|
|
 | | From: | enigma2005team at aol.com | | Subject: | New Code | | Date: | 23 Jan 2005 02:47:22 -0800 |
|
|
 | I'm co-inventor of what we hope is a new kind of code.
We've posted an example of encrypted text at www.enigma2005.com, and
would love it if people would come round and try to decode it. We've
even offered cash to the first person to crack the code.
Hope you have a go!
Martin Coleman
|
|
| | From: | infobahn | | Subject: | Re: New Code | | Date: | Sun, 23 Jan 2005 14:41:02 +0000 (UTC) |
|
|
| enigma2005team@aol.com wrote:
> I'm co-inventor of what we hope is a new kind of code.
>
> We've posted an example of encrypted text at www.enigma2005.com, and
> would love it if people would come round and try to decode it. We've
> even offered cash to the first person to crack the code.
> Hope you have a go!
Some observations:
Firstly, your competition offers no way for an entrant to
verify before submission that his decryption is correct.
It is not difficult to construct a cipher that, for a
given ciphertext, renders different plaintexts depending
on what key is used. AC2 gives some guidance on how to
do that.
Secondly, you rightly pay homage on your site to the
Bletchley Park people who cracked Enigma. Enigma was
designed to be secure even if the enemy had access to
"the machine". That is, the designer of Enigma was
realistic enough to understand that attackers would
not content themselves with ciphertext! In peacetime
an Enigma machine can be stolen. In wartime, it can
be captured. We wouldn't want to change our whole
way of communicating just because someone half-inched
a secret box from some insecure embassy or consulate
in Outer Wherever. So if you want your algorithm to
be taken seriously, you'll have to publish it. The
security should lie in the combination of a strong,
secret key and a strong, public algorithm.
Thirdly, a few observations on the cipher. I presume
the line breaks are not part of the cipher? If not,
then the cipher alphabet appears to be a base-36
encoding. (I am surprised you didn't go for base
64, which would probably have been easier to code up!)
For a ciphertext of the size you provided, I would
expect each character in your 36-character set to be
used approximately 70 times - i.e. 1/36th of the time.
In fact, one character is used three times as often
as that, and two more are used twice as often.
Here are all the characters that appear more than
1/36th of the time:
hex asc freq % (2dp truncated)
77 w 239 9.37
68 h 204 8.00
61 a 150 5.88
6B k 124 4.86
64 d 116 4.55
67 g 114 4.47
62 b 92 3.61
75 u 92 3.61
35 5 84 3.29
6D m 82 3.21
63 c 81 3.17
69 i 81 3.17
74 t 80 3.13
79 y 77 3.02
I suspect that a digraph analysis would also prove fruitful.
|
|
| | From: | John A. Malley | | Subject: | Re: New Code | | Date: | Sun, 23 Jan 2005 16:13:02 -0800 |
|
|
| enigma2005team@aol.com wrote:
> I'm co-inventor of what we hope is a new kind of code.
>
> We've posted an example of encrypted text at www.enigma2005.com, and
> would love it if people would come round and try to decode it. We've
> even offered cash to the first person to crack the code.
> Hope you have a go!
A lot of numbers appear in the ciphertext, and at first glance, they
might be preserving the spaces between plaintext words, sentences and
paragraphs. Single digits represent word spaces, double digits represent
sentence endings, and the relatively infrequent triple digits mark new
paragraphs.
A quick check of the Gutenberg web site's recent top 100 downloads and I
see James Joyce's "Ulysses." That novel had sentences like the patterns
I see in the ciphertext, but again, this is just a hunch on what to rule
in/out on the way to identifying the system used.
But that's what pops into mind with 5 -10 minutes of examining the
ciphertext and its context (i.e. the web site from which the text came,
and which texts get downloaded most frequently. )
John A. Malley
102667.2235@compuserve.com
|
|
| | From: | Morten Dahl | | Subject: | Re: New Code | | Date: | Sun, 23 Jan 2005 15:29:13 +0100 |
|
|
| enigma2005team@aol.com wrote:
> I'm co-inventor of what we hope is a new kind of code.
>
> We've posted an example of encrypted text at www.enigma2005.com, and
> would love it if people would come round and try to decode it. We've
> even offered cash to the first person to crack the code.
> Hope you have a go!
>
> Martin Coleman
>
Just a note on your challenge: for a better result you should follow
Kerckhoff's principle, see
http://en.wikipedia.org/wiki/Kerckhoffs'_principle
|
|
| | From: | Gianna Stefani | | Subject: | Re: New Code | | Date: | Sun, 23 Jan 2005 13:19:37 +0000 |
|
|
| enigma2005team@aol.com wrote:
> I'm co-inventor of what we hope is a new kind of code.
>
> We've posted an example of encrypted text at www.enigma2005.com, and
> would love it if people would come round and try to decode it. We've
> even offered cash to the first person to crack the code.
> Hope you have a go!
>
> Martin Coleman
>
I took a quick look at your cipher.
I wonder did you try frequency analysis and sequence analysis on the
text before publishing?
--
Gianna Stefani
|
|
|
