Hacker News new | ask | show | jobs
by negativity 4400 days ago
I dunno, it's still pretty debatable, since e-mail isn't guaranteed to be encrypted over the wire, which leaves people open to MITM attacks.

So consider a situation where someone receives a password in plain text, and the password never expires and never gets changed by the user.

All things considered, a token is a token, so whether the "password" is sent in plaintext, or whether a nonce hex key is provided by e-mail, anything sent by e-mail should have a shelf life, even if it's a relatively long one of like 30 days.

Ideally, it should expire in hours or minutes. If they don't get around to it fast enough, you have the user's e-mail, just tell them you need to send them another, because the last one expired. That way, you're forcing a live user to interact with the system, and act quickly, to establish proper authentication credentials.