[TUHS] OT: Patternless Encryption

[TUHS] OT: Patternless Encryption

[Carl Lowenstein]

> X-From: mirian at trantor.cosmic.com (Mirian Crzig Lennox)
> Subject: Re: [TUHS] OT: Patternless Encryption
> Date: Tue, 8 Jul 2003 16:29:59 +0000 (UTC)
> 
> On Tue, 08 Jul 2003 15:31:08 GMT, John P. Willis <jwillis at coherent-logic.com> wrote:
> >
> >Just curious to hear the opinions of the many wise people here...
> >What is the likelihood of an encryption system in which the resulting data
> >has no pattern,
> 
> Such a thing exists, it is called a one-time pad.  Generate a
> completely random key as long as the plaintext, and then XOR each
> successive bit of the key with the corresponding bit of the plaintext.
> The result is indistinguishable from random noise; only someone with
> an identical copy of the key can decrypt it (using precisely the same
> method of course).
> 
> > and one character of encrypted data may stand for many
> >different characters when decrypted?
> 
> Assuming you mean "one character of encrypted data might represent any
> one of several different characters of plaintext" (not "one
> character's worth of encrypted data represents multiple characters
> worth of plaintext), this is indeed the effect of a one-time pad.
> Just don't ever reuse that key; promptly destroy both copies after
> use.
> 
> --Mirian

This is hardly the place for a long discussion on such topics, but
one might want to look at the FAQ for the net news group sci.crypt.

    carl
-- 
    carl lowenstein         marine physical lab     u.c. san diego
                                                 clowenst at ucsd.edu

[TUHS] OT: Patternless Encryption

[Carl Lowenstein]

> From: "Joel Martinez" <president at coherent-logic.com>
> To: <tuhs at minnie.tuhs.org>
> Subject: Re: [TUHS] OT: Patternless Encryption
> Date: Tue, 8 Jul 2003 11:35:24 -0600
> 
> Is it possible to do this with a fixed length key?
> 
> > Such a thing exists, it is called a one-time pad.  Generate a
> > completely random key as long as the plaintext, and then XOR each
> > successive bit of the key with the corresponding bit of the plaintext.
> > The result is indistinguishable from random noise; only someone with
> > an identical copy of the key can decrypt it (using precisely the same
> > method of course).

For various degrees of security, depending on the length of the key.
Keys are not used directly for encryption, but are used to generate
cryptographically secure pseudo-random sequences.  

As a starting point, look at
    < http://www.mindspring.com/~schlafly/crypto/faq.htm >

    carl
-- 
    carl lowenstein         marine physical lab     u.c. san diego
                                                 clowenst at ucsd.edu

[TUHS] OT: Patternless Encryption

[Joel Martinez]

Is it possible to do this with a fixed length key?

> Such a thing exists, it is called a one-time pad.  Generate a
> completely random key as long as the plaintext, and then XOR each
> successive bit of the key with the corresponding bit of the plaintext.
> The result is indistinguishable from random noise; only someone with
> an identical copy of the key can decrypt it (using precisely the same
> method of course).
>
> --Mirian
> _______________________________________________
> TUHS mailing list
> TUHS at minnie.tuhs.org
> http://minnie.tuhs.org/mailman/listinfo/tuhs
> 

[TUHS] OT: Patternless Encryption

[list-tuhs]

On Tue, 08 Jul 2003 15:31:08 GMT, John P. Willis <jwillis at coherent-logic.com> wrote:
>
>Just curious to hear the opinions of the many wise people here...
>What is the likelihood of an encryption system in which the resulting data
>has no pattern,

Such a thing exists, it is called a one-time pad.  Generate a
completely random key as long as the plaintext, and then XOR each
successive bit of the key with the corresponding bit of the plaintext.
The result is indistinguishable from random noise; only someone with
an identical copy of the key can decrypt it (using precisely the same
method of course).

> and one character of encrypted data may stand for many
>different characters when decrypted?

Assuming you mean "one character of encrypted data might represent any
one of several different characters of plaintext" (not "one
character's worth of encrypted data represents multiple characters
worth of plaintext), this is indeed the effect of a one-time pad.
Just don't ever reuse that key; promptly destroy both copies after
use.

--Mirian

[TUHS] OT: Patternless Encryption

[John P. Willis]

Just curious to hear the opinions of the many wise people here...

What is the likelihood of an encryption system in which the resulting data
has no pattern, and one character of encrypted data may stand for many
different characters when decrypted?