But what I'm planning to do is similar to what you're doing in Ditto (if I understand it correctly).
- Client interacts with the server using Mastodon API. - When the user performs some action, the server creates an activity and sends it to the client. - Client signs the activity and sends it to outbox (per FEP-ae97).
So, I think implementing FEP-ef61 in Ditto may make sense.
@alex Yes, it should be possible. Fediverse projects will likely end up using Ed25519 keys, but with FEP-ef61 it is easy to support many different types of keys and signature algorithms.
@alex Many Mastodon users want data portability, so I wouldn't rule that out. But even if Mastodon devs will implement it, they certainly won't be the first.
Right now FEP-ef61 is just an idea. Some parts are not quite elegant, (especially inbox discovery), W3C standards which FEP-ef61 is based upon are unstable, likely this proposal is not compatible with JSON-LD... So there's a lot of work to do (and your input is welcome!).
@arcanicanis@silverpill I don't think that much thought was put into it. It's the same curve as Bitcoin. It works well enough for Bitcoin, where actual money is at stake, so it couldn't be a bad choice for shitposting. Even if others may possibly be better.
Wait, secp256k1 (as in Koblitz curve, as used in Bitcoin/Ethereum, only meant for a specific narrow use and supposedly has much more ways to be easily misused, making key compromise easier) versus P-256 (secp256r1) or the much simpler/performant Curve25519? Was there a reason for adopting specifically secp256k1 in Nostr above all others?