Hacker News Reader

Comment by nickik | Hacker News Reader

Surprised to see no mention of GNUNet. GNUNet arguable has the most complete stack and they have a GNU Name System so they can go beyond public key as direct identifiers.

I would really love to see matrix use GNUNet and GNU Name System as their P2P layer. It just hides so much information by default and its looking for a 'killer app'.

View on HN

Would be great if a security-head would write up a deep dive on Gnunet's codebase.

I think the long-standing assumption is that it could do what it advertises if the users were there. But it'd be nice to know what the actual state of the implementation really is.

and then rewrite it in Rust.

I've been excited about GNUnet in theory for a very long time, and they have some great ideas... but despite that, I've very rarely been able to run any of their actual software and have it work reliably.

And that's understandable, given that their website states prominently: "GNUnet is still undergoing major development. It is largely not yet ready for usage beyond developers." But it started around two decades ago; at a certain point I stopped expecting that I'd be able to use it anytime in the near future.

That's why I would love somebody like Element to look into it and potentially actually using it, helping solve the problems that prevents it form being commercial.

I just like the model of having multiple layers of encryption, helping hide metadata, having a name system and routing across different protocols.

Reply on news.ycombinator.com