Code Breaking

[dedndave]

i see where a few guys cracked this old cypher
http://www.sciencedaily.com/releases/2011/10/111025102320.htm



which is pretty cool

it caught my interest...
i followed some links and came across this page on the FBI site
they are looking for public help in cracking this code
http://www.fbi.gov/news/stories/2011/march/cryptanalysis_032911




i am a little surprised that they were unable to crack it
i see a few patterns repeated several times (NCBE, PRSE, and others)
i also noticed a lot of E's - and that nearly all lines end in E
and, i noticed that the E's are often paired with only a few letters (BE, NE, RE, SE)

the guy wasn't particularly well educated, although that doesn't mean he wasn't smart
the text appears to have been written in some haste
what i mean is, he probably generates the code from an algorithm in his head

they give another link if you have a solution
http://forms.fbi.gov/code

[TmX]

So it's not cracked yet?
I'm not sure what makes it so difficult, because they are just plain alphabets/numbers, and not alien symbols.
Honestly I know practically nothing about cryptanalysis, though.

[clive]

The 'E' terminate words.
Each line stands on it's own.
The second looks like a recipe, or directions. You'd need to see how the location related to possible landmarks.
It looks to be written in a non-linear fashion, ie the writer moved back and forth.

[dedndave]

yes - that was my first thought - that E's were spaces
or - the last letter (or first letter if reversed) forms the key for that word
that might cause the first or last letter of a word to always be E

as for back and forth - that would make it difficult to do
and hand-writing analysis would have revealed it
i thought about backwards   :P

the parens, hyphens, and circled areas may be important to figuring out what type of document it is
i am not sure that's going to be much help breaking it

[dedndave]

Honestly I know practically nothing about cryptanalysis, though.

that may be an advantage in this case
the FBI people are probably some of the best
they wouldn't publicize it if it was "standard" stuff

as assembly language programmers, we tend to think in terms of algorithms
this may be a little "outside the box" for someone trained in crypto   :U
that's why i posted it - there are some sharp minds here that, if nothing else, might offer some ideas

[clive]

Some of the R's and E's look different.

One interesting thing that's hard to see with this, and that would be useful, is what characters have been replaced/corrected. If there are some consistent errors it could shed some light on the mechanism of the cipher.

You're right though Dave, this is being done in the head.

[dedndave]

yes - that implies it is reasonably simple
however, he may be using his mother's middle name or his girlfriends phone number to form keys   :P
the FBI doesn't want to give out a lot of info - and there's the challenge

it would have been nice if they had shared the results of hand-writing analysis, though   :P


i might add - if someone solves this, it is probably best not to post the results - or the method used
at least, until the case is closed

[vanjast]

It seems..
- As if he was calm and calculating when he wrote this.
- He had a well organised, structured (if not mathematical) mind
- He looked educated (maybe not book smart) but educated.

Looks interesting...
:8)

[Ficko]

I just saw this very interresting but have no time to tuckle with it. :toothy

But I would approach it like a "password cracker" does for ZIP or ARJ.

If you manage to get just one word right the algo can be reversed relativelly easily.

1/2 MUNDDLSE

sound like 1/2 miles that's can be taken as the anchor.

[Farabi]

Yeah, just check some of the pattern, if it is repeated a few times and it is consistens, yes it is, it is readable. I saw that most of human manuscript is having a consistent pattern, except for some of the ancient asian manuscript, it is completely unreadable, since it is not a text for me, obviusly, it is a picture, even it should have a meaning. But since there is no clue not for single word, I cant understand it. It is different with Hebrew, Greek and arabic, the pattern is consistent. And After I checked how the prononciacion, youll surprised with many of the similiarity. Yeah, we used to talked on the same language thousand years ago, and used one currency, GOLD.

[Farabi]

http://www.newscientist.com/article/mg21228415.500-inside-the-race-to-crack-the-worlds-hardest-puzzle.html?

Interesting. I had something on my mind. Line up the lines, and since it was printed using a printer, the font size and type got to be constant. We can separate each piece and line it into a linier arrangement. By using an AR system like OpenCV we can get which line up set which had a meaning words. Some document got to be had a same lines on each pages, it cannot be wrong.

[jj2007]

Dave,

When you are ready with the toys above, go here. They have a job for you.

[Farabi]

Dave,

When you are ready with the toys above, go here. They have a job for you.

It look familiar, "EB 04" did not it jmp 4 bytes?
Craps, that look like a hex code of a dechipher,  :green

[fearless]

GCHQ CanYouCrackIt Solution explained:

http://gchqchallenge.blogspot.com/

[dedndave]

i was looking at that
it appears to be some sort of obfuscated code
you could probably make an EXE and stick that in the .CODE section with DB's
however - run it at your own risk

one line of code i saw in there was...

Code Select Expand

BA EF BE AD DE              mov edx, 0DEADBEEFh