Hacker News Reader

Comment by jongjong | Hacker News Reader

I've always said over and over that password managers are a bad idea and kept getting down-voted each time. It centralizes the risk as it creates a large target for hackers to focus on. A good alternative is to use rules-based passwords; each website/app has a different password based on a common non-obvious rule which depends on the website, app or company you're using.

View on HN

You realize that if one password leaks, then the attacker learns the rule, and the rule can be used to get anything else? You haven't really made any improvements over standard shitty passwords. The passwords are different, but that doesn't matter.

The legendary researcher Dan Kaminsky had a hash of his password leak when some hacker forum got breached. Turns out the name of the forum was in the password, and by changing that, his password to his twitter, his blog, and his VPS with all his servers were all using the same pattern. It was a bad day for him.

That's why the rule needs to be non-obvious. The rule must be a secret in itself and there is no way that a hacker would figure out the rule from just a single or a couple of passwords.

For example, if I give you a sequence of 2 points, there are an infinite number of functions which could pass through these two points. A hacker who knows 2 numbers would not be able to figure out the other points unless they know the function.

Actually it's weird (to the point of being suspicious) why none of the password managers I know of use this principle for generating passwords from a single master password; then they wouldn't even need to store any of peoples' passwords at all. You wouldn't need password manager companies since it could be implemented as a purely front-end component (e.g. a standalone Chrome extension which doesn't connect to any server). The user would only have to remember a single master password.

You can basically do this using hmac-sha256 hashes to produce an infinite number of new hashes from a single secret seed and it's impossible to figure out hashes based on their sibling hashes and also impossible to figure out the secret seed (preimage) from any hash.

I tried that but it's just really hard to keep up over time -- e.g. I used a rule based on the domain name but domain names change somewhere often. Toss in things like "ugh, which of my three emails did I use on this site" or "which high school teacher did I say was my favorite for this site" and it ends up being a big hairy mess that screams for an encrypted place to stick my notes.

Also, what I consider "non obvious" isn't that non-obvious. Given enough of a sample size, a committed attacker can guess a lot of rules. And if the prize (a crypto wallet) is big enough, they might be motivated enough to give it a go.

If the domain name,, company name or whatever changes, you can change your password too. Also, it doesn't have to be domain names though.

Also when there's a breach and you need to change your password, you have to make an exception to your rule. And remember it for that specific site.

By rule, do you mean something like PW = PBKDF2(TLD || master_password), or something like “mysecurepw!amaz0n”? I see the latter a lot, and it makes me quite nervous – it seems quite pattern/AI-iteratable.

I have used that "insecure" system for 20 years. However, over the time I added additional several additional obfuscation levels, especially for accounts with monetary value. And then there are of course site password rules that force me to use various variations to my scheme. Additionally I use a different email address for every account. Those with monetary value are not predictable.

Of course all this became unmemorizable for less frequently used accounts over the years. So I had to use a password manager. However, I store only the rule/hint to generate the email address and password, not the data itself.

Could dedicated hacker with a keylogger learn my scheme? Absolutely yes. Am I afraid of any automated attack? No. I guess the keylogger is the real threat here. Once I am infected with that, the dedicated hacker can save their effort to learn my schemes. (Where I use keylogger in a wide sense, covering accessing your clipboard / browser input fields.)

Reply on news.ycombinator.com