clarification of terms

[coolaj86]

Seed

1. Structure

Is every random 256-bit (32-byte, 64-character) hex string a valid seed, or does it have some constraints?

In other words:

If I run openssl rand -hex 32 a million billion trillion times, will all outputs always be valid seeds?

2. Usage

Do we know anything about the mechanism that turns a seed into a secret?

For example: is it based on a typical CSPRNG? Or a curve? Or something else?

Secret

1. Structure

Is every random 32-byte array a valid secret? Or are there some constraints?

IIRC, the code just iterates through 32-byte CSPRNG buffers until it finds one that doesn't fail to generate a public key.

[wmerfalen]

1) No. 2) AugSchemeMPL::KeyGen is the function that's used, but I know nothing about it's internals, so: No. Secret, 1) I don't know

[coolaj86]

AugSchemeMPL::KeyGen

So is a seed truly a seed? Or is it just a secret key, and any secret can be used to generate another secret?

What's the technical, literal difference between a seed and secret as used in the code?

[wmerfalen]

A seed creates a secret. A seed is a seed. A seed is not treated like a secret. A secret can be used to generate other key pairs by generating the Nth index. The literal difference between a seed is that it is never used as a secret. A seed is used to generate a secret.