Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?
Encryption Security

Please Patiently Ponder Purported Poe Puzzle 195

grouchomarxist writes: "Salon has an article about a cryptograph attributed to a certain W. B. Tyler, possibly a pseudonym for Edgar Allen Poe. There is a $2500 prize for the person who solves the cryptograph." The Gold-Bug , which rates a mention in the Salon article, was by far the most spell-binding story in my old Horace Mann Reader, and it's the tale that first turned me on to The Divine Edgar. Could it be that the reason this cryptograph has remained unsolved for so long is that it is actually insoluble? Now that would be the ultimate posthumous practical joke. Even if you have no intention of trying to solve it, take a look -- the cryptograph itself is strangely hypnotic.
This discussion has been archived. No new comments can be posted.

Please Patiently Ponder Purported Poe Puzzle

Comments Filter:
  • ...by someone with a lot of time on their hands. Anyone over at Andover wanna give it a shot?
  • Okay, replace the most repeated letter with 'e' & see whatcha got. Work from there. I want a cut :)
  • I'm sure somebody will probably figure it out in the next 12 hours. Won't be me since I'll be asleep.
  • by BigTed ( 78942 ) on Wednesday March 08, 2000 @11:58PM (#1215912)
    An exam I took last year - Stage Three Electrical Engineering :-)
  • by DgtlGhost ( 155814 ) on Thursday March 09, 2000 @12:00AM (#1215913)
    Anyone else get the idea that MAYBE this thing is up-side-down, or backwards? If it is Poe's, then that would make the Salutation a signature instead. Just a thought. I'm not a cryptographer, but I've watched them on Tv...
    Hey, new open source project, FreePoe! Set this thing up on Distibuted and lets get cracking! Damn, it's late.


  • I would hope that Poe/Tyler would be more creative than simple letter replacement. I know that I came up with that one in grade school, so hopefully an intelligent person with an interest in cryptography would try to do a slightly better job than that! =)

    A more light suggestion... maybe we should get D.net involved with this! =) I still have a few CPU cycles to spare!
  • There's 105 letters in this code (upper + lower case upside down and right way up + spaces)

    But back to your point, for most simple substitution codes, you can also use the rule that t is the second most repeated letter, and most of the time "t?e" maps on to "the". Doesn't everyone know this?
  • by telematx ( 17553 ) on Thursday March 09, 2000 @12:08AM (#1215916)
    Or perhaps two messages in one, since some letters go one way and others another. Or maybe it is one message, split in the middle and looped back on itself.

    Someone should split it in two and run frequency counts on each section.
  • by Lionfire ( 103856 ) on Thursday March 09, 2000 @12:22AM (#1215921) Homepage Journal
    Just a thought -- perhaps the first puzzle, which was apparently reasonably easy to solve, is a clue to this second puzzle?

    It could be used as a form of "key" to solve the second...

    Or maybe that's just my sick twisted mind enjoying the idea of having people struggle to understand something left behind for 150 years while the clues are sitting right there on the same page :)

  • by luckykaa ( 134517 ) on Thursday March 09, 2000 @12:30AM (#1215925)
    My syntax here is "!" -> Upside down

    !TIA is repeated at least twice
    !A!mL is repeated at least twice
    These are probably "The" and "And"

    !i!rz and !irz are both in there

    b, !b and K are the only symbols on their own.
    Multiple representations of "I" and "A"?
  • Things like that seem just a little too obvious. That is a good way to solve those little cryptograms in the daily newspaper. But, it is a great idea to get some thought flowing. But $2500 is a lot of money to give away if this little beauty was that easy.

  • by Anonymous Coward
    It is obviously a multiply interleaved boustrophedonic text. If you don't know what that means, you have no hope of solving it. I could probably solve the problem given a week or so of hard work, its fairly obvious just from looking at the typesetting as to how it should be solved. But alas, I don't have the time at the moment. The solution is fairly trivial.
  • by Anonymous Coward on Thursday March 09, 2000 @12:51AM (#1215928)
    It says,
    I poured hot grits down my pants while Natalie Portman was petrified and naked with ninjas and pancakes.


    Way ahead of his time, Mr. Poe was.
  • by Red Pointy Tail ( 127601 ) on Thursday March 09, 2000 @12:59AM (#1215933)
    My guess is that it is a letter substitution plus translations, governed by the characteristics like Big/Small, Capital/Not, Reversed/Not.

    I bet 100+ years ago they didn't have networked supercomputers like we did now, so it should be a cinch. So someone with access the juice, please key in the schema and churn

    1. Define degrees of freedom schema
    a) Capital or not (0,1)
    b) Big, Small (0,1)
    c) Reversed, or not (0,1)
    d) The letter (1..26)

    2. Key in the data in this schema

    3. Run the damn thing
    Using a standard dictionary substitution methods for the letters d), using various translations for a), b) and c).
  • Shouldn't be that hard to solve the puzzle. If it's not solved in 12 hours then it's a fake in my Opinion. I don't solve it but I think there are very good people out there on the net with a lot of time for such games.
  • by Anonymous Coward on Thursday March 09, 2000 @01:00AM (#1215935)
    There's 105 letters in this code (upper + lower case upside down and right way up + spaces)

    You missed something obvious. The puzzle contains 6 alphabets. It uses normal and inverted instances of Upper case, lower case, and small caps.
    There are 156 possible letters in those 6 script systems altogether (157 including the space char). I haven't counted the letters in the ciphertext, but if there are only 105 symbols in the original text, there would not be enough repeats to form any obvious line of attack. It would all be brute force decoding based on an understanding of the encoding method. Good luck.
  • by Anonymous Coward
    For example, if I defined the letter A to mean "Please Patiently Ponder Purported Poe Puzzle", you'd have no way of telling what the encrypted code A means unless I told you. The cryptographic method might not be that strong, but there just isn't enough data to be decrypted. I wouldn't waste my time with this.
  • You're kidding about d.net right?

    All the cpu cycles in the world won't help if you don't know how the cypher was encoded in the first place...Well, assuming the cycles aren't used for some as-of-yet-unknown artificial intelligence.

  • If some Slashdot reader happens to solve this, why not donate some of that money to the OSS community? Or at least to projects that support strong cryptography :P

  • There's 105 letters in this code (upper + lower case upside down and right way up + spaces)

    He also uses 3 or 4 different letter sizes. Propably only obfuscation, but you never know... (onless you solve it that is ;)

  • Look at the occurence B and !B (! is reversal, both small size)

    ... probably a and A.

    I think it is a direct mapping, i.e. If X -> A then X! -> a or if X -> a then X! -> A.

    If you look, you will see all the Bs occuring in midword (where it can't possibly be capital) is B and not B!.

    Now lets figure out what big/small, capital/small means....
  • whoops, wouldn't work.

    i!Rz and iRz are both on. back to the drawing board....
  • seems to me this would suggest a word that is three letters, except the center letter is change, anyone know of three letter words were the center letter changes and still makes sense?
  • by Yardley ( 135408 ) on Thursday March 09, 2000 @01:30AM (#1215945) Homepage
    I have assuredly found an admirable resolution to this, but the margin is too narrow to contain it.

    And perhaps, posterity will thank me for having shown it that the ancients did not know everything.
  • Uou're kidding about d.net right?

    Yeah, I was kidding. But you never know, it could be encoded using DES-III. That Poe fellow was ahead of his time! =)
  • Someone had moderated it as "Interesting"! That is pretty funny.

  • Don't discount the possibility that the inversion of letters is just obfuscation, or maybe it means that you should swap with another letter in the word. Tyler's First Cryptograph suggests that the writer likes scrambling as well as substitution.
  • by theGamer ( 137984 ) on Thursday March 09, 2000 @01:46AM (#1215951)
    and !W!c
    are all repeated twice.

    as for subtile changes there's:
    !dhb and dhb
    !irz and !i!rz
    (all lowercase, hmmm....)

    I have been staring at this damn thing for an hour and it baffles me.

    A few thoughts: could some of the characters be earmarks for jumping around between? Perhaps caps or upside-down indicates that the value of the letter should be added or subtracted.

  • ...and although I wouldn't want to call Poe 'twisted' as such, he was definitely, well, shall we say 'interesting' ? :)


    # Using Linux in the UK? Check out Linux UK [linuxuk.co.uk]

  • No, simply because it's often difficult to tell a good troll from someone who's serious, and of course one's person't funniest joke ever is another's yawn.
  • It looks to me as if the case and orientation don't matter at all. I think whoever made it may have just thrown those in there to create another step or two for the solving. Not many people can read at different angles like that. Since it is english it would be intuitive to try to read it like a paragraph of prose. It has occured to me that the message may be backwards or that it could even be a "jumble" where you have to unscramble all of the words. I don't think this is the case at all. The O'rt' appears twice from my breif look, and in the same order, and there are no other 3 letter groupings with the same letters. Damn this is fun!
  • Although this isn't completely on topic, you might want to check out http://members.aol.com/s6sj7gt/mikerav.htm

    This is a transformation of the poem "The Raven" converted so the number of letters in each word match the digits in PI. Talk about people with too much time on their hands. :)

  • by doomy ( 7461 ) on Thursday March 09, 2000 @01:56AM (#1215956) Homepage Journal
    Maybe we are looking at the puzzle in the wrong way. I've been doing a lot of 3d work lately and the first time I saw this [bokler.com] puzzle, I felt that it had something to do with a depth.

    Look at how some letters extrude out and some go in and some lie in most unusual manner. Also see the cases. There are fonts uses of varying sizes. This gives a very strong feel for depth. The first two letters. Cap D and R with the R being half the size of D, pushing it in. Then the mix of cases and the big OGXEW right up front.

    If someone could lay this puzzle in a 3d representation, I'm pretty sure we a solution would open up.
  • And everyone is automatically assuming it is written in english. Oh my, oh my.

    Well that is what you get being all alone out there far away all by yourself and so on...

  • Why 'the' or 'and', why not 'Poe', it's a letter adressed to him, isn't it?
  • by Anonymous Coward
    It's a message from their server saying 'Please don't /. me'
  • The NSA seems to be a great place to help with this. Great PR-stunt for them too. And since they cost a ton of money a year it might be nice to see that they actually do have the power to solve this. In the end this doesn't come down to the money but to the honour of being the first.

  • by cybrpnk ( 94636 ) on Thursday March 09, 2000 @02:10AM (#1215961)
    Poe IS cool but the REAL 19th century code mystery is called the Beale cypher. This guy sent the letter below and three code sheets to a friend. Beale's codes were based on documents available in the early 1800s - one was the Declaration of Independence. One code sheet talked about a vault of gold in Bedford County (Virginia, I think) when decoded, and the second listed who got what share of it (not reproduced here). The final code sheet with the gold's location, reproduced below, HAS NEVER BEEN BROKEN....

    Details at:

    http://treasurehunt.miningco.com/hobbies/treasur ehunt/bllet.htm

    The Original Uncoded Letter From Beale

    St. Louis, Mo., May 9th, 1822.

    Robt. Morris, Esq.:

    My Esteemed Friend: - Ever since leaving my comfortable quarters at your house I have been journeying to this place, and only succeeded in reaching it yesterday. I have had altogether a pleasant time, the weather being fine and the atmosphere bracing. I shall remain here a week or ten days longer, then "ho" for the plains, to hunt the buffalo and encounter the savage grizzlies. How long I may be absent I cannot now determine, certainly no less than two years, perhaps longer.

    With regard to the box left in your charge, I have a few words to say, and, if you will permit me, give you some instructions concerning it. It contains papers vitally affecting the fortunes of myself and many others engaged in business with me, and in the event of my death, its loss might be irreparable. You will, therefore, see the necessity of guarding it with vigilance and care to prevent so great a catastrophe. It also contains some letters addressed to yourself, and which will be necessary to enlighten you concerning the business in which we are engaged. Should none of us ever return you will please preserve carefully the box for the period of ten years from the date of this letter, and if I, or no one with authority from me during that time demands its restoration, you will open it, which can be done by removing the lock. You will find, in addition to the papers addressed to you, other papers which will be unintelligible without the aid of a key to assist you. Such a key I have left in the hands of a friend in this place, sealed, addressed to yourself, and endorsed not to be delivered until June, 1832. By means of this you will understand fully all you will be required to do.

    I know you will cheerfully comply with my request, thus adding to the many obligations under which you have already placed me. In the meantime, should death or sickness happen to you, to which all are liable, please select from among your friends some one worthy, and to him hand this letter, and to him delegate your authority. I have been thus particular in my instructions, in consequence of the somewhat perilous enterprise in which we are engaged, but trust we shall meet long ere the time expires, and so save you this trouble. Be the result what it may, however, the game is worth the candle, and we will play it to the end. With kindest wishes for your most excellent wife, compliments to the ladies, a good word to enquiring friends, if there be any, and assurances of my highest esteem for yourself, I remain as ever,

    Your sincere friend, T.J.B.

    Beale Code Page II - (This is the decrypt; the original is on the web page above)

    I have deposited in the county of Bedford, about four miles from Buford's, in an excavation or vault, six feet below the surface of the ground, the following articles, belonging jointly to the parties whose names are given in number "3," herewith:

    The first deposit consisted of one thousand and fourteen pounds of gold, and three thousand eight hundred and twelve pounds of silver, deposited November, 1819. The second was made December, 1821, and consisted of nineteen hundred and seven pounds of gold, and twelve hundred and eighty-eight pounds of silver; also jewels, obtained in St. Louis in exchange for silver to save transportation, and valued at $13,000.

    The above is securely packed in iron pots, with iron covers. The vault is roughly lined with stone, and the vessels rest on solid stone, and are covered with others. Paper number "1" describes the exact locality of the vault so that no difficulty will be had in finding it.

    THE LOCALITY OF THE VAULT - (This has never been decoded)

    71,194,38,1701,89,76,11,83,1629,48,94,63,132,16, 111,95,84,341.
    975,14,40,64,27,81,139,213,63,90,1120,8,15,3,126 ,2018,40,74.
    758,485,604,230,436,664,582,150,251,284,308,231, 124,211,486,225.
    401,370,11,101,305,139,189,17,33,88,208,193,145, 1,94,73,416.
    918,263,28,500,538,356,117,136,219,27,176,130,10 ,460,25,485,18.
    436,65,84,200,283,118,320,138,36,416,280,15,71,2 24,961,44,16,401.
    39,88,61,304,12,21,24,283,134,92,63,246,486,682, 7,219,184,360,780.
    18,64,463,474,131,160,79,73,440,95,18,64,581,34, 69,128,367,460,17.
    81,12,103,820,62,116,97,103,862,70,60,1317,471,5 40,208,121,890.
    346,36,150,59,568,614,13,120,63,219,812,2160,178 0,99,35,18,21,136.
    872,15,28,170,88,4,30,44,112,18,147,436,195,320, 37,122,113,6,140.
    8,120,305,42,58,461,44,106,301,13,408,680,93,86, 116,530,82,568,9.
    102,38,416,89,71,216,728,965,818,2,38,121,195,14 ,326,148,234,18.
    55,131,234,361,824,5,81,623,48,961,19,26,33,10,1 101,365,92,88,181.
    275,346,201,206,86,36,219,324,829,840,64,326,19, 48,122,85,216,284.
    919,861,326,985,233,64,68,232,431,960,50,29,81,2 16,321,603,14,612.
    81,360,36,51,62,194,78,60,200,314,676,112,4,28,1 8,61,136,247,819.
    921,1060,464,895,10,6,66,119,38,41,49,602,423,96 2,302,294,875,78.
    14,23,111,109,62,31,501,823,216,280,34,24,150,10 00,162,286,19,21.
    17,340,19,242,31,86,234,140,607,115,33,191,67,10 4,86,52,88,16,80.
    121,67,95,122,216,548,96,11,201,77,364,218,65,66 7,890,236,154,211.
    10,98,34,119,56,216,119,71,218,1164,1496,1817,51 ,39,210,36,3,19.
    540,232,22,141,617,84,290,80,46,207,411,150,29,3 8,46,172,85,194.
    39,261,543,897,624,18,212,416,127,931,19,4,63,96 ,12,101,418,16,140.
    230,460,538,19,27,88,612,1431,90,716,275,74,83,1 1,426,89,72,84.
    1300,1706,814,221,132,40,102,34,868,975,1101,84, 16,79,23,16,81,122.
    324,403,912,227,936,447,55,86,34,43,212,107,96,3 14,264,1065,323.
    428,601,203,124,95,216,814,2906,654,820,2,301,11 2,176,213,71,87,96.

  • The post I'm replying to started at 0 and went down to -1 because it said "Why isn't slashdot running a story on OpenAL [openal.org]? Is this a MS conspiracy?"

    However, I checked out the link to OpenAL. It looks interesting. Excerpts:

    OpenAL, the Open Audio Library, seeks to become the audio counterpart to OpenGL for audio. With OpenAL it is possible to create three-dimensional sound across many platforms, such as Linux, Apple Macintosh, Windows and more, with quality suitable for professional projects like games and multimedia applications.
    OpenAL is being sponsored by commercial software and hardware companies to help make open source work. Here is a current list of the member and partner companies:
    • Creative Labs
    • Loki Entertainment Software
    But you gotta wonder if these people are serious whan they name part of their library "Al-Core".

    This certainly is offtopic here, but in case the AC is right and Slashdot is not posting this for some sinister reason, I figured I'd make it seen using my +1 bonus.

    Still, there's no reason to assume the worst; we don't even know if this has been submitted as a story, so don't call me a paranoid conspiracy theorist or anything :-)

  • I read Slashdot with several stories in different windows. When I switched to this one, I got it confused with the previous story's comments.

    That must be the answer. The puzzle is an example sendmail.cf.
  • by Anonymous Coward
    Aren't some of the spaces upside down too?
  • To encyrpt implies the possibility of decryption; unless this guy new that over a hundred years later, several thousand idiots would still be trying to solve that which never meant anything ;)
  • ...37734 5P33|!!! So all we need is to get a skr1pt k1dd33 to decode it for us!
  • by Anonymous Coward on Thursday March 09, 2000 @03:41AM (#1215969)
    No, simply because it's often difficult to tell a good troll from someone who's serious..
    Yeah, well I was serious. Go look up "boustrophedon" at britannica.com if you think I just made it up. Poe was a clever bastard, and his text obviously contains clues to its own solution as a means to humiliate the solver so they'd see how OBVIOUS the clues were once it was solved. Remember, this was Poe's Parthian shot.
    The text runs in two directions, suggesting the regular characters are one text stream, and the upside-down characters are running back the other direction, interleaved with each other. Each type font (upper case, lower case, and small caps, plus inversions) could be a separate text stream, possibly there are 6 different interleaved streams, running in boustrophedonic fashion. I could work it out by brute force, but I don't have time. Good crypto cracking takes time, time I haven't got.
    So go ahead and label it a troll, it only shows your own ignorance. Maybe if you'd taken a few liberal arts and humanities courses in college, you'd know about this stuff.
  • That's always a possibility, but in order to encode a puzzle in such a manner, the author would have to have a pretty good understanding of 3D perception and transformation, right?
    The question is: would an accomplished LITERARY genius necessarily have good spacial comprehension? If so, then it would be a trait nobody had known about until now.
    If not, maybe the puzzle is an attempt at 3D encoding, but not a very good one, so it would be easy to crack. Or utterly impossible if he didn't do it right!
  • Like this maybe-Poe cryptograph, though, the Beale has also never been solved, or at least not provably so. Some hunters have said that they located the destination (claiming this or that cave, well, hollow under West-Virginian bar, etc.).

    The most convincing report I've heard says that the Beale cipher was intentionally broken. That is, the first two sheets *were* soluble, with some work, but that's only a tease to make the 3rd one that much more frustrating.

    And since I haven't solved it, I'm inclined to bitterly agree.:)
  • Since I'll probably not spend much time working on this myself (don't want to get fired from my real job that pays the bills), I'll throw out my first thoughts about the cryp.

    Perhaps the most obvious observation is that there are really four different sets of characters in the message:
    right-side-up UPPER CASE
    right-side-up lower case
    up-side-down UPPER CASE
    up-side-down lower case

    If the message isn't obsenely difficult to decode (as with the first message) then there are probably simply four different sets of rules to use with each of the four character types.

    A starting place might be to begin with shorter words or common repetitions of characters and begin there.

    Simple character replacement isn't all that difficult to decode, but character replacement when you've got 4 times as many encodings could be much more difficult. For instance, the upside "R" and "q" and rightside up "l" and "B" could all be the same letter for instance, making it nearly impossible to identify patterns or repeting characters. Someone want to make some estimates as to the total number of possible encodings?

    Is it (26!)^4 or something crazy like that?

    (26!)^4 = 2.65 x 10^106

  • Flamebait?!
  • But he certainly looks like he's trying to be 31337 here...
  • I have discovered a truly remarkable solution to this cypher which, unfortunately, is too large to fit in the available bandwidth to this server...

  • Another thing to consider is that Poe had to write all of the right-side-up letters first, leaving blanks for the upside-down letters. Then he had to take the paper out of the typewriter and write the upside-down letters. You can notice how he didn't quite align everything properly between the 3rd and 2nd to last letters on the second line (a small-caps u and a upside-down/backwards lower-case c).
    I would vote against the guy who said to separate the letters by their type (like a 3d puzzle) because Poe wouldn't have gone to all that trouble just to have the letters be easily separated.
    Wouldn't it be naughty if the orientation/case was meant to mean something like the following:

    LC: leave in current position
    UC: Move one space away.
    SmallCap: Move two spaces away.
    Orientation Normal: Move to the right.
    Orientation Reversed: Move to the left.

    After you do all that, then you are left with a regular cipher.
  • Hey! Look at that poe cipher, row 7 from the bottom, slightly right of the middle. It says 'eToY'!

    Darn. Just as one dispute was settled, the next begins. Now etoys.com will have to sue Poe postumly for using something that almost looks like their company name... :)

  • Could it be that the reason this cryptograph has remained unsolved for so long is that it is actually insoluble?

    But of course! The molecular structure of a cryptogram is polar, meaning that it is insoluable in water. However, if you stir it up enough, it might look like it dissoved.

    Of course, that still doesn't mean it insolvable. ;)

  • Since you haven't solved anything ... why don't YOU give some money to the OSS community?

    At least, he who will solve it will deserve it ...

  • That last paragraph is the thing to remember, certainly.

    But you got moderated down to Flamebait for saying THAT?

    Wow, I hope this one comes up in my M2.

  • Dudes, this man has something to say. And he could be onto something...
  • Perhaps the most obvious observation is that there are really four different sets of characters in the message:
    right-side-up UPPER CASE
    right-side-up lower case
    up-side-down UPPER CASE
    up-side-down lower case

    Actually there are more letters than that:
    I presume that with upside down you mean rotated 180 degrees. So that leaves us with:
    -The four you mention
    -small UPPERCASE
    -Also possible is upside down

    This last one is especially sneaky with symmetric letters, but you may notice that in for example in the letter A the right leg is thicker than the left leg. I haven't found any upside down letters that are non symmetric, but there are some A's that could be upside down

    On a different note, what also strikes me is that only capitols (small and large) are used in the To Edgar part. Maybe this is a clue?

  • I repost the appropriate, learnèd and interesting post from our fellow A.C. I'll probably get moderated down, but anyway.

    A Hint:

    It is obviously a multiply interleaved boustrophe donic [britannica.com] text. If you don't know what that means, you have no hope of solving it. I could probably solve the problem given a week or so of hard work, its fairly obvious just from looking at the typesetting as to how it should be solved. But alas, I don't have the time at the moment. The solution is fairly trivial.

  • It's Textmode Quake running on Ada Lovelace's mechanical computer! I remember reading on freshmeat last century [freshmeat.net] that aalib had been ported to it!

  • Would a LITERARY genius necessarily have good spacial comprehension???? Hell yes! Pull two or three pages out of a book and hold them up to a light. You now have letters in different orientations and depth. It would not be easy to crack because the words are intertwined in each other. There is no letter to letter correlation, which is what was hinted to in the last puzzle. This may be a step in the right direction.

  • by FirstEdition ( 79762 ) on Thursday March 09, 2000 @04:48AM (#1215988)
    Did anyone notice that the smaller, broken cypher was encoded using alphabetic substitution, but each word was reversed.

    "....Another clue was taken from Tyler's correspondence to Poe in which he discussed the difficulty of deciphering text that was written backwards ("eht", rather than "the"), and spaces and punctuation are omitted."

    We should not rule this out when/if attempting a brute force dictionary attack.

    Secondly, all symbols in use seem to be in one of 8 alphabets:

    * upper/lower case
    * small/normal size
    * right/upside down

    What about reading off the characters in each of the alphabets, (ignoring spaces & other alphabets), to create a stream of "normalised" characters. ie. start reading all upper/small/right, then upper/small/upsidedown, etc. From there, attack the concatenated stream as an alphabetic substitution cypher, allowing reversed words.

    Of course, we don't know which order to do the alphabets in.

    Hmmm. in the best Bletchley Park tradition, we could run all arrangements of 8 character streams in parallel.

  • So, has anyone tried compiling it yet? Heh heh... ;)
  • This is for all the crypto-newbies out there like myself:

    If we are to believe that Poe was lazy about the writing of the cipher, then we might also believe that Poe was lazy about the encryption schema aswell. In such case we may find that an successful cryptanalysis must also involve an examination of the process used to actually WRITE the cipher on a piece of paper... How would Poe have had to write this cipher down? How would he have done the type-setting?

    This involves more careful thought and use of the more arcane tools of cryptanalysis. I was a big fan of Brute force until I used it to try and crack a 500 year old cipher (the Vigenere). The Kasiski test yeilded a key word of 6... the key space was 2 to the 26th... the execution time was 3 days... the result was garbage. (reason: I was using the Mutual Index of Coincidence of English to tell me if I cracked the cipher)

    Perhaps my code was just buggy.

    The point is: How do you Brute Force something you know nothing about... and How do you know if your answer is "Correct"?

    in example Here is a cipher:


    what's the decryption? Not a lot to go on is it? Perhaps Poe's cipher isn't insoluable, we just don't have enough to go on.

    - // Zarf //
  • If there are so many alphabets, maybe they all map to the same letters, but should be translated one after the other.

    What are the frequencies in each alphabet?

  • Another crypto inspired author and slashdot favorite, Neal 'Cryptonomicon [slashdot.org]' Stephenson, has his own
    crypto challenge [eruditorum.org].
    It is certain to be harder than this one.

    I wasted half a day, and have since given up.

    Mind you the Poe cypher has been around 150 years, Stephenson's Eruditorum challenge has resisted solution for 1 (and counting.)

    Good Luck!

  • one

    spring to mind.

  • Hold 3 pages up to a light, you now have letters in different orientations... but not depth, unless you're using really thick paper. The three sheets will just appear superimposed but it would be very difficult to tell which one was in back, in the middle, or in front.
    I didn't say that this WASN'T a possible explanation, just not likely, since I don't think Poe was reknowned for his ability to think spatially (assuming he was the author?). Maybe he could, but he just wasn't known for that specifically.
    It seems to me that there would be a more "algorithmic" solution...
  • by snicker ( 7648 ) on Thursday March 09, 2000 @05:46AM (#1215995) Homepage Journal
    I bet given time we could come up with marginally reasonable algorithms to transform the cypher into, say, the receipt for Toklas Brownies , the Book of the Subgenius, my home phone number... after all, it could be compressed as well as encrypted. Without any idea of what the plaintext could even BE (what language? what alphabet?) how can we expect to know we have the right answer when we get one?

    Now there's an idea for an encryption algorithm, one that yields a false plaintext if an incorrect key is used...

    Hey, and does anyone care about Negroponte's challenge from the Being Digital hardcover? (Oooh triple encrypted! It's probably just "Yay! Digital!" over and over...)

    yeah, ok. back to work
  • It's the results of this year's obfuscated C code contest.

    It's a UU-encoded pic of Mrs. Poe.

    It's the vendor key track off the DVD of Poe's first movie.

    It's a stolen copy of my first Latin homework assignment. (As you can see, I didn't do very well on it.)

    It's a ruse to distract the predecessor of the NSA from the real secret messages, his poetry.

    It's the Universal Turing Machine's representation of itself.

    It's the transcript of a Ravenish poem about a human who sits on the bookcase and keeps saying "moreover". (That explains the repeats.)

    It's a plot summary for the entire X-Files series. (Actually that's just a WAG, based on the fact that they make equal sense to me.)

    It's Mr. Poe's public key, used as a .sig to the other cryptogram.

    It's the message those two famous Byzantine generals kept sending back and forth.

    It's what he got when he happened to write more /dev/inkpot with his quill.

    It's the way everyone used to write, before the bookmaking cartel decommoditized the protocol.

  • This reminds me of a really excellent book I read last year called The Gold Bug Variations [amazon.com], by Richard Powers. Great geek literature involving genetic coding, Darwinian theory, art history, programming, classical music and the "science of love."

    Hey! Someone else who's read it and likes to write book reviews (not me) should submit it to Slashdot.

    Sorry for the amazon URL. I only buy computer books online, so I wasn't sure what other site was more acceptable to link to.

  • On a different note, what also strikes me is that only capitols (small and large) are used in the To Edgar part. Maybe this is a clue?

    I suspect that the author started coding using uppercase letters first, and when he ran out of uppercase he switched to lowercase or smallcaps.

    Maybe we can indeed use it.

    Furthermore I propose, should we solve this thing together here, to donate the $2500 to the Gutenberg project. It only seems fitting.

  • But if you solved the Beale cipher, would you announce the solution or just go get the gold and keep quiet about it?
  • I find the spacing of the puzzle intriguing. Why aren't all the "words" spaced evenly. The left hand side of the bottom few lines are especially spaced differently. Are there two spaces between some of the letters, or is that just a result of his having to turn the paper over to type the upside down letters? Could this mean that 'space' is actually another letter?
  • You seem awfully confident. Remember, this has sat around for 150 years unsolved and I doubt it's just some simple substitution or even a Vignere cipher.

    For instance, Poe/Tyler could have used any piece of writing as a key so unless you have a perl script that does a word by word analysis of every bit of published writing from Poe's time, you could be in for a long haul.

    Check out The Code Book by Simon Singh. He talks
    about a piece of code that supposedly leads to several million dollars worth of gold buried in the hills of western Virginia that has never been solved despite the efforts of 1000's of people.
    A human being who wants to keep something secret from computer analysis probably can, as long as they don't want anyone else to read it either. You can use several different keys, each being a very obscure text, for instance.

    Of course, I could be wrong. This could be a very simple cipher, but something tells me it probably isn't.
  • We don't know who made these ciphers, there's no good reason to assume that the person who wrote them was into literature at all, let alone whether he or she was a literary genius and least of all whether it was Poe or not.
  • Well, as long as we're talking brute force (and it's a pretty small text, too) you might as well simply check if some percentage of the translation is in /usr/dict/words.
  • ...for those who haven't looked at the other, solved cipher (known plaintext makes it a lot easier, of course...) I've been reminded of something else you can do with your code to make it difficult to break - spelling mistakes don't obfuscate the plaintext very much, but can really confound attempts at decryption without keys.

    The plaintext for the first few lines reads as follows:

    ``The soul secure in her existence smil at the drawn dagger and defies ints point the stars shall fade away the sun himoelf grow dim with age anh nature...''

    A couple of the 'e's there are periods, whilst the rest are commas. Just to put a bit more of a spin on the problem...
  • I tried seperating the alphabets, but it's very difficult because certain characters are impossible to tell apart from eachother. For instance, the lowercase 'c', 'o', 's', 'v', 'w', 'x' and 'z' are almost identical to their smallcaps counterparts.

    Also, the aligment of the alphabets on certain lines makes it extremely hard to tell what alphabet a letter belongs to. For instance, try to separate the lowercase 'b' from upside-down lowercase 'q', lowercase 'd' from upside-down lowercase 'p', lowercase 'n' from upside-down lowercase 'u', lowercase and smallcaps 'o' from upside-down lowercase and upside-down smallcaps 'o', lowercase and smallcaps 'x' from upside-down lowercase and upside-down smallcaps 'x', and lowercase and smallcaps 'x' from upside-down lowercase and upside-down smallcaps 'z'.

    Poe must have known this would be a problem when he constructed the puzzle. Perhaps ambiguous characters default to a certain alphabet, or form an alphabet by themselves?

  • Given that letter substitition is beeing used, I think we can rule out that the words are written backwards. I base this in the fact that at the end of line 10 in the crypto is the word "RVV" (in small capitals).. This could probably translate to "see" or "all" or something like that.. If this was a word written backwards, that would mean that the word would start with two of the same letters.. And I don't know of any such word in english (or in any other language)

    Isn't it logical??

  • I've seen CS professors write less clear code than that.

    The crypto thingy looks like it's in good ol' scary Perl anyways...I'd compile it but I'm no good at anything but assembly anymore...damn cprE classes.

  • After close examination of the piece, I noticed something. The 9th Row, the 4th word from the left indeed does say Ninja! And one word over says "diky"! In theory, this unknown word may be part of a "secret ninja code language", and this word may just happen to mean: "Pancake" in that "secret ninja code language"! Oh this is a glorious day!

    Thank you Mr Poe!
  • The molecular structure of a cryptogram is polar, meaning that it is insoluable in water.

    Humm... Doesn't it work the other way?

  • If we manage to 'solve' the puzzle, how will we know that our solution is correct?
  • True. He could also have gone by the same concept of The Great Gadsby, and not used the letter 'e' at all, regardless of how he encrypted it.
  • Anybody ever consider the posibility of another language being in use here. Who said the intended message is in modern English? I guess this is a debate better left to someone with a better Literature and History person to validate this (if in fact it was done by Poe), but I wouldn't discount it.

    I intend to look at this in detail later. Don't expect to crack it, but its fun to spin the wheels ocassionally. (maybe it'll move me towards a career with the NSA or CIA) :-)

  • I'm looking at the first cryptograph and its solution on "http://www.bokler.com/eapoe.html", but I don't see a one-to-one correspondance between the symbols in the cryptograph and the letters in the solution. For example, if spaces aren't represented in the code and the first few characters are "THESOULSEC", then a comma stands for a 'T' the first time and an 'S' the second time.

    Would someone please explain the key to this cryptograph? Is it one-to-one, or something more complex?

  • Two thoughts on this...

    First, what if you write the whole thing around a rectangular rod (i.e. elongated cube): Take each four-letter combination and write it vertically, then match up the vertical sections, then look at each section and rewrite it in whatever direction makes the text readable. This could then be encrypted using standard replacement techniques. Has anyone done a count on whether each type of character has the same number of occurrences in the cipher??

    Second, I had flashbacks of an old 3-2-1-Contact episode where they wrote a code around a broomstick, i.e. wrap the paper around a broomstick, then write the message across it, then unravel the paper and transcribe the character pattern. Then you tell your friend how far apart the letters should be and what size broomstick you used and they can decode the message by reversing what you did...

    Just my 2c
    The "Top 10" Reasons to procrastinate:
  • I know this is offtopic here, but I looked at it and at first glance, it looks like hex to me. It's a 22x22 matrix of characters ranging from 0-6 and b-f, 12 characters in all. This falls within the definition of hex. I haven't done much analysis yet, but it looks to me to be converted to hex after being created in either decimal or binary (or I could be completely way off base).
    Scott Jones
    Newscast Director / ABC19 WKPT
    Commodore 64 Democoder
  • The words are written backwards, but you still start the message in the upper-left hand corner. There is no word-delimiting symbol.
  • It's a little more complex. The words are backwards and there are spelling mistakes.
  • eel
  • Remembering all the kinds of encrypt forms I learned as a kid, one of the easiest was to throw in dummy characters (things that never meant anything but helped make things look more confusing). There's a few things to look at here:

    1) Possibly, some characters (maybe the ones in all caps, the ones upside-down, or whatever) are just thrown in there to confuse the decryptor.

    or 2) The different alignment of characters was meant to allude to different meanings. For example, if this letter could have been secretly written by Poe himself, maybe he based the text on allusions to his different works, or maybe on only one alone. Perhaps each letter alluded to a different word in a work of his.

    perhaps 3) The combination of the allignment and case of the letters was only meant to throw people off, and instead means absolutely nothing.

    Thinking about Poe's work, this particular code seems to fit. In almost all of his works, he has two sides, the light and dark. The fact that the code has right-side-up, upside-down, upper, and lower case seems to blend in with it. Come to think of it, perhaps it is two messages blended into one?
  • by grappler ( 14976 ) on Thursday March 09, 2000 @09:23AM (#1216037) Homepage
    First off, I know that this COULD be in a different language, meaningless random gibberish, or based on some horribly obscure document. However, I doubt it. Here's my thoughts:

    Here's what gets me - no letters are on top of each other. While this sure looks like it was fed into a typewriter and typed over several times with lowercase, smallcaps, and uppercase, upside down and not, it looks like gaps were left by someone who knew that when he got to that exact spot later, he would need a letter there.

    Also, I took your advise and looked up boustrophedon (great word by the way) - it means that every line would start at the same side of the page that the last ended on and go in the opposite direction "as the ox plows". If it is written this way, you would have to turn the paper upside down to type the other direction, so the stream would consist of right side up characters from alternating lines, and upside down from the other lines, and then go back the other way.

    I think it is more likely that he did something like this:

    o Type the entire right side up portion of the message, leaving gaps where he knew the upside down letters would go.

    o Turn the paper around

    o Type the rest of the message. Done.

    The fact that gaps had to be left implies that some planning went into this before the "final copy" was made. He probably had some rough drafts worked out. Also, he was into cryptograms that people worked out as puzzles, not codes that would be unbreakable. I think it's unlikely that there's some obscure text that he used as a key. If I were making a serious attempt at this, I would look for the following:

    o Words broken up differently from the spaces in the message.

    o Words written backward, or the entire message backward. Or vise versa.

    o Dummy characters. Especially at the end (beginning?)

    o Different substitution alphabets for different kinds/orientations of characters in the message. (This is doubtful IMHO)

    o boustrophedonic writing, like the guy said.

    I'm betting it's cheap tricks like this. You can be surprisingly criptic just doing that.

  • by gothiac ( 58624 ) on Thursday March 09, 2000 @09:44AM (#1216042)
    I have a couple of comments. I've been trying to get into Poe's head here, and I am finding that he would specifically be looking for ways to spoil the typical way of solving these cyphers. Like backward words in the previous example.

    Since he solved hundreds of these, he is really in tune with what is used to solve these (which are the trails that most of you seem to be following.) I would expect him to try to foil all or most of those techniques.

    I believe that it is solveable. It's a bit too perfect to not be (see what I say later.) And from what scant information I have read that Poe has said about these puzzles, he seems like he would be quite opposed to publishing one that didn't work. And man, did he spend some time typing it in just for a hoax.

    6 character sets, or a mutiple of 6, gets us dangerously close to 26 with a multiplier of 4 (6 * 4 ?). Maybe conincidence.

    You all know this already, but I see 6 classes of 26 characters: lg cap normal, sm cap normal, lg lower normal, lg cap inverted, sm cap inverted, lg lower inverted.

    The spaces _look_ like they really could represent spaces. Frequency is about right, and spacing of words seems right.

    Take a look at the first 3 lines of the cypher and you will find that there are almost no repeated characters (if you provide that he is using 6 "alphabets.") Repeats come in to play after you go on for a while. I think that there is an alphabet cycling routine going on here, so that the frequency would be perfectly flat! There is NO WAY that this is conincidence.

    Notice that, unlike normal words, I don't see any words with repeated letters. This is not simple substitution.

    Although I haven't really dug in yet, it seems to be that the different character sets happen more in some instances and less in others. The existence of OGXEW in the very beginning seems to me to be a clue of his algorhythm getting "started" and not into perfect hiding quite yet.

    For the 2-d or 3-d theories out there (my theory was that upside down letters were to be read from the bottom to the top, etc.), it's just too darn hard to do such perfect flat freqency if you are dealing with multiple threads of meaning throughout the message. Probably would have taken a computer.

    Poe is trying to heavily mask frequency analysis.

    Well, that's my $.02.. I'd be interested if anyone could build on this.
  • Good point. In that case, I don't really see any reason why it should be bostrophedonic at all. It could just as easily be regular right-to-left text, in which case I'm betting it is.

  • Since the original jpeg would give an OCR program fits, I gave it my best shot. Am currently encoding it as a two character sequence to handle each letter's orientation. (will be a sub-post).
    For now, here it is, letter orientations ignored.

    To Edgar A. Poe, Esq.
    ----------- original sequence a square ---------
    ------------------------------------------------ --

    A frequency counter.
    foreach (sort(keys(%freq)))
    $byvalue{$temp} .= "$_,";
    foreach (reverse(sort(numeric(keys(%byvalue)))))
    print "$byvalue{$_} has a frequency of $_ \n";
    sub numeric
    if ($a ;
    for ($i=0; $i 90;
    print chr($j);
    print "\n";
    print "---- NEXT SHIFT ----\n";
  • About a week ago, I asked my brother to make me a peanut-butter and jelly sandwich, since he was making himself one - i figured it wouldn't be too much trouble.

    He said no.

    10 minutes ago, he knocks on my door and hands me a PB&J on potato rye. As he is walking out the door, I say "thank you"... but with a mouth full of PB&J, it sounded something like "TbRfaamQz".

    Thats when i figued it out... I figured out the encryption algorighim!!!!

    After discovering the secret to PB&J encryption, reversing the cycle and finding the cypher for reverse PB&J encoding was quite easy.

    Now go make yourself a sandwich, read the encrypted page, and enjoy! I won't spoil the answer here, you'll enjoy it much more if you taste it for yourself :)

    A wise man speaks because he has something to say.
    A foolish man speaks because he has to say something.
  • This really is the right track, imho- problem analysis. And it's not so much looking at the cypher itself, but it's presentation. Published in an article which had been solving cyphers, and with stories which go on to explain the methods of solving (simple) cyphers, if those techniques worked on this, then the author (Poe or Tyler) was just giving another exercise to the readers. But it's not just another exercise, unless the exercise is to let the aspiring codebreaker know that the previous techniques don't work.

    Personally, nothing inspires me more than a challenge. (but don't send me any 'challenges', I'm a lazy-assed procrastinator, too.) And that's really what this cypher does- it challenges you to think about new things. If 'simple' cyphers can be cracked by symbol frequency, then mask the frequency somehow, by not having a 1-to-1 correspondence between symbols.
    -One means of doing this using the Lewis Carrol encryption algorithm. (look it up!:) I don't think that's at work here, because of the lack of repeated characters- in a Lewis Carrol code, you could see the same character repeated several times, each appearance corresponding to a different character because of the cypher key. (that'll make more sense if you looked it up.;)
    -Another means of 'flattening the frequencies' is at work here. Whether or not it's some alphabetic rotation as proposed won't be known until it's cracked. ;) But it seems that this is another part of problem analysis which could yield some good starts- figure out the problems resulting from trying to map 26 characters up to 156. I pose to the reader, how do you choose what set of 26 characters to use, and then what character? (so I pop open the calculator and start shuffling numbers... nothing really conclusive.)
    -Following the last idea of character mapping, then there's the matter of building words, and building the paragraph line by line. Maybe the words are encoded, and then their order is determined by another algorithm. (And here I'll give thanks to the Anonymous Coward who offered 'boustrophedon', and to dictionary.com for having the definition!) This compounds the problem of decryption, because finding out how a word is encoded is now foiled by it potentially being upside down and backwards. So in trying to solve the character-mapping problem, you have to think of every which way each word could be turned.

    That should be enough thought-provoking schlock for anybody either looking at the problem, or at encryption in general.

    "Holy exponential permutations!" -cryptomancer

  • While small Caps/lowercase are still undistinguishable upside down characters are aligned a little lower than 'normal' characters (even small ones) so they can be distinguished with a ruler. I think though that Poe thought about this and used ambigous characters only for one meaning. For example i didn't see an upside down 'x' in the text, but i did find small caps 'T' and 't' (even next to one another). I did some counting and believe (as some others here) that it's a polyalphabetic ciphre, maybe with some added difficulties, since in the letters i counted (only large upside) none stood out being used particular often and most were used about 9 times throughout the text.
  • Actually, it was just an early entry in the Obfuscated DeCSS Contest [nttg.net].

"If it's not loud, it doesn't work!" -- Blank Reg, from "Max Headroom"