🚀 go-pugleaf

In article <1992Jul8.171702.15889@ucsu.Colorado.EDU> cuffell@spot.Colorado.EDU (CUFFELL TIMOTHY MICH) writes:
>In article <1992Jul8.000947.12891@colorado.edu> bear@tigger.cs.Colorado.EDU (Bear Giles) writes:
>>In article <1992Jul7.212920.6452@colorado.edu> weverka@csn.org (Robert T. Weverka) writes:
>>>As I understand it the banks use code to transfer money, and these
>>>codes are about as secure as it gets.
>>
>>Money transfers between banks are encrypted with DES.  Some people
>>have concerns that DES has been compromised by the NSA, but that is
>>unproven.  (Not necessarily unfounded: IBM, the originator of DES,
> ^^^^^^^^
>
>Not true!!!!  Shamir (the 's' in RSA, the name of what is probably the most
>secure cryptological system to date)  has recently cracked a non-trivial
>special case of DES, and is close to a general solution.

As I understood it, his special case required a "choose-text" attack and
reduced the number of necessary from an obscenely large number to a number
merely ridiculous.  Something like from 2^56 to 2^48 tries.

That is _not_ close to a general solution.  (His approach would have
been close to a general solution except DES uses 16 iterations and
this approach only works for 15-or-fewer iterations.  The designers
of DES basically said: "We knew about that approach..."  This was
discussed on sci.crypt back in December, as I recall).

>While a computerized and secure election is possible, the logistical 
>consideration still make the old fashioned ballot box the best choice.

I agree.  DES is irrelevant anyway since the method I suggested on
sci.crypt/alt.security involves public-key encryption.


Bear Giles
bear@fsl.noaa.gov