Re: an scos2 test...

On 7/27/2024 12:20 AM, Richard Harnden wrote:

On 27/07/2024 01:20, Ben Bacarisse wrote:

"Chris M. Thomasson" <chris.m.thomasson.1@gmail.com> writes:
>

On 7/26/2024 2:32 AM, Richard Harnden wrote:

On 26/07/2024 06:15, Chris M. Thomasson wrote:

On 7/25/2024 8:54 PM, Rich wrote:

scos2 65 33
>
0ZGS XB sJ@ lH ~i<8/

>
For some reason I am getting a plaintext of:
>
~U9o My 8/g v[ Ym9\;
>
using my impl and Rich's original scos2 impl.

The key is wrong, try:
28 60
0ZGS XB sJ@ lH ~i<8/

>
Indeed it works like a charm. It had to be a "key issue". Humm... That
would be a fun test? Try different keys and log "readable" results wrt
decrypted plaintext? ;^)

>
Yes.  Somewhere I have a program that just tries all keys.  I was going
to get it to stop when English (or C) letter frequencies were found but
it turned out simpler just to eyeball the output.  As a human, you can
spot a decrypt a mile off and, if I remember correctly, SCOS has "close
decrypts" that I could spot but which would look, statistically, like
plaintext.
>

 I have one that tries to find consistent incr's in 3-letter word-blocks.   Hopefully the correct base and incr bubble to the top.
 It's better with longer texts, but ...
 $ cat test.scos
0ZGS XB sJ@ lH ~i<8/
 $ ./decode_scos test.scos | ./scos
./decode_scos: Not sure, but ...
 > -28 -60
 > Yes, it can be read.
  > -71 -60
 > =~Bu :C ]}, [~ A~}#s
 

The key space is not that massive, so it can be brute forced for sure. However, its interesting wrt the plaintext. What if the plaintext is generated from a random source?