# Cracking WWII-era codes - code found on a pigeon's leg in Surrey

A recent BBC article entitled WWII code 'may never be cracked' posted a code:

AOAKN HVPKD FNFJW YIDDC
RQXSR DJHFP GOVFN MIAPX
PABUZ WYYNP CMPNW HJRZH
NLXKG MEMKK ONOIB AKEEQ
WAOTA RBQRH DJOFM TPZEH
LKXGH RGGHT JRZCQ FNKTQ
KLDTS FQIRW AOAKN 27 1525/6

According to the article the code was found on the leg of a pigeon in a chimney in Surrey a few weeks ago.

Do we have enough information about WWII-era codes (presuming that is when it comes from), such as the one above, to figure out what these long-forgotten codes might say?

EDIT: This was discussed on slashdot.

-
An article on BBC indicates that someone may have successfully cracked the code. –  Brian M. Hunt Dec 17 '12 at 14:03

NOTHING IS UNCRACKABLE! It may take a mind a thousand years, or a computer 500. But as long as there are people with curiosity, this will be cracked.

-
As others will undoubtedly point out, if we do not know what the message says (or contains) we have no way of proving or even indicating what the decrypted message is. It is likely that we will have a set of statistically probable decryptions, but that it will remain impossible to verify correctness. –  Brian M. Hunt Dec 17 '12 at 1:11
Some things are "uncrackable" even with infinite computing power. –  Thomas Dec 17 '12 at 1:44

In principle even a one time pad isn't unbreakable if the pad is a short sequence that is repeated. If that were the case then a brute force search comparing each output message with the statistical likelihood of that message existing in English should yield an answer.

Seems that we really need to know more about the encryption methods that may have been used to produce this message.

Even if it is a one time pad and one that that is not repeated, if the pad was taken from a commonly available book (say) then once again a brute force search may be possible (in principle) using the same maximum likelihood approach. The only completely hopeless case is if the one-time-pad is not repeated and no longer exists anywhere.

-
If the pad is a short sequence that is repeated, or if it's taken from a book, then it's not a one-time pad. –  Thomas Nov 27 '12 at 4:17
Sure, I'll improve the wording accordingly. The point remains that a random pad, if repeated, may be breakable. Though in this case the message is very short, therefore it's very likely to have been covered by a single pad. –  locster Nov 27 '12 at 9:23
The definition of the one-time-pad is that the pad is exactly as long as the message. The pad cannot repeat (that is the basis of the security proof of the one-time-pad). On the other hand, if the same pad is used twice to encrypt two different messages, then yes, you are in trouble. Used properly (which includes: do not reuse the pad), the one-time-pad is unconditionally secure - it cannot be "broken" without knowledge of the pad (which is fully random and as long as the message it encrypts). So unless you can find the pad scribbled somewhere in some german bunker, no, you are screwed. –  Thomas Nov 27 '12 at 10:09
I'm agreeing with you. If you take the strict definition of one-time-pad then yes, but we don't know that it is a one time pad by that definition, that's essentially the thought train I'm following here. We don't know for sure that it's a random pad at all in fact - it's highly likely but not certain. –  locster Nov 27 '12 at 10:38
Well in that case, it's not a "one-time-pad" anymore. I see your point though. It could be anything since we don't know which encryption algorithm was used. –  Thomas Nov 27 '12 at 10:43

A problem with this message is that the image in a press release from GCHQ differs slightly from one in a New York Times article, although the paper, with its tears looks the same.

The image below shows the two side by side, the one on the left from the NY Times and the right from GCHQ. There is a slight difference in the style of writing, i.e. more sloping to the right in the GCHQ image. There are also some word differences: word 7 is GOVFN from GCHQ and FOVFN in the other.

That the message was found in Bletchingley, Surrey may be a bigger clue.

-
There are several scans of the message floating around, with varying original quality and postprocessing. Looking at the Times version you show, it's pretty clear that someone took a poor quality scan or photo and tried to make it clearer by overdrawing some of the letters; you can still see the original text in the background if you look closely. The new GCHQ scan has much better quality (though still not perfect), and most importantly is in color. –  Ilmari Karonen Nov 25 '12 at 19:28
However, you may be onto something with the G/F thing: for example, the BBC transcript quoted in the question gives the 24th group as "FQIRU", but comparing the first letter with the other Fs and Gs in the rest of the message makes it pretty clear that it should be "GQIRU" instead. Similarly, the 17th group should probably be "UAOTA", not "WAOTA". –  Ilmari Karonen Nov 25 '12 at 19:33
Just how long has this message been floating around? CNN and others say it was discovered 'recently' whereas the NYT article says it was found in 1982! –  SPA Nov 27 '12 at 17:28

I believe that the key lies in the final 6 number 1525/6. In this period the German Mathematician Albrecht DÜRER published "The Four Books on Measurement" the third of which picks up on the geometric construction of the latin alphabet. Albrecht DÜRER was also famous for a magic square which is the same as Sudoko puzzles and would therefore have the intelligence and ability to work with alphnumerical codes. http://www.taliscope.com/Durer_en.html It may be a long shot, however it also seems very coincidental. It may be that the key to the code is one of his books and that 27 is a page reference to one of is publications such as "The Four Books on Measurement" which were released in 1525. Not a solid answer but a path worth following.

-
27 is also the number of blocks of five-letter code words –  SPA Nov 25 '12 at 11:38