[curves] Another try at point compression

[Samuel Neves]

On 16-12-2014 18:38, Trevor Perrin wrote:
> So I don't think it's worth much added complexity to have an encoding
> that "eliminates the cofactor for all practical purposes".

Insofar as this proposal could contribute to further fragmentation of elliptic curves and point formats, I agree. It has
been a slow process to move to curve25519, despite the near 10 years it has been around as the sole non-NIST proposal. A
larger curve that fits right into curve25519's shoes thus makes sense.

That said, I'm excited about this. Mike's idea effectively eliminates cofactors from the equation, which makes it
possible to treat all these curves as prime-order. Cofactors are probably one of the last technical reasons that keep
Weierstrass curves being considered; this could be a convincing argument to finally move away from them.