Using Scrypt for user/password derived key #721
Replies: 3 comments 4 replies
-
I have spoken to two cryptographers who claim that argon2 only looked better than scrypt during the window of time of that competition (was it 2014? I forget), and subsequently they believe scrypt is better. PBKDF2 is not state of the art. |
Beta Was this translation helpful? Give feedback.
-
I've taken a stab at this and clients seem to work well, I'll leave this here for comment for awhile. Insufficiently long passphrases are an inherent weakness as with anything else, but I think there's room to design an improved UX that provides acceptable hardness.
|
Beta Was this translation helpful? Give feedback.
-
Interesting idea. Obviously you can't click "I forgot my password", and you're ultimately still giving the app your private key. But this idea is appealing to me since I'm building Nostr into a legacy system. |
Beta Was this translation helpful? Give feedback.
-
Allowing users to log in with a more traditional username/password combination, or a human readable long string, would offer better UX to users in many cases. The former at minimum being less glitchy with browser-based password managers.
I was thinking back to how GUN did this with PBKDF2, and upon a little research found Argon2 is the alleged successor. EDIT seems that
scrypt
may be even better.There are considerations over my head with using it to create seeds for nostr compatible keypairs, but from what I do understand it sounds doable.
I wanted to throw the idea out there short of a proposal as I'm not qualified to write it.
I'll kick in on a bounty if others think this conceptually has merit.
Beta Was this translation helpful? Give feedback.
All reactions