[sci.crypt] Vendor's Responsibilities on seized equipments

brnstnd@kramden.acf.nyu.edu (Dan Bernstein) (01/11/91)

In article <1991Jan10.041906.27901@ddsw1.MCS.COM> karl@ddsw1.MCS.COM (Karl Denninger) writes:
> A key is used to encipher each byte.  After each byte is ciphered, the key 
> is permuted, and depending on the results of that permutation a random number
> of bytes of "garbage" (random data) from 0-n is added to the output file.

I assume you do this after compression so that the original data doesn't
have any immediately noticeable redundancy. Then this system is secure
if and only if the ``random number'' is cryptographically secure and the
``garbage'' is both random and independent of the key. I find the first
assumption exceedingly unlikely.

---Dan

spm2d@newton.acc.Virginia.EDU (Steven P. Miale) (01/15/91)

In article <1991Jan10.041906.27901@ddsw1.MCS.COM> karl@ddsw1.MCS.COM (Karl Denninger) writes:
>In article <13899@uudell.dell.com> sblair@upurbmw.dell.com (Steve Blair) writes:
>I'd chuckle, giggle and snicker - and then give them the source to the 
>program.  First and foremost, I doubt if even the best computer they have 
>right now would help, given that the user of the cryptosystem in question 
>used a reasonable key on the file(s) in question.

Actually, I have another method a bit similar to the one you used:

To each character in the file, a character from the password is added. Then,
the number of times the encryption program has gone through the password
is subtracted. Add to this the character number in the password, and
XOR it with the CRC of the password. Makes for lots of fun. I'm not
sure of the *actual* secureness of this algorithm, and I would be pleased
if some of the code experts out there would tell me if it is actually secure
or if it could be easily broken. My program accepts passwords of any
length, and most of mine are over 20 characters...

Steven Miale
spm2d@virginia.edu
 

jmc@DEC-Lite.Stanford.EDU (John McCarthy) (01/15/91)

I'd use DES even if I believed NSA could break it.  They wouldn't
reveal an ability to break DES in order to help the Secret Service
convict a two bit Sysop.  On the other hand, if you're a spy, ask
the Soviets for a suitable cipher.