Rev.1 of "Private communications over public infrastructure" is up.

Hey everyone! How’s it going? I’m introducing Rev.1 of my article "Private communications over public infrastructure." The first version walked through NIP-04, NIP-17, Marmot, and Double Ratchet, exploring whether private communications over public infrastructure are even possible and whether privacy and encryption are the same thing. The short answer was that it’s mostly not possible because relay metadata does most of the work.

This revision is a rewrite of some parts and an expansion. The biggest change is the addition of three new protocols to the survey: Concord, Nymchat, and Cordn. I’ve also added more nuances about the concept of "Sovereignty" and included a helpful table that puts all the protocols and their features together. I hope you enjoy reading it!



The highlights:

- Covered the in-progress Marmot v2 draft and what it changes regarding privacy.
- The NIP-4e section is now fully written out instead of being a stub.
- Added subscription filters as a key piece of metadata: the filter a client sends to a relay is itself observable, and this nuance now runs through the entire piece.
- A new framework section defines the tradeoff dimensions upfront (forward secrecy, post-compromise recovery, and others), with a comparison table mapping all eight protocols across them.
- Introduced a two-observer model: external observer vs. relay operator, because some protocols hide more from one than the other.
- The conclusion now clearly distinguishes between exit sovereignty (the ability to leave a relay) and control sovereignty (governing the substrate), reframing what Nostr actually delivers.

naddr1qvzq...