Talk:Type B Cipher Machine

Untitled

Note: comments on this page have sometimes been most recent at the top.

---

Ortolan88: Actually, the Purple machine (called Alpabetical Typewriter with a date indicator by the Japanese) was made of electrical stepping switches, not rotors. It doesn't appear to have been influenced much by the Enigma design, or indeed by any of the rotor machines. The Japanese Navy Captain who desinged them all followed his own star, though there is some mention of a Polish Army Captain who went to Japan to help in crypto design in the '20s. Perhaps an influence? Leo Rosen at SIS was asked to build an equivalent after the SIS team (under Frank Rowlett) figured Purple out, and in an ironic coincidence, chose exactly the same stepping switch the Japanese had used.

On another subject, there is an unsigned comment below talking about digital v analog. The following may help.

All of the rotor machines, and the Purple and its predecessors (Red and the M machine) are actually digital in the sense that a substitution (on either encryption or decryption) is digital. Either q is substituted for f or it isn't. Whether it should have been is another question having to do with how well an analog machine (rotors rotating, electrical contacts making or breaking properly) behaves digitally. Exactly similar considerations underly digital circuits in modern computing equipment. They are implemented using transistors of various types, all of which work by controlling the flow of electrons from hither to thither. A meaningful flow isn't any particular number of electrons for a digital 1 or exactly zero for a digital 0 -- if that happens to be the assignement at that particular logic gate just then. There are always guard bands. Thus, 'a 1 is any voltage greater than such and less than so', while 'a 0 is less than this and greater than zero'. The value of such is always greater than this by some amount, and if the voltage falls between such and this, there will be trouble. A good bit of digital electrical design is devoted to ensuring that ambiguity of this sort doesn't happen. Reshaping circuits, regenerating flipflops, Schmitt triggers, are some of the dodges used.

For a long time, T(ransistor)T(ransisitor)L(ogic) in which one value is nominally +5VDC and the other 0VDC was the standard. In recent years the larger of the two voltages has become smaller. +3.3VDC is used in the PCI bus and +5VDC was used in the first version of the PCI bus. This saves power, extends battery life, and is otherwise good. It also squeezes guard bands and makes tighter circuit design mandatory for digital cicuits, lest the ambiguity get to be too much.

Hope this helped some,

ww.

---

What do you mean, "analog computer"? Wouldn't it need to be digital? Remember, "digital" does not have to be computerized. A mechanical adding machine can be digital.

---

Well, the the purple machine is like the Enigma. It consists of a bunch of interlocked letter wheels. The interlocking relationship changes after each letter is typed. In effect, it is exactly the same technology as an odometer, which is certainly analog.

A digital computer works on numbers. An analog computer works on mechanical relationships. I would be interested in hearing of any mechanical adding machine that was digital. An analog computer, such as an odometer, is usally single purpose while a digital computer is multipurpose. Ortolan88

---

Sure thing -- I have used (obsolete) mechanical digital adding machines to add, subtract, multiply, and, with enormous effort, take square roots. The numbers were represented as digits using a series of wheels, one for each power of ten. Babbage's Analytical Engine was another example of a mechanical digital computer.

I'd also disagree with you about an odometer being analog. If the number is stored as a set of digits, then the representation of the number is digital, even if the representation of each digit is analog. Note also that the digits (except the lowest) move in discrete steps.

Here are four examples, showing the full gamut:

Mechanical / digital: Babbage's Analytical Engine

Mechanical / analog: Mechanical rotating-wheel integrators, slide rules

Electronic / digital: nearly all modern computers, pocket calculators

Electronic / analog: Op-amp based 1960s and 1970s analog computers

Note also that there is nothing requiring a digital computer to be programmable, or Turing-complete (even though nearly all are). For example, FPGA codebreaking engines and galactic physics simulators are special-purpose electronic digital computers.

The Anome

---

Hello, The Anome,

Yes, but, and I am way out of my depth here, I am pretty sure that you're supposed to make a distinction between the representation of a value and the way that value is calculated and stored. You have to agree that a mercury thermometer is analog, but it represents values digitally.

As for the odometer, there's a little thing on the axle that transfers the rotary motion from the axle to the odometer, which replicates that motion on some other wheels that just happen to have numbers printed on them and an escapement mechanism (for those discrete steps). When you change your tire size, you have to change a little gear, else the odometer, and its companion analog computer, the speedometer, won't read right. To me that is the essence of "analog".

The odometer doesn't "store" anything. The original Mauchly-Eckert computer, a chunk of which is on display at Harvard U., used odometer mechanisms for display, but all they did was tell you how many times the relay they were attached to had fired, the data, I believe, was stored in memory.

(Fools rush in, I don't know Jack about this, other than having seen the display. My dad was a classmate of John Y. Atanasoff (who taught Mauchly and Eckert everything they knew), but I'm just a tech writer.)

A regular old-fashioned clock is analog too, by my definition, digital representation, discrete steps (some), and all.

That said, you probably know more about this than I do. There is no wikipedia article on computer history, but your outline above certainly belongs there. I think you'd have to have a pretty rigorous definition of the difference between analog and digital to do so, however. I have a little trouble with your roadmap because I think it is mixing the two. For instance, I think, and so does the wikipedia article on analog computers that the slide rule is analog.

What do you think of the wikipedia article on analog computers. It seems to agree with me. (I think it's fabulous!! Actually, it seems a little thin to me, but I couldn't fix it.)

There is no article on "digital computer"; that is redirected to Computer, which is longer, but pretty slapdash in organization. It says analog computers have "no discrete or digital computational ability". By the same token, the purple and Enigma machines didn't compute or store anything. To change the code, you moved the wheels.

Fun computer history fact:

• On warships there used to be an organization of humans called "the computer", often made up of members of the ship's band, who sat at a table in the depths of the ship during combat, performing simple math and passing along the results to the appropriate next person for the next calculation, which work in summary produced aiming trajectories and the like for the ship's gunners. The original log tables were computed the same way.

Best regards, Ortolan88

---

Shouldn't the title of this page be "Purple cipher"? Since it really was a cipher and not a code (even the article says this). - Crenner

Engima?

A book I just bought (The Codebreakers) and this site both say that PURPLE was based upon the Enigma. Can someone reference the counter-claim? BrokenSegue 03:54, 22 February 2006 (UTC)

We currently say, "The German Enigma machine was unrelated to PURPLE, though there have been published claims that PURPLE was merely an Enigma copy of some sort". The first part is too strong: PURPLE and Enigma are related in the sense that they are electromechanical cipher machines that produce a polyalphabetic substitution with a long period. They are also similar in that they use a cascade of simpler elements to create a constantly shifting wiring "maze". However, PURPLE is significantly different to Enigma (and all the other rotor machines) by using stepping switches, as opposed to wired rotors. I have references which assert that PURPLE and its sister machines JADE and CORAL are unique amongst crypto machines for using this component (It's not clear that Kahn was aware of this when he write Codebreakers in 1967, as he mentions "coding wheels" in his description of PURPLE. While Enigma could be described as having code-wheels, PURPLE really can't.) It would be false to say that PURPLE was an Enigma copy, pretty much by inspection. As to being based on Enigma, or derived from it, I think we should treat assertions like that with caution, simply because of how much we now know about PURPLE and Enigma's internals, but I haven't found a direct source countering the claim directly. I do have a source which says that the Japanese bought every crypto device on the market (Damm, Korn, Hagelin, Hebern, Kryha...); and adapted those principles for their own machines. PURPLE is as close to the Hebern machine (or other rotor machines) as it is to Enigma. — Matt Crypto 09:42, 22 February 2006 (UTC)

Ach, you know so much. Alright I'll keep looking for information (yeah Kahn's book is outdated, I noticed several instances where he is wrong or misleading). BrokenSegue 20:35, 22 February 2006 (UTC)

I've moved the section here for now. — Matt Crypto 09:42, 22 February 2006 (UTC)

The German Enigma machine and PURPLE

The German Enigma machine was unrelated to PURPLE, though there have been published claims that PURPLE was merely an Enigma copy of some sort. In fact, the PURPLE machine was a Japanese development, one of a series designed by a Japanese Navy captain, though there seems to have been some assistance by at least one Polish officer prior to the 1930s. There is some evidence that the Germans shipped several military Enigma machines to Japan by submarine late in the War; several sources suggest that they never arrived.

Enigmas are rotor machines whereas PURPLE was constructed using telephone relays, I think by Fujitsu. I have heard "knowledgeable" historic claims that this made for a slightly weaker machine. Whereas apparently no examples survived the war (main page) by Operation Paperclip, somewhere might exist tallies or production counts. I have a reference for the sub (sunk off Norway), but they likely had them Japan long before the end of the way (in the German embassy in Tokyo and likely else where). 143.232.210.46 (talk) 23:28, 8 June 2009 (UTC)

I think that the claims about them being unrelated are way too strong. The Purple machine is almost exactly like the enigma machine except that it uses a different part to do the same job. It may not contain rotors like the enigma machine does, but it has banks of stepping switches which are wired in a way so as to achieve exactly the same effect. It's also based off of the Red cipher machine and the Red cipher machine used a rotor. They switched to stepping switches because Red had trouble with dirty contacts on its rotor. KeithyIrwin (talk) 06:48, 3 February 2010 (UTC)

I don't think that last statement is true, and anyway, if you read the RED article you'll see that it was based on a completely different type of rotor technology than ENIGMA (and indeed, that one of the reasons it is different is that they correctly didn't trust the Japanese not to steal the mutiple rotor technology used in the ENIGMA-style machines). Mangoe (talk) 10:37, 3 February 2010 (UTC)

I think you both are right: the principle of the cipher is the same or similar, while the implementation in hardware is different. --Ayacop (talk) 08:52, 28 October 2010 (UTC)

Yep, AFAIK. The mechanics were different, but the output result was much the same. TREKphiler ^{any time you're ready, Uhura} 02:04, 29 October 2010 (UTC)

Japanese cipher machine & device

I write it down for reference. Tuve74 11:07, 13 January 2007 (UTC)

A classification by a encipher method

	Kryha type (half-rotor)	Enigma type (rotor)	PURPLE type (stepping-switch)	M-138 type (strip)
Ministry of Foreign Affairs	Angooki Taipu-A, RED (Type A Cipher Machine)		Angooki Taipu-B, PURPLE
Japanese Navy	91-shiki injiki 1-gata (Type 91 print machine model1) 91-shiki injiki 2-gata 91-shiki injiki 3-gata	3-shiki kaejiki, GREEN (Type 3 converter)	97-shiki injiki 1-gata, JADE 97-shiki injiki 2-gata 97-shiki injiki 3-gata, CORAL	3-shiki SB ban (Type 3 strip board)
Japanese Army		92-shiki injiki 97-shiki injiki 1-shiki 1-go injiki	Jidou insatu musen densinki (Automatic wireless teleprinter)
Japanese-German navy collaboration		TIRUPITU (TIRPITZ), OPAL

• Ministry of Foreign Affairs was granted machines from the Navy.
• Navy's injiki had three models, model1 for use at ground base, model2 for vessel and model3 for attache at oversea.
• Navy 3-shiki kaejiki was a copy of Enigma, which Kana-letter modified.
• All Army machines and Navy 3-shiki kaejiki were not put to practical use.
• TIRPITZ was made in Germany. It is a special type of Engima, 4-rotor and no plug-board.

Thankyou for the information, Tuve74; we well probably put this in an article somewhere, possibly the underdeveloped Cryptography in Japan. What source are you using for the above? — Matt Crypto 18:59, 13 January 2007 (UTC)

I am a Japanese, who study Japanese military cipher machines. The above table is translated to English from http://ja.wikipedia.org/wiki/日本の機械式暗号, contents 2.

At first, I needed to edit the table from many articles, both in English and Japanese. The source articles are also listed same wikipedia, contents from 6.1 to 6.3. Can you read Japanese? - Tuve74 21:00, 13 January 2007 (UTC)

Why Romaji?

Why did crypto machines not use Kana? Jim.henderson 04:38, 17 January 2007 (UTC)

If Kana-type used, somebody has to translate Japanese kana-text in English text at an embassy or a consulate. By use of Roman-type, A diplomatic note in English text can just send and receive it. Of course, Kana-type is convenient and used by communication of the Japanese Navy inside. Tuve74 12:32, 19 January 2007 (UTC)

Thank you. I didn't realize the diplomatic side of cryptography was so much separated from the military side despite the Navy doing the work. Somebody once told me Japanese Army crypto work was handicapped by the need to translate everything into Italian. That fellow's confusion between the Roman alphabet and the Italian language warned me that he didn't know what he was talking about, but now I see that a similar problem actually arose in diplomatic communication, presumably due to the difficulty of transliterating foreign languages into the familiar Kana. Anyway with modern computers using ISO character standards I guess it's all easier. Jim.henderson 16:40, 20 January 2007 (UTC)

Several other things to consider:

1> If kana was to be used, the coding table would be absolutely huge for even a simple substitution programme. Thus any chance of portability at that time would go out the window. 2> Upon deciphering the Kana, you would have a VERY SPECIFIC decipherment of the code being transmitted. Leaving the concerned party(ies) at even more risk. With romaji, there is an element of still having to translate what was indeed the intent of the message. ie: there are many more combinations of Romaji characters that would be meaningless. 80.5.219.60 (talk) 18:17, 27 May 2016 (UTC)

Who was the mathematician?

Answer to Ww, he was Teiji Takagi (高木　貞治)

The articles are below.

• http://ja.wikipedia.org/wiki/高木貞治
• Hiroshi Kimura, "World War II and Teiji Takagi "(第二次世界大戦と高木貞治), 17th Symposium on the History of mathematics 2006, Institute for mathematics and Computer science Tsuda College.

"Before and during the war, when the doubt of the printing machine (PURPLE) was decoded in Ministry of Foreign Affairs. Ministry of Foreign Affairs telegraphic communication section without technical knowledge referred to the navy, and the navy quoted evaluation of Teiji Takagi, and concluded that "such thing is impossible".

Tuve74 01:02, 27 January 2007 (UTC)

Citation about Teiji Takagi was unaware of the basics of cryptanalysis

There is (was) no articles that he was an expert for cryptanalysis. According to the article of Hiroshi Kimura above-mentioned, "No Japanese mathematician aware of relations of a cryptology and mathematics before november 1940". But it is almost impossibility to prove that they ware not. How am I good?

• In 1937, the Japanese navy was going to evaluate of combinations of key after the PURPLE was completed, not before. The navy looked for a well-qualified mathematician for evaluation.
• The Japanese mathematicians were not able to give a clear evaluation, Teiji Takagi was one of them also.
• After having examined for two or three days, Teiji Takagi informed the navy that "Key combinations is more than 1,000,000,000,000. More detailed calculations are not possible". The navy was satisfied with his answer, because he was a most famous mathematician in those days. Tuve74 11:58, 25 February 2007 (UTC)

Poor readabilty

The artiacle is not very readable. Its hard to understand if all the references to navy or amry refer to the american army or the japanes army. Seems like it wasnt writen by a native english speaker. Please somebody improve it.

Yes, it's mostly a poor translation from Japanese, and stylistic improvements from someone who knows the subject would be a good thing. No, it's mostly about the Japanese code making authorities, and when it's about the American code breakers, that's generally pretty clear. Jim.henderson 13:41, 18 May 2007 (UTC)

No Mention of Genevieve Grotjan Feinstein?

According to the NSA website, Genevieve Grotjan Feinstein was instrumental in the breaking of the Purple cipher, yet the article does not mention her... --ukexpat 14:00, 20 September 2007 (UTC)

"The Story of Magic", by Rowlett, also indicates that she was the person who made the most important breakin (although it was a team effort, and others had made major contributions before). It's described on pp. 151-153 (1998 hardback edition). Noel (talk) 22:19, 1 November 2013 (UTC)

Truly Amazing

What has become truly amazing to me is how thoroughly Japanese cryptographic material was destroyed during and especially near the end of WWII. With over 100,000 Enigma machines made, admittedly it had a commercial start, we have somewhere a warehouse full of captured machines of various varieties, and a small market economy of traders (e.g., eBay) and various associated peripheral artifacts (e.g., code key books), yet not one JN25 machine (what we call), not one complete Purple or Red and few captured artifacts survive. And the firms which made these machines still survive (not one blueprint survives?). Hard to say how many lower level intelligence officers survive (I think I know the son of a US MIS officer who likely debriefed Japanese officers during and after WWII and had access to post-war collected materials <the son was born at Ft Meade>). It is true that the historic NSAers call the Enigma fans "Enigma" fetishers. Yes, I have read Budiansky's book, and live not far from Two Rock (and need to try to locate Rebeccia what's her name cited in his book). I raised this with David Kahn at a meeting in October, but he didn't know what to say. Ditto Jim Bamford who was also at the meeting. 198.123.51.178 (talk) 23:37, 13 April 2010 (UTC)

not one JN25 machine?? JN25 was, IIRC, a book code? Noel (talk) 16:18, 19 January 2014 (UTC)

What? no!

This article is poorly written and needs more citations. what is this "Further Reading" section, I do not see it on other articles. —Preceding unsigned comment added by 24.211.171.23 (talk) 21:21, 26 February 2011 (UTC)

What was Blue?

The article says

Clearly, the Purple machine was more secure than Blue,

but doesn't mention Blue again anywhere. What was it?