There are parts that could be tricky in pure Erlang / Elixir, for instance anything that requires constant-time arithmetic.

One way to avoid that would be to wrap an existing Rust library like k256 with Rustler. I’m neither a cryptographer nor much of a Rust dev, but this example seems like it wouldn’t be too much hassle to hook up.

Hey! Did a rough implementation that I think covers your use-case

https://github.com/davidarmstronglewis/k256

alias K256.Schnorr

signing_key = Schnorr.generate_random_signing_key()
message = "This is some content to sign"
assert {:ok, signature} = Schnorr.create_signature(signing_key, message)
assert {:ok, verifying_key} = Schnorr.create_verifying_key(signing_key)
assert :ok = Schnorr.validate_signature(message, signature, verifying_key)

There’s a few things I’m not happy about in this first pass

1. I’m not pre-compiling anything for downstream users, that’ll probably be necessary if this is being published to Hex.
2. signature and verifying_key are all lists, but I think people would expect binaries or some other opaque blob. Not sure the best way to do this atm.
3. Missing specs.
4. READMEs are whatever was generated by default, docs are minimal

Let me know your thoughts and if you’d like I can run through anything you’d like to know more about

Again, not a cryptographer, don’t know much about nostr, but enjoyed the challenge here.

Thanks, I will try this right away and potentially start a nostr lib today, thanks to you!

I Agree about the 4 improvements, in the same order. Binaries would be the best as input.