Words are the least secure way to generate a password of a given length because you are limiting your character set to 26, and character N gives you information about the character at position N+1
The most secure way to generate a password is to uniformly pick bytes from the entire character set using a suitable form of entropy
Edit: for the dozens of people still feeling the need to reply to me: RSA keys are fixed length, and you don’t need to memorize them. Using a dictionary of words to create your own RSA key is intentionally kneecapping the security of the key.
That’s only really true if you’re going to be storing the password in a secure vault after randomly generating it; otherwise, it’s terrible because 1) nobody will be able to remember it so they’ll be writing it down, and 2) it’ll be such a pain to type that people will find ways to circumvent it at every possible turn
Pass phrases, even when taken with the idea that it’s a limited character set that follows a semi predictable flow, if you look at it in terms of the number of words possible it actually is decently secure, especially if the words used are random and not meaningful to the user. Even limiting yourself to the 1000 most common words in the English language and using 4 words, that’s one trillion possible combinations without even accounting for modifying capitalisation, adding a symbol or three, including a short number at the end…
And even with that base set, even if a computer could theoretically try all trillion possibilities quickly, it’ll make a ton of noise, get throttled, and likely lock the account out long before it has a chance to try even the tiniest fraction of them
Your way is theoretically more secure, but practically only works for machines or with secure password storage. If it’s something a human needs to remember and type themselves, phrases of random words is much more viable and much more likely to be used in a secure fashion.
Thanks for that! I recommend anyone who wants to minimize risk to follow their instructions for self-hosting:
Is the source code available and can I run my own copy locally?
Yes! The source code is available on Github. Its a simple static HTML application and you can clone and run it by opening the index.html file in your browser. When run locally it should work when your computer is completely offline. The latest commits in the git repository are signed with my public code signing key.
We are talking about RSA though, so there is a fixed character length and it isn’t meant to be remembered because your private key is stored on disk.
Yes the word method is better than a random character password when length is unbounded, but creating secure and memorable passwords is a bit of an oxymoron in today’s date and age - if you are relying on remembering your passwords that likely means you are reusing at least some of them, which is arguably one of the worst things you can do.
Most of my passwords are based around strings of characters that are comfortable to type, then committing them to muscle memory. There’s a few downsides to this:
If I need to log in to something on mobile and don’t have a proper keyboard with me, it’s tough to remember which symbols I’ve used
I share some of my logins with friends and family for certain things, if they call and need to re-enter a password, it’s usually impossible to recite it to them over the phone (most of my shared logins have reverted back to proper words and numbers to make it easier for the others)
If I lose an arm, I’ll probably have to reset all of my passwords.
But yeah, words alone provide plenty of possibilities. There’s a reason cryptocurrency wallets use them for seed phrases.
And even with that base set, even if a computer could theoretically try all trillion possibilities quickly, it’ll make a ton of noise, get throttled, and likely lock the account out long before it has a chance to try even the tiniest fraction of them
One small correction - this just isn’t how the vast majority of password cracking happens. You’ll most likely get throttled before you try 5 password and banned before you get to try 50. And it’s extremely traceable what you’re trying to do. Most cracking happens after a data breach, where the cracker has unrestricted local access to (hopefully) encrypted and salted password hashes.
People just often re-use their password or even forget to change it after a breach. That’s where these leaked passwords get their value if you can decrypt them. So really, this is a non-factor. But the rest stands.
It’s still a rather large pool to crack through even without adding more than the 1000 most common words, extra digits, minimal character substitution, capitalization tweaks, etc
so you are saying 44 bits of entropy is not enough. the whole point of the comic is, that 4 words out of a list of 2000 is more secure then some shorter password with leetcode and a number and punctuation at the end. which feels rather intuitive given that 4 words are way easier to remember
If you know the key is composed of English language words you can skip strings of letters like “ZRZP” and “TQK” and focus on sequences that actually occur in a dictionary
we are talking about RSA keys - you don’t memorize your RSA keys
if you rely on memorizing all your passwords, I assume that means you have ample password reuse, which is a million times worse than using a different less-secure password on every site
Words are the least secure way to generate a password of a given length because you are limiting your character set to 26, and character N gives you information about the character at position N+1
The most secure way to generate a password is to uniformly pick bytes from the entire character set using a suitable form of entropy
Edit: for the dozens of people still feeling the need to reply to me: RSA keys are fixed length, and you don’t need to memorize them. Using a dictionary of words to create your own RSA key is intentionally kneecapping the security of the key.
That’s only really true if you’re going to be storing the password in a secure vault after randomly generating it; otherwise, it’s terrible because 1) nobody will be able to remember it so they’ll be writing it down, and 2) it’ll be such a pain to type that people will find ways to circumvent it at every possible turn
Pass phrases, even when taken with the idea that it’s a limited character set that follows a semi predictable flow, if you look at it in terms of the number of words possible it actually is decently secure, especially if the words used are random and not meaningful to the user. Even limiting yourself to the 1000 most common words in the English language and using 4 words, that’s one trillion possible combinations without even accounting for modifying capitalisation, adding a symbol or three, including a short number at the end…
And even with that base set, even if a computer could theoretically try all trillion possibilities quickly, it’ll make a ton of noise, get throttled, and likely lock the account out long before it has a chance to try even the tiniest fraction of them
Your way is theoretically more secure, but practically only works for machines or with secure password storage. If it’s something a human needs to remember and type themselves, phrases of random words is much more viable and much more likely to be used in a secure fashion.
Generally people don’t memorize private keys, but this is applicable when generating pass phrases to protect private keys that are stored locally.
Leaving this here in case anyone wants to use this method: https://www.eff.org/dice
And if you don’t feel like using physical dice:
https://diceware.rempe.us/#eff
Thanks for that! I recommend anyone who wants to minimize risk to follow their instructions for self-hosting:
We are talking about RSA though, so there is a fixed character length and it isn’t meant to be remembered because your private key is stored on disk.
Yes the word method is better than a random character password when length is unbounded, but creating secure and memorable passwords is a bit of an oxymoron in today’s date and age - if you are relying on remembering your passwords that likely means you are reusing at least some of them, which is arguably one of the worst things you can do.
Okay, that’s fair… Not sure how I missed that context but that’s totally on me
You didn’t have to call me out like that.
Most of my passwords are based around strings of characters that are comfortable to type, then committing them to muscle memory. There’s a few downsides to this:
If I need to log in to something on mobile and don’t have a proper keyboard with me, it’s tough to remember which symbols I’ve used
I share some of my logins with friends and family for certain things, if they call and need to re-enter a password, it’s usually impossible to recite it to them over the phone (most of my shared logins have reverted back to proper words and numbers to make it easier for the others)
If I lose an arm, I’ll probably have to reset all of my passwords.
But yeah, words alone provide plenty of possibilities. There’s a reason cryptocurrency wallets use them for seed phrases.
One small correction - this just isn’t how the vast majority of password cracking happens. You’ll most likely get throttled before you try 5 password and banned before you get to try 50. And it’s extremely traceable what you’re trying to do. Most cracking happens after a data breach, where the cracker has unrestricted local access to (hopefully) encrypted and salted password hashes.
People just often re-use their password or even forget to change it after a breach. That’s where these leaked passwords get their value if you can decrypt them. So really, this is a non-factor. But the rest stands.
That’s fair
It’s still a rather large pool to crack through even without adding more than the 1000 most common words, extra digits, minimal character substitution, capitalization tweaks, etc
Good luck remembering random bytes. That infographic is about memorable passwords.
You memorize your RSA keys?
you memorize the password required to decrypt whatever container your RSA key is in. Hopefully.
Sure but we aren’t talking about that
I think this specific chain of replies is talking about that actually… though it is a pretty big tangent from the original post
“can you string words to form a valid RSA key”
“Yes this is the most secure way to do it”
“No, it’s not when there is a fixed byte length”
-> where we are now
the direct chain I can see is
“can you string words to form a valid RSA key”
“I would hope so, [xkcd about password strength]”
“words are the least secure way to generate random bytes”
“Good luck remembering random bytes. That infographic is about memorable passwords.”
“You memorize your RSA keys?”
so between comments 2 and 3 and 4 I’d say it soundly went past the handcrafted RSA key stuff.
Sounds like a good point, but claiming that “Words are the least secure way to generate a password 84 characters long” would be pointless.
and some people will try to just hold a key down until it reaches the length limit… which is an even worse way to generate a password of that length
so you are saying 44 bits of entropy is not enough. the whole point of the comic is, that 4 words out of a list of 2000 is more secure then some shorter password with leetcode and a number and punctuation at the end. which feels rather intuitive given that 4 words are way easier to remember
No im saying if your password size is limited to a fixed number of characters, as is the case with RSA keys, words are substantially less secure
Not if you’re considering security gained versus difficulty of remembering.
You don’t memorize RSA keys
That’s why you need lots of words. (6) If you combine that with a large word list it gets very secure.
There is no point in a password cracking attempt during which the attacker knows the character at N but not the character at N+1
If you know the key is composed of English language words you can skip strings of letters like “ZRZP” and “TQK” and focus on sequences that actually occur in a dictionary
Edit: Oops forgot what the topic was.
we are talking about RSA keys - you don’t memorize your RSA keys
if you rely on memorizing all your passwords, I assume that means you have ample password reuse, which is a million times worse than using a different less-secure password on every site
Derp. Forgot where I was.
I find passphrases easy to remember and I have several. I appreciate the concern, but I understand basic password safety.