[chickopozo]

>Outside of this, no one should be using MD5 in this day and age, I would even recommend against Sha256 and other variants in that family, though I have a lot of respect for Blowfish, etc.

@shanelija SHA-256 is a secure cryptographically secure hash function. It is not and was never intended nor should be used as a key derivation function.

Given time and email are not unique, they should not be used as a salt. Concatenation of a salt is not the correct method either, not for KDF and not for message authentication.

One of the aims of a hash functions is to be fast, really fast, super fast while making it computational infeasible to: make a message that will hash to a given value alter a message that when hashed will produce the same or a give value. Find two different messages with the same hash value.

bcrypt, scrypt, PBKDF2 are perfect for password storage.

YOU are not off the hook yet until you repeat these 2 sentences 10 times: A secure crypto algorithm/system/function/whatever MUST continue to be secure even if the algorithm/system/function/whatever is public.

It is perfectly fine to homebrew crypto given one key condition, you never ever ever ever use it for things that matter.

[dhimes]

shanelija said s/he uses bcrypt. Apparently s/he just makes the passwords stronger than the user makes them before hashing.