Talk:Transposition cipher

Untitled

Bold text

This isn't Railfence

The system described under the section "Railfence" is not the Railfence system as I've usually seen it described.

The system described here is a columnar transposition, written down and read left-right. The Railfence system as I've always seen it used was based on a rising-and-falling pattern:

ABCDEF

A				E
	B		D		F
		C

AEBDFC --jdege 19:00, 18 January 2007 (UTC)

Yes, that's what I've understood it to be too. I've removed the section for now until we can clear it up. — Matt Crypto 19:09, 18 January 2007 (UTC)

I've done some websearching, and I've found that there are a number of sites that use "Railfence" as was used on this page:

http://www.cs.trincoll.edu/~crypto/historical/railfence.html

http://www.braingle.com/brainteasers/codes/railfence.php

OTOH, there are sites that refer to "Railfence" as I've always thought it was used:

http://members.aon.at/cipherclerk/Doc/Railfence.html

http://www.woodmann.com/fravia/railfe.htm

The NSA's CryptoKids site describes Railfence only for height=2, in which case the two systems yield identical results:

http://www.nsa.gov/kids/ciphers/ciphe00012.cfm

As for who is right? Seems to me that write-by-coumn/read-by-row method that was described here is equivalent to the write-by-row/read-by-column that we see described in the section on Columnar Transposition. The older references seem to use Railfence as I had thought it meant. I'm working through Caxton Foster's "Cryptanalysis for Microcomputers", from 1982, and I'm absolutely certain that it uses Railfence to refer to an up-and-down transposition. And I'm pretty sure that Helen Gaine's "Cryptography", from 1939, uses the same.

The canonical answer on the Web, though, is probably the American Cryptogram Association, and they're solidly on the up-and-down side:

http://www.cryptogram.org/cdb/aca.info/aca.and.you/chapter_09.pdf#RAILFE

--jdege 19:26, 18 January 2007 (UTC)

I've just noticed that the current page contains a reference to the Rail Fence section, which has been removed. --jdege 23:52, 26 February 2007 (UTC)

I've edited Rail_Fence_Cipher to correct this misconception. If folks could look at it, check it for spelling, grammar, form, and fix what needs to be fixed, I'd appreciate it.

Perhaps, when it's right, we could consider using it to restore the Railfence section, here.

--jdege 02:30, 27 February 2007 (UTC)

Opening Sentence

"... a transposition cipher changes one character from the plaintext to another (to decrypt the reverse is done)."

If you read on, the writer's intention becomes clearer but the opening line is a bit of a muddle. Could I suggest: a transposition cipher changes the order of characters in a message. Example: THIS IS A TRANSPOSITION can become HTSI SI TA ARSNOPISITOIXN.

--Steve 00:18, 6 November 2007 (UTC)

"... a transposition cipher swaps the positions of characters in the plaintext in predetermined way, so that the ciphertext constitutes a permutation of the plaintext."

--jdege 15:39, 6 November 2007 (UTC)

I like this much more than the current sentence, however characters are not always the units that change position in a transposition cipher. (It's probably a better idea to follow the lead of the Substitution cipher article and call them "units of plaintext." Also, is "swap" the right word to describe the transposition of a transposition cipher? Someone who read that might infer that specific pairs of units are switched, when they can really be disarranged by any regular system.

I think that this will work well:

"... a transposition cipher is a method of encryption by which the positions of plaintext units are shifted according to a regular system, so that the ciphertext constitutes a permutation of the plaintext.

hac (talk) 02:04, 25 December 2008 (UTC)

Question

---

Who copied from who? The info on columnar transposition looks identical to that from this site. http://www.braingle.com/brainteasers/codes/columnartrans.php

-- anon 15:26, 12 January 2008

If you look carefully, that website credits Wikipedia. — Matt Crypto 09:28, 13 January 2008 (UTC)

Looping the scrambling process

"Looping the scrambling process on the binary string multiple times before changing it into ASCII characters would likely make it harder to break."

A transposition of a transposition is just a transposition. If the transposition produces a binary string that can be interpreted as ASCII then necessarily the input alphabet hasn't been expanded by repeating the process. The only attack I can see as being made more difficult is the case where the attacker knows the specific scrambling technique and knows a weakness in that technique (perhaps using known plaintext) but doesn't know the number of times the scrambling has been applied. 173.243.41.226 (talk) 23:17, 23 February 2014 (UTC)

Grille Ciphers

The claim that grille ciphers were used only at the start of the Great War is questionable. David Kahn in The Code-Breakers, 1996 edition, page 308:

"At the end of 1916, transposition messages again appeared in German military communications."

French cryptographers recognised turning grille ciphers, in January 1917, and devised attacks.

"The grilles lasted four months." (page 309)

--Steve (talk) 02:34, 20 January 2016 (UTC)

Cages

cages in this context? thanks scope_creep (talk) 22:40, 16 October 2017 (UTC)

Kryptos K3 algorithm

Kryptos K3 was enciphered by arranging letters in a grid, rotating the grid 90 degrees, arranging them in a second grid, then rotating that. That doesn't seem to fit into any of the listed ciphers, I was wondering if anyone knew of a name for that algorithm?

Driver

Now felling happy 👌💪❤ 203.144.88.36 (talk) 22:42, 27 March 2023 (UTC)

Driver

thank you my life 203.144.88.36 (talk) 22:44, 27 March 2023 (UTC)

Self-reference

In the "General principle" section, the article makes a reference to itself by using "THIS IS WIKIPEDIA" as the text that is encrypted. Per Wikipedia:Manual of Style/Self-references to avoid, I suggest changing the text to something like "HELLO WORLD" to avoid a self-reference.