Careful. Salt quality is almost totally unimportant. And there are no well-known, well-reviewed password hashes that aren't randomized. The lesson I'd take away from an article like this is: "don't DIY your password hash".

Historically, every discussion I've seen that tried to acknowledge the "importance" of salts devolved into people defending their DIY "secret-salt" schemes. "Important salts" seem like an invitation to reject bcrypt, scrypt, and PBKDF2.

I think part of the problem is the term. I think we should stop talking about "salts", and instead just say whether a given password scheme is randomized or not.