Password Difficulties

Hey folks, passwords are hard to choose!
 
It boils down to this: I can't remember as many bits as the TLAs can
crack by brute force.
 
Starting with a bunch of coin tosses I tried ways of coding them: hex,
ASCII, and words off word lists.
 
Horrors!  The hex is too long, the ASCII is too long and too obscure,
words words chosen by those bits too many and too obscure.
 
Sorry, there is no way regular people are going to remember pass words
or phrases with more than about 50-bits worth of information in
them--and even doing that well is going to be rare.
 
We need to slowdown password testing?
 
Obvious things come to mind.  1) Try to pair up short passwords with
slow hardware, like a smartcard that can only consider a few passwords
a second.  2) Try to hide behind an expensive operation.  (Does
encrypting my private key 1,000,000-times equal encrypting it once
with a key 20-bits longer?)
 
What do we do?  (What are you folks doing right now?)


-kb, the Kent who occasionally considers practicalities


--
Kent Borg                                                  +1 (617) 776-6899
kentborg@world.std.com                                
kentborg@aol.com                                      
          Proud to claim 31:15 hours of TV viewing so far in 1994!

>   Hey folks, passwords are hard to choose!

?

>   It boils down to this: I can't remember as many bits as the TLAs can
>   crack by brute force.

I generally choose things like (no, this is not a real one):

Rare steak tastes good when it is cooked over a wood fire. better than
chicken. better than fish. good with worcestershire sauce.

this is for a pgp passphrase, of course.

I find it not to be a problem remembering a sentence character for
character.

>   Starting with a bunch of coin tosses I tried ways of coding them: hex,
>   ASCII, and words off word lists.

>   Horrors!  The hex is too long, ....

>   Sorry, there is no way regular people are going to remember pass words
>   or phrases with more than about 50-bits worth of information in
>   them--and even doing that well is going to be rare.

?

josh

On Sat, 2 Jul 1994, joshua geller wrote:
> [. . .]
> >   It boils down to this: I can't remember as many bits as the TLAs can
> >   crack by brute force.
> 
> I generally choose things like (no, this is not a real one):
> 
> Rare steak tastes good when it is cooked over a wood fire. better than
> chicken. better than fish. good with worcestershire sauce.

You can improve entropy even more, and still keep it memorable, by doing 
something such as the following:

Rare 513AK tastes g))d when it is c))K#D over a wood fjord. 
BETTERthanCHICKEN....

Using poor or improper English--or some other language--will also help. 
So now, we might have:

Viva dA5 bu0n) Rare 513AK tastes w3#l it when 15 c))k#D....

You, of course, will have to be the judge of how much mutilation you can 
remember.

And note that, while such changes will help with passphrases, any 
sophisticated dictionary/algorithm-based password (>8 charcters) cracker 
will be able to guess most of them. "f43d" is no more secure than "fred." 
Better to hit random keys on the keyboard or use a true random number 
generator--flip a coin 56 times to get a 7-bit ASCII string, more if you 
get control characters--to get your eight characters, and just force 
yourself to remember it. Even something like "g&*3VkjH" is memorable--I 
did use that one for a couple weeks some months ago.

Speaking of which, are there any /bin/passwd plugins that use 
passphrases rather than passwords? Or should I be a good cypherpunk and 
write some code?

> [. . .]
> josh

b&
--
Ben.Goren@asu.edu, Arizona State University School of Music
 net.proselytizing (write for info): Protect your privacy; oppose Clipper.
 Voice concern over proposed Internet pricing schemes. Stamp out spamming.
 Finger ben@tux.music.asu.edu for PGP 2.3a public key.

> It boils down to this: I can't remember as many bits as the TLAs can
> crack by brute force.

Have you *tried* to memorize these long passphrases?  I pick ones that
are substantially too complex for me to memorize in one trial.  So I
write the candidate passphrase on paper until I have a grasp on it,
then burn the paper, scatter the ashes (yes, literally), and begin to
use the passphrase.  My experience is that once I've successfully
remembered a phrase two or three times, I will not forget it.

This approach is vulnerable to anyone who is able to snoop around my
belongings, but at that point they might just as well do what they
did to Ames.  I hardly think I warrant this kind of attention.

   Eli   ebrandt@hmc.edu

On Sat, 2 Jul 1994, Eli Brandt wrote:

> > It boils down to this: I can't remember as many bits as the TLAs can
> > crack by brute force.
> 
> Have you *tried* to memorize these long passphrases?  I pick ones that
> are substantially too complex for me to memorize in one trial.  So I
> write the candidate passphrase on paper until I have a grasp on it,
> then burn the paper, scatter the ashes (yes, literally), and begin to
> use the passphrase.  My experience is that once I've successfully
> remembered a phrase two or three times, I will not forget it.
> ... 

I have actually tried memorizing truly random passwords of 8 characters 
or longer (generated with a paranoid program similar to PGP 2.6's 
excellent technique).  I've found that if I review it enough, that I find 
patterns and mnemonic clues in such passwords that help me to remember 
them.  I don't imagine too many people will go through that effort, so I 
still think that a longer pass phrase that sort of "makes sense" is 
better for a PGP key.  Still, I do use the truly random passwords on 
publicly accessible Unix systems like CSN, since that makes dictionary 
attacks improbable.