Why is encode_to_curve non-uniform?

[matthiasgeihs]

curve25519-dalek/curve25519-dalek/src/edwards.rs

Line 688 in 015707a

/// `encode_to_curve` is a nonuniform encoding from byte strings to points in `G`. That is,

Can someone shed some light on:

1. Why is encode_to_curve non-uniform?
2. Why is hash_to_curve uniform?

[tarcieri]

That is how they are defined in RFC 9380. See in particular Section 10.4:

https://www.rfc-editor.org/rfc/rfc9380.html#security-considerations-encode

Which one you pick depends on your use case. If you need a random oracle encoding as defined earlier in section 2.2.3, then encode_to_curve is unsuitable. But it's useful if you just want a fast, deterministic mapping from an input string to a curve point.