Last night I tried a simple experiment.

There is a spike at period 15 and 5 for the 340, so I tried to replicate that. Similar to the cryptography books of the day, I used multiple 6×5 inscription rectangles. I alternated reading from the inscription rectangle column 1, 3, 5, 2, 4, and 6 in that order. Transcribed and encoded with about 63 symbols, using Jarlve’s plaintext library and making one symbol map to three plaintext. It was fairly difficult to make a message with a spike at period 15 comparable to the 340, and even when I did, there weren’t spikes at period 5 or period 10.

Then I made some messages with the same key, but no transposition at all. And it was very easy to make a message where period 1 spiked as high as the 340 period 15 spike.

EDITED PICTURE:

List of route transposition issues

I am still pretty convinced that the 340 is a route transposition, but it must be messed up somehow and a combination of some of the below. Not necessarily in this order.

1. Complete inscription rectangle with gibberish in one corner of the message, or possibly in some other area of the message.

2. Incomplete inscription rectangle, with or without gibberish in one corner of the message, or possibly in some other area of the message.

3. Gibberish row in the transcription rectangle; see above. The period 19 bigram < S scores the highest, and is in the left half of the bottom row. This to me is evidence that the left half of the bottom row may not be gibberish. But the right half of the bottom row has ZO^AIK in it, which looks like a signature to me. I have a difficult time believing that this is accidental.

4. Gibberish column in the transcription rectangle. Transcription around gibberish column, but maybe more likely gibberish row after transcription. Maybe it is period 14 or period 18. We worked on gibberish rows, but they do not make a message nearly as difficult to solve as a gibberish column, which causes a lot more misalignments. We should work on that some time.

5. Regular transcription skips or nulls. A gibberish column would be regular transcription nulls, only spaced at 17.

6. Random transcription skips or nulls.

7. Some or all of the pivot symbols are nulls. I recently updated my skip / null detection heatmap to make it more efficient, and to test for multiple conditions like "increase period 30 bigrams and decrease period 29 bigrams." Using the forgoing test, the areas of the pivots showed positive while most other areas of the message showed negative. I need to update for multiple skips or nulls in groups, say for example, delete groups of three symbols in all areas of the message and see what happens.

8. Multiple inscription rectangles, but only a few larger ones. I really like how I can paste any size message into the program, because with multiple inscription rectangles, I can untranspose them and stack them vertically.

9. Multiple transcription rectangles, something we haven’t worked on yet. If so, how many steps did he take? Could he have transcribed into the 17 x 20 message, but into different separate rectangles in the 17 x 20 grid? Or would there have been an extra step, where he transcribed into different separate rectangles, then transcribed again into the 17 x 20 grid?

10. Inscription into a large rectangle, then rearrangement of rows, then transcription. This could create a lot of period 15 / 19 repeats. We have experimented a little bit with solving a message similar to this before, testing the program to see if it could solve with only fragments. I am not sure what you did exactly. But I have an idea.

Let’s say we untranspose a message and it is flipped. I believe that the program links the rightmost symbols with the leftmost symbols in the next row for looking up and scoring the n grams. What about a solve mode that recognizes a special symbol, like maybe a "." or something else. And have the program lookup 1 grams, then 2 grams, then 3 grams, then 4 grams, then 5 grams before and after the ".". Like it would do at the beginning and end of the message.That way, even if the rows are not in the correct order, maybe the program will not fight itself by trying to link the rows together.

Maybe the program already does this. I am not sure.

Or perhaps transcription vertically, then rearrangement of columns, then transcription. Basically the same idea.

11. Gibberish row or column in the inscription rectangle, which would essentially be a set of regular nulls that would still allow for a lot of bigram repeats, but cause a lot of misalignments.

12. Three big rectangles. Inscribe into rectangle 1, transcribe into rectangle 2, then transcribe again into rectangle 3.

13. One or more polyphones. EDIT: He did that in the 408.

Maybe there are other issues. What bothers me is that he drafted this thing very meticulously, even scratching out the K and replacing it with a backwards K. Why would anyone do that knowing that there is some problem with the transposition making the message so difficult to solve? Why would anyone do that if there were polyphones? I mean, what difference does one symbol make? Could the transposition issue be something so simple and we haven’t thought of or tried it yet?

EDIT:

14. Poly literal transposition, inscribing into a geometric shape, and reading off the symbols two or more rows at the same time, not necessarily adjacent rows, then transcribing into the transcription shape. Possible way to detect by comparing period 15 / 19 repeat symbols with other periods for a spike.

Take your time and let me know what you think. I will post some other ideas about solving the message soon.

EDIT: Maybe I should name my messages so that they indicate what kind of messages they are, instead of "smokieX". Let me know what you think.

I don’t think that we are going to untranspose the entire message and get a solve. So here is a possible way to solve the 340:

Idea of how to solve the 340

1. Check for suspected transposition periods reading message in eight directions.

2. Check cycles and identify any possible polyphones, or symbols that could possibly be merged later. Also look for long, nearly perfect cycles where the symbol that is out of sequence could be indicator of disruption.

3. Draft the message into columns = suspected period.

4. Transpose at 90 degrees so it reads horizontal. Make mirrored, flipped, mirrored and flipped versions.

5. Make a test message with similar symbol count distribution and cycle score, but without transposition. Draft into same shape as transposed message. Explore AZD capabilities with regards to minimum grid sizes that can be solved. Longer rows and fewer columns is much better than shorter rows and more columns.

6. Try to solve the message. Identify small groups of words that reveal context or parts of phrases. Try different areas of the message instead of the entire message. Scroll through the message with the minimum count of rows needed for a partial solve. Scroll through the message with the minimum count of columns needed for a partial solve. See item 10 in the post above. Examples:

7. Try to detect possible distortions. Nulls, skips, strings of gibberish symbols. Try to solve areas of the message that have more repeat symbols highlighted and fewer distortions.

8. If partial solve, populate a key and apply to remaining part of message. Figure out what the encoder did by looking at the plaintext and rearranging.

9. Try another period. Work down from most predominate, to less predominate.

10. Expand possible polyphones, and merge symbols that cycle together, ranking from highest to lowest. Maybe do this at the beginning of the procedure instead of at the end of the procedure.

I’m not exactly sure how virtualization works, if that is the kind of virtualization I think you are referring to. Is it such that you can set up a rack where each rack unit holds a bunch of CPU’s and RAM which then can be accessed as a whole through a single main computer?

Yes, that’s basically the idea. Most virtual host providers allow you to dynamically allocate some number of CPUs (each with some number of cores) and RAM based on your application needs and budget. You could set up an application on a low-powered configuration, and then one day decide to ramp up the number of CPUs and cores by making some configuration changes without needing to install the application on new hardware.

As an example, imagine running AZDecrypt on the 128-CPU "X1" instance on Amazon with 1,952 GB of RAM: https://aws.amazon.com/ec2/instance-types/x1/

Downside: Processing time costs almost $4,000 per hour!

Another option is to learn CUDA and take advantage of the thousands of cores that come with relatively inexpensive graphics cards: https://en.wikipedia.org/wiki/CUDA It would be interesting to see how quickly thousands of GPU cores could work on decryption tasks.

An example from 5 years ago of a 25 GPU system that can crack passwords in seconds: https://securityledger.com/2012/12/new- … n-seconds/ "In a test, the researcher’s system was able to churn through 348 billion NTLM password hashes per second"

At Nick’s site I came forward with some observations on the Beale ciphers for anyone who’s interested: http://ciphermysteries.com/2015/12/22/e … ent-349674

Very interesting observations – I think you have discovered something new about B3 that nobody has noticed. Well done! Have you read this paper?

http://zodiackillerciphers.com/2013-3-0 … -2-BIT.pdf

What do you think of his conclusions? I attended his presentation at a crypto symposium and his reasoning that the Beale ciphers are a hoax was rather persuasive.

Thanks for the information doranchak. The new Ryzen chips from AMD should also perform very well with AZdecrypt.

What do you think of his conclusions? I attended his presentation at a crypto symposium and his reasoning that the Beale ciphers are a hoax was rather persuasive

A non-random string like "ABCDEFGHIJKLMN" could as well be filler or a hint and does not exclusively pinpoint the cipher as a hoax which is the easy way out. I think the B1 and B3 are interesting and are worthy of further research, which sadly has been dampened by the hoax theory.

Hey smokie,

I like all your points and here are a few comments with added categories to make sense. Ideally, we should try to come up with examples of every point (some we already have) so that my transposition solver can help and be helped.

I am still pretty convinced that the 340 is a route transposition, but it must be messed up somehow and a combination of some of the below. Not necessarily in this order.

We owe it to ourselves to get at least to some reasonable bottom of the issue. It indeed must be messed up somehow.

Aperiodical nulls:

1. Complete inscription rectangle with gibberish in one corner of the message, or possibly in some other area of the message.

2. Incomplete inscription rectangle, with or without gibberish in one corner of the message, or possibly in some other area of the message.

6. Random transcription skips or nulls.

7. Some or all of the pivot symbols are nulls. I recently updated my skip / null detection heatmap to make it more efficient, and to test for multiple conditions like "increase period 30 bigrams and decrease period 29 bigrams." Using the forgoing test, the areas of the pivots showed positive while most other areas of the message showed negative. I need to update for multiple skips or nulls in groups, say for example, delete groups of three symbols in all areas of the message and see what happens.

13. One or more polyphones. EDIT: He did that in the 408.

Yes, stuff with corners can be hard. I have ran the same tests in the other thread and the areas of the pivots indeed return stronger here. My filler test finds both the pivot rows and columns and even down to the smaller individual pivot segments as most likely. I am really not convinced though, the pivot axes positions are a successive multiple of 39 and are period 39 repeats, though that does not rule out the pivots out as being potential filler.

To generalize further, for me, a null could be 1. added, 2. skipped or 3. polyphone. In case 1 it needs to be removed at the position, in case 2 it needs to be added at the position and in case 3 it needs to be expanded at the position. Though I think that generalizing polyalphatism under nulls only make sense if it is periodical (part of the cipher).

Periodical nulls:

3. Gibberish row in the transcription rectangle; see above. The period 19 bigram < S scores the highest, and is in the left half of the bottom row. This to me is evidence that the left half of the bottom row may not be gibberish. But the right half of the bottom row has ZO^AIK in it, which looks like a signature to me. I have a difficult time believing that this is accidental.

4. Gibberish column in the transcription rectangle. Transcription around gibberish column, but maybe more likely gibberish row after transcription. Maybe it is period 14 or period 18. We worked on gibberish rows, but they do not make a message nearly as difficult to solve as a gibberish column, which causes a lot more misalignments. We should work on that some time.

5. Regular transcription skips or nulls. A gibberish column would be regular transcription nulls, only spaced at 17.

11. Gibberish row or column in the inscription rectangle, which would essentially be a set of regular nulls that would still allow for a lot of bigram repeats, but cause a lot of misalignments.

The string "ZO^AIK" only has 2 matches with "ZODIAC" so it could as well be random with our brains filling in the rest. I have a null operator in the transposition solver (currently unused) which could get all of these but the argument searchspace is rather large: add or remove null, start position, end position, period. I will have to check if it is in range of brute force. That could be interesting.

Multiple inscription and/or transcription rectangles:

8. Multiple inscription rectangles, but only a few larger ones. I really like how I can paste any size message into the program, because with multiple inscription rectangles, I can untranspose them and stack them vertically.

9. Multiple transcription rectangles, something we haven’t worked on yet. If so, how many steps did he take? Could he have transcribed into the 17 x 20 message, but into different separate rectangles in the 17 x 20 grid? Or would there have been an extra step, where he transcribed into different separate rectangles, then transcribed again into the 17 x 20 grid?

10. Inscription into a large rectangle, then rearrangement of rows, then transcription. This could create a lot of period 15 / 19 repeats. We have experimented a little bit with solving a message similar to this before, testing the program to see if it could solve with only fragments. I am not sure what you did exactly. But I have an idea.

All very interesting.

Stacked inscription and/or transcription rectangles:

12. Three big rectangles. Inscribe into rectangle 1, transcribe into rectangle 2, then transcribe again into rectangle 3.

I like this allot.

Let’s say we untranspose a message and it is flipped. I believe that the program links the rightmost symbols with the leftmost symbols in the next row for looking up and scoring the n grams. What about a solve mode that recognizes a special symbol, like maybe a "." or something else. And have the program lookup 1 grams, then 2 grams, then 3 grams, then 4 grams, then 5 grams before and after the ".". Like it would do at the beginning and end of the message. That way, even if the rows are not in the correct order, maybe the program will not fight itself by trying to link the rows together.

That is a very good idea smokie. I have previously applied it to solving anagrams or keyed ciphers by treating each cipher line/row independently. This functionality can be added to AZdecrypt if you need it, let me know. Another option would be to adjust the ngrams, reformat the input corpus into rows of 17 characters and randomize the row order.

EDIT: Maybe I should name my messages so that they indicate what kind of messages they are, instead of "smokieX". Let me know what you think.

I would prefer that you keep the numbering system and use a look up table, you could add it at the start of this thread.

Update on the transposition solver. For week 2 I allowed allot more randomness and in general this did no good. The multiple inscription rectangles ciphers (smokie46 to 48) solved quite well, some of them did not which is probably just a matter of time. Smokie, feel free to make another suite.

Format: AZdecrypt score _ Index of coincidence times 10000 _ multiplicity times 1000

340:
week 1: 21151_748_185
week 2: 21115_738_185

340_flip(17,20):
week 1: 21071_685_185
week 2: 21088_762_185

340_mirror(17,20):
week 1: 21175_727_185
week 2: 21017_745_185

340_offsetcolumnorder(tp,17,20,1):
week 1: cipher not included
week 2: 21030_771_185

340_periodcolumnorder(utp,17,20,2):
week 1: cipher not included
week 2: 20914_715_185

340_reverse:
week 1: 21152_739_185
week 2: 20916_732_185

441:
week 1: 20763_699_142
week 2: 20656_685_142

dorabella:
week 1: cipher not included
week 2: 23443_708_229

doranchak_route1:
week 1: cipher not included
week 2: 21655_738_185

feynman1:
week 1: 21228_757_60
week 2: 21131_757_60

glurk_cc1:
week 1: 25125_653_185
week 2: 25071_656_185

glurk_cc2:
week 1: 24194_739_176
week 2: 24075_733_176

jarlve_cc1:
week 1: 21033_750_185
week 2: 20977_762_185

jarlve_chunksize_easy:
week 1: cipher not included
week 2: 20271_780_173

jarlve_chunksize_hard:
week 1: cipher not included
week 2: 20707_798_185

jarlve_route1:
week 1: 22161_620_64
week 2: 21188_630_64

jarlve_route2:
week 1: 25220_605_73
week 2: 25223_605_73

jarlve_route3:
week 1: 23893_724_67
week 2: 21566_730_67

jarlve_route4:
week 1: 22214_713_185
week 2: 21416_730_185

jarlve_route5:
week 1: 24930_641_67
week 2: 24895_641_67

kryptos3:
week 1: 25275_663_74
week 2: 25206_667_74

reunion:
week 1: cipher not included
week 2: 23804_762_246

smokie9:
week 1: 21098_732_185
week 2: 20987_744_185

smokie10:
week 1: 20899_750_182
week 2: 20649_747_182

smokie11:
week 1: 21958_766_185
week 2: 21910_767_185

smokie12:
week 1: 21240_726_182
week 2: 21195_737_182

smokie13:
week 1: 22451_757_185
week 2: 21432_735_185

smokie14:
week 1: 21137_707_185
week 2: 20796_723_185

smokie15:
week 1: 21189_731_185
week 2: 20918_719_185

smokie23:
week 1: cipher not included
week 2: 21797_635_176

smokie38:
week 1: cipher not included
week 2: 20884_717_191

smokie39:
week 1: 21207_763_185
week 2: 20931_718_185

smokie39_expand(7):
week 1: 22567_706_252
week 2: 22454_698_252

smokie40:
week 1: 22682_692_61
week 2: 22593_692_61

smokie41:
week 1: 20943_730_67
week 2: 21872_702_67

smokie42:
week 1: 21274_663_67
week 2: 20920_625_67

smokie43a:
week 1: 23812_650_70
week 2: 21966_650_70

smokie43b:
week 1: 23812_650_70
week 2: 22165_649_70

smokie43c:
week 1: 23812_650_70
week 2: 22730_663_67

smokie44a:
week 1: 22703_671_129
week 2: 22669_699_129

smokie44b:
week 1: 23275_693_135
week 2: 21831_754_135

smokie44c:
week 1: 23118_644_132
week 2: 22614_640_132

smokie45a:
week 1: 21780_731_182
week 2: 22119_699_182

smokie45b:
week 1: 21109_756_191
week 2: 21270_703_191

smokie45c:
week 1: 23720_630_188
week 2: 22739_664_188

smokie46a:
week 1: cipher not included
week 2: 24099_728_67

smokie46b:
week 1: cipher not included
week 2: 23005_684_70

smokie46c:
week 1: cipher not included
week 2: 21317_671_70

smokie47a:
week 1: cipher not included
week 2: 24149_684_120

smokie47b:
week 1: cipher not included
week 2: 19525_761_141

smokie47c:
week 1: cipher not included
week 2: 23770_613_123

smokie48a:
week 1: cipher not included
week 2: 23801_728_170

smokie48b:
week 1: cipher not included
week 2: 20994_724_185

smokie48c:
week 1: cipher not included
week 2: 24001_768_176

I don’t think that we are going to untranspose the entire message and get a solve. So here is a possible way to solve the 340:

I need some time to digest and interpret your attack plan, there is allot of conditionality.

Basically the attack plan centers around trying to solve areas of a message instead of an entire message, so as to avoid disruptions. With a partial solve leading to a key, it would theoretically be possible to solve the distorted area by hand. It would require the "." idea, or, maybe cleaner and easier, just to treat individual rows as independent of each other, regardless of the length of the rows, and not using a special symbol.

How is it going with smokie39? This message was made with the attack plan in mind. Maybe you have a decent solve, I am not sure. I would like to show it to you if you do not.

I will work on some new messages soon. I want to make some with gibberish columns. And I will make the lookup table, a better idea.

It would require the "." idea, or, maybe cleaner and easier, just to treat individual rows as independent of each other, regardless of the length of the rows, and not using a special symbol.

I’ll add it to my to do list.

How is it going with smokie39? This message was made with the attack plan in mind. Maybe you have a decent solve, I am not sure. I would like to show it to you if you do not.

I’m not working on it anymore so go ahead.

My shuffle tests show that with the 340, rows 3 and 18 are more likely gibberish. I find that interesting because one is the third row down from the top and one is the third row up from the bottom. If I mirror row 18, the count of period 19 bigrams goes up. Not so much with row 3. So I decided to make a message that has two rows that are flipped, one polyphone, and two hand made pivots.

Smokie39

It is a complete inscription rectangle, 34 columns x 10 rows, creating a spike at period 10. I transcribed into the 17 x 20 rectangle LRTB, but when I got to rows 3 and rows 18, I transcribed RL instead of LR.

The key has one polyphone to simulate the +. It is more cyclic at the top and becomes less cyclic toward the bottom. The pivots are just hand made and create a few polyphones but no distortion.

When I tried to solve it myself, I just drafted in up into 10 columns, then transposed it at 90 degrees. Then tried to solve various parts of the message, including the section in the middle that does not have any distortions at all. It is a pretty big section ( bold outline ), but the yellow shaded cells are the only ones that the solver found the correct plaintext for.

Anyway, that is basically the attack plan, except that I only show one area that I attempted solve instead of a sliding area. I didn’t think that the message would be so difficult, but only two rows transcribed the opposite direction and one polyphone made it so difficult that even when I tried to solve the part of the message that wasn’t distorted, it still wouldn’t solve. You may want to mess around with this idea. Here is the message again, and the solution, from Lord of the Flies again.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
12 18 19 20 15 21 22 23 24 25 1 26 27 10 17 28 25
29 7 30 4 31 32 33 34 35 20 36 13 20 37 38 39 6
40 2 27 41 12 7 11 5 12 33 9 14 41 42 43 40 25
23 15 8 44 45 41 36 16 40 46 47 48 49 1 4 17 18
24 19 50 21 51 7 52 37 28 6 44 27 34 53 54 41 20
33 55 54 7 50 47 14 13 38 2 7 56 53 9 23 57 58
35 5 52 30 59 16 8 60 20 24 2 42 7 61 11 14 48
21 2 40 7 11 20 15 27 23 2 28 32 29 53 38 49 16
37 31 7 39 44 14 54 28 6 25 57 17 55 15 14 18 32
19 41 57 56 8 62 7 26 27 10 57 30 32 33 35 40 7
7 12 24 27 6 63 20 47 14 57 4 47 30 2 3 6 38
48 35 27 7 15 43 53 13 37 49 25 15 23 37 16 32 42
38 34 7 45 20 63 52 53 58 25 54 4 30 55 42 56 16
59 28 60 61 17 7 21 7 3 20 42 35 7 26 6 12 5
61 9 14 7 8 24 20 51 43 19 4 30 4 25 27 39 63
63 15 15 36 56 21 62 34 55 14 16 42 35 7 18 20 7
44 50 28 25 15 40 41 30 2 38 5 1 38 21 34 7 37
44 32 55 19 41 36 7 57 59 41 11 56 14 20 60 62 32
7 22 27 25 3 23 61 36 28 43 43 1 30 60 40 61 38

T H E B O Y W I T H F A I R H A I R L O W E R E D H I M S E L F D O
W N T H E L A S T F E W F E E T O F R O C K A N D B E G A N T O P I
C K H I S W A Y T O W A R D T H E L A G O O N T H O U G H H E H A D
T A K E N O F F H I S S C H O O L S W E A T E R A N D T R A I L E D
I T N O W F R O M O N E H A N D H I S G R E Y S H I R T S T U C K T
O H I M A N D H I S H A I R W A S P L A S T E R E D T O H I S F O R
E H E A D A L L R O U N D H I M T H E L O N G S C A R S M A S H E D
I N T O T H E J U N G L E W A S A B A T H O F H E A T H E W A S C L
A M B E R I N G H E A V I L Y A M O N G T H E C R E E P E R S A N D
B R O K E N T R U N K S W H E N A B I R D A V I S I O N O F R E D A

Anyway, for the next suite I was going to make messages with gibberish columns or partial columns. Is there anything else in particular that you would like instead? Just checking.

I didn’t think that the message would be so difficult, but only two rows transcribed the opposite direction and one polyphone made it so difficult that even when I tried to solve the part of the message that wasn’t distorted, it still wouldn’t solve. You may want to mess around with this idea. Here is the message again, and the solution, from Lord of the Flies again.

Okay,

In the next codebox I marked all the disruptive elements in your cipher (rows, polyphone and horizontal pivot parts) as number 99 while renumbering the others to 1 and then untransposed period 10. Especially this kind of vertical disruption is very hard for the solver since it can’t find many long uninterrupted sections. There are 61. After expanding these you end up with a cipher multiplicity of 0.361, though it does solve since there’s enough structure left. Interesting scheme, also in regards to the 340. And thanks for sharing the full details. I need to think about it.

1  1  1  1  99 99 1  1  1  99 1  1  1  1  1  1  1
1  1  1  1  1  1  1  1  99 1  1  1  99 99 1  1  1
1  1  1  1  99 1  1  1  1  1  1  1  1  99 1  1  1
1  1  1  1  1  1  1  1  1  1  1  1  99 99 1  1  1
1  1  1  1  99 1  1  1  1  1  1  99 1  1  1  1  1
1  1  1  1  99 1  1  1  1  1  1  1  99 99 99 1  1
1  1  1  1  99 1  1  1  1  1  1  1  1  1  1  1  1
1  1  1  1  99 99 1  99 1  1  1  1  99 99 1  99 1
1  1  1  99 99 1  1  1  1  1  1  1  1  1  1  1  1
1  1  1  1  99 1  1  1  1  1  1  1  99 99 1  1  1
1  1  1  99 99 1  1  1  1  1  99 1  1  1  1  99 1
1  1  1  1  1  1  1  99 1  1  1  99 99 99 1  1  1
99 1  1  99 99 99 1  1  1  1  1  1  1  1  1  99 1
99 99 1  1  1  1  1  1  1  1  1  1  99 1  1  1  1
1  1  1  99 99 1  1  1  1  1  99 1  1  1  1  99 1
1  99 1  99 1  1  1  1  1  1  1  1  99 1  1  1  1
1  1  1  99 99 1  1  1  1  1  1  1  1  1  1  1  1
1  1  1  1  1  1  1  1  99 1  1  99 99 1  1  1  1
1  1  1  99 99 1  1  1  1  1  1  1  1  99 1  1  1
1  1  1  1  1  1  1  1  1  1  1  99 99 1  1  1  1

Anyway, for the next suite I was going to make messages with gibberish columns or partial columns. Is there anything else in particular that you would like instead? Just checking.

Sure, go ahead. And thanks in advance.

hi. still reading everything, trying to keep out of your way and taking a break at the same time.
i need to know how to utilise the zodiac pattern drawer so i can take the route chosen and drop it into az decrypt as symbols rather than position numbers(1-340), can anyone alter the programme to accommodate such a request.
cheers.

Smokie 49ABC is 1:1 substitutes, with one column mostly or all gibberish. The inscription rectangle is complete, but I made them so that they all make up at least 323 symbols. One corner in most cases, or possibly one row, at the top or bottom of the message is not transposed. EDIT: Most of the following messages, including smokie50 and smokie51, are close to 340 with only a few gibberish symbols in one corner.

A

1 2 3 4 5 6 7 8 3 9 4 10 3 11 12 6 10
13 13 14 4 15 11 15 13 16 7 14 11 14 3 13 17 14
13 15 18 4 5 15 3 3 1 12 4 14 5 6 12 17 1
3 13 9 3 8 19 2 2 6 4 6 12 16 5 12 20 15
10 10 12 4 5 4 6 21 4 2 16 11 14 3 9 3 12
10 13 6 13 14 4 6 4 16 11 4 14 3 22 19 3 15
21 12 22 11 8 13 13 13 3 5 6 23 3 14 4 15 6
12 4 12 15 5 3 18 15 5 7 10 23 15 18 3 2 9
10 3 6 10 5 23 11 13 13 10 5 3 12 3 15 18 10
11 19 21 10 5 4 5 6 21 8 8 10 5 5 1 5 3
18 5 9 2 6 21 5 15 3 8 9 13 12 21 14 9 3
4 15 21 9 10 18 23 18 12 11 8 18 4 12 13 6 18
4 11 13 3 11 5 3 15 8 15 13 23 12 14 3 6 12
15 16 20 13 21 18 13 4 8 6 8 4 20 14 6 7 12
3 15 3 3 4 16 9 15 10 13 3 4 3 4 13 14 12
6 8 7 3 5 18 21 14 3 8 2 15 15 5 10 14 12
3 16 11 3 10 5 15 4 1 15 10 16 1 4 10 10 7
5 21 4 12 3 10 11 3 4 18 10 9 9 3 6 4 6
10 10 13 3 6 13 3 1 3 12 10 15 11 14 14 15 8
12 12 15 2 2 4 16 12 18 4 16 3 18 12 8 21 3

B

1 2 3 4 4 5 6 7 7 8 1 9 2 10 11 12 8
10 7 13 1 4 8 14 12 15 6 16 10 1 3 7 10 8
14 6 7 1 8 17 7 16 15 5 2 11 8 2 8 17 8
17 2 5 12 6 13 14 9 14 8 5 18 6 9 12 7 10
6 2 9 12 19 14 2 8 8 7 9 12 20 18 21 20 7
2 6 17 6 12 1 9 2 12 5 14 3 1 7 13 5 8
14 22 7 7 5 11 2 4 8 9 12 19 23 22 2 22 12
8 10 6 4 6 6 4 5 4 8 10 1 7 14 8 14 6
7 5 22 2 5 13 1 14 8 19 2 19 24 8 1 2 10
9 1 12 20 5 6 23 20 21 7 5 6 12 5 7 10 7
17 9 5 2 7 6 14 22 7 9 7 6 10 8 2 10 22
5 7 5 1 19 13 6 22 8 7 5 7 2 2 12 2 4
12 22 8 6 6 8 2 13 11 7 1 2 9 9 12 10 8
2 10 2 8 1 4 1 10 14 17 2 13 2 13 8 2 7
16 7 10 13 6 11 5 7 5 5 17 5 1 2 23 5 10
8 13 14 8 7 17 12 10 1 9 1 19 9 12 8 15 5
12 6 24 13 20 9 6 2 7 2 4 22 4 1 1 8 10
7 2 20 11 8 5 6 2 22 13 6 6 8 8 2 4 5
5 7 9 1 15 5 19 12 9 13 13 1 12 11 6 1 20
19 10 5 2 9 7 6 15 7 5 20 1 4 14 2 1 8

C

1 2 3 4 3 5 6 1 7 8 9 10 11 8 3 2 12
13 13 4 1 13 13 14 2 11 6 11 1 3 1 15 3 9
11 1 16 17 5 3 18 3 19 20 17 6 6 7 11 2 4
13 1 16 15 4 18 15 10 9 14 20 11 2 3 17 6 21
6 1 13 6 16 17 11 16 3 16 1 9 1 13 2 22 3
17 13 1 11 17 18 3 6 1 11 5 15 8 11 4 3 13
13 17 6 6 8 1 6 17 9 4 21 11 13 14 13 18 4
13 8 9 13 23 14 4 17 10 17 17 9 4 6 11 21 1
6 15 11 18 2 24 22 8 2 5 1 1 14 1 17 23 17
15 21 3 11 16 1 10 13 2 6 13 6 1 2 1 6 16
9 11 17 13 9 6 1 17 13 6 17 2 1 21 4 9 11
2 8 13 18 11 5 6 9 12 13 1 17 6 2 11 13 16
16 16 16 8 18 14 22 16 5 2 6 20 3 15 25 13 22
2 1 5 1 10 1 5 8 13 1 3 5 4 9 5 13 8
16 19 1 22 16 5 14 16 1 5 8 2 13 17 1 13 8
2 1 13 5 4 4 17 17 17 4 2 16 1 1 8 2 17
3 13 22 23 5 2 11 16 15 1 13 1 15 2 6 1 10
17 10 9 5 11 3 18 9 23 18 16 5 3 18 2 1 13
8 3 13 17 6 17 13 21 11 13 5 16 23 22 17 9 3
17 23 15 13 8 16 17 1 14 17 5 1 1 2 6 1 3

The other messages will be very soon as I have the hard work done already.