z0zw3cppemhxue69uhkummn9ekx7mp0qyghwumn8ghj7mn0wd68ytnhd9hx2tc46s0t0

I that's because ipv6 doesn't really solve the fundamental issues and it required a tremendous amount of cooperation to even get support for it. Backwards compatibility is a must but there seem to be ways to integrate with existing software without having to rewrite it all. (as is the case with something like reticulum)

Ok well i'm looking forward to a proposal or maybe even some kind of demo? What did you have in mind?

I want to validate some prerequisites for this to work before trying to put it all together. - The first thing i need to know is if I can indeed intercept traffic using the local ipv6 trick. - Then some basic network discovery where nodes announce themselves. This doesn't scale but is good to have for local testing. - Only if that works, the longer distance routing (either source-routing or landmark-routing) becomes possible. I don't know if it should be part of a NIP, it feels more like its own thing.

DNS is definitely a problem, but still builds on IP. The short-term solution to me is fixing DNS The long-term one is decentralising routing