[curves] The great debate over point formats

[Trevor Perrin]

On Tue, Jan 28, 2014 at 11:44 AM, Watson Ladd <watsonbladd at gmail.com> wrote:
> On Tue, Jan 28, 2014 at 11:41 AM, Michael Hamburg <mike at shiftleft.org> wrote:
>> Do you mean the whole Edwards-x-coordinate, or just the sign?
>
> Good catch: just the sign is probably the best choice.

That's similar to point compression, which some people have avoided
due to patent fears [1].  But the patent in question expires in July
29 2014.  Since it will probably take that long to get any new curves
deployed, that doesn't seem like a problem.

It would be interesting to see an analysis of how this "unified
format" affects performance and code complexity vs simply using 2
separate formats, one for ECDH and one for signatures, as is done for
curve25519 / Ed25519.


Trevor


[1]
http://www.google.com/patents/US6141420
http://cr.yp.to/patents/us/6141420.html