search

bitwiseshiftleft / sjcl

Stanford Javascript Crypto Library
http://bitwiseshiftleft.github.com/sjcl/
Other
7.19k stars 988 forks source link

non-weiestrass curves #147

Closed pera closed 10 years ago

pera commented 10 years ago

Would be possible to use alternative curves like Montgomery and Edward? I would like to use curve25519..

cheers

bitwiseshiftleft commented 10 years ago

Yes. But SJCL isn’t designed as a complete toolbox that implements every primitive; we selected just a few which we hope will be good enough, so as to reduce complexity. That’s why there’s no Curve25519.

On Jan 13, 2014, at 4:46 PM, pera notifications@github.com wrote:

Would be possible to use alternative curves like Montgomery? I would like to use curve25519..

cheers

— Reply to this email directly or view it on GitHub.

pera commented 10 years ago

I understand, reduce complexity in a relative new crypto library seems like the right decision... but still, it would be awesome to construct Montgomery and Edwards curves with sjcl.ecc.curve(). Especially after the lastest (and not so) revelations concerning NIST's recommended ECC. thanks

alax commented 10 years ago

@pera The only revelations concerning the NIST's recommendations are regarding Dual_EC_DRBG, which is a random number generator based on elliptic curves. It has absolutely nothing to do with the asymmetric cryptography curves.

pera commented 10 years ago

@alax Yes, s/ECC/EC CSPRNG/, sorry. I didn't try to say that SJCL included any known backdoored curve, but only that some people are moving away from NIST recommended curves because they don't trust them anymore, and that currently SJCL is only able to construct Weiestrass form curves. Though afaik there is no evidence of any potential risk on using those parameters...

Nilos commented 10 years ago

I created a ticket to keep track of all curve addition requests here: #158 Closing this ticket for cleanness of backlog.