Talk:Enigma machine/Archive 1

Enigma Rotor Details
Please see the enigma rotor details page discussion.

PBS doc and three letter message setting. clarity?
I seem to remember from the PBS documentary on the subject that the three letter rotor setting was sent only in the army and airforce versions of the code, but in the naval enigma, the three letter setting was not sent, but was determined by codebooks, and that made it more difficult to break, needing the capture of codebooks to do it. Does someone else remember the same? --AN


 * Indeed the Navy Engima was harder to break from the beginning and throughout. It encrypted the intial settings in a way that the Army, Air Force, Abwehr, and SD Enigma operating procedures did not. And, indeed, captured material from U-boats, weather trawlers, and the small vessel Krebs in the Loftein Islands helped considerably with the Navy Engima.


 * As well, the operator selected three letter start position setting (sent twice in succession) was not used in the Navy. Even the Army abandoned it relatively early on. My memory is that the Poles had to cope with that, among other things, before it became obvious they'd have to throw in the towel.


 * Yep, the Navy procedure was different from quite early on. The doubled three-letter indicator procedure was abandoned by the Army and Air Force nets in May 1940 (except for a Norwegian key, apparently). The Poles had already handled an earlier change of indicator procedure before the war; more precisely, instead of using a "ground setting" to encrypt the indicator twice, they chose their own initial setting, and sent it along with the message in the clear. This stymied the Polish techniques (the "characteristic" method) for only a month or two, and they quickly responded with the Bomba and the Zygalski sheets. What really stopped them (while they were in Poland) was the introduction of two new wheels and the increasing of the number of plugboard connections; or maybe it was that they got invaded ;-) Anyway, once the British had punched them a larger collection of perforated sheets they could carry on in France. &mdash; Matt 12:58, 13 Aug 2004 (UTC)


 * So the PBS documentary is correct, but if it implied that the Army, Air Force, or Navy Enigma procedures remained the same for very long it is essentially wrong.


 * Hope this helps.


 * ww

Enigma and disambiguation
We need to make Enigma a disambiguation page.


 * There's already a link to Enigma Browser
 * A link also exists for Enigma Variations
 * Also, there is the musical group Enigma (though no entry as of this writing)

Recommend moving to Enigma machine

(related comments squeezed together by Mulad, May 29, 2003)


 * This has been done. &mdash; Matt 05:48, 29 Jul 2004 (UTC)
 * The "band" is now Enigma (musical project), but i'm not quite prepared to confusing things by fixing your now double-redirected link to it, OK? --Jerzy(t) 00:10, 2004 Nov 16 (UTC)

German name
I've just noticed that this article doesn't mention *anywhere* that Enigma isn't the real name of the machine, it was a name given to it by one of the groups that tried to crack it. I'll try and track down the real German name, but I'm posting this in case anyone happens to know it without needing to look it up. --AW


 * Poking around the web, I've found references that indicate the German name for the Enigma machine was "Schlüssel M". This may be specific to the naval version of the machine; it's not clear to me if other versions had a different name.


 * See, e.g.:
 * 
 * 


 * Ortonmc 20:29, 8 Sep 2003 (UTC)


 * Adam:


 * I've just noticed your comment about the name Enigma. Sorry it took me so long. It's my understanding that the name actually was chosen by Scherbius (or someone at the firm) and the product was offered commecially under Enigma (or some German name meaning Enigma). After Scherbius's death, the firm essentially went broke and was reorganized under the name (spelling is shakey, this is from memory) Heimsoth and Reinke. This version of the company continued in operation till the end of the War. Enigma machines (there were many variations) were eventually made by several firms, including the Olympia typewriter company (!) as production demands increased beyond what H&R could manage.


 * On the cryptanalysis side, the British Post Office (the Bell System of Britain) (in particular the Dollis Hill Works) made various bombes (akin to the Polish designs of the '30s) and assorted other code breaking machinery (eg, the Heath Robinsons and the Collossi) for the people at Bletchley Park, while in the US, NCR (and some others) made several bombe variants for Army and Navy (in quantities that astonished the British), and late in the war made quite a few special purpose machines (generically called RAM -- but not meaning random access memory, that came later) for both Army and Navy and (shortly after the War) for the ASA which resulted from the merger of the Army and Navy former girls' schools. The individual names for these things are quite varied, and I haven't yet come across an account of which one was what and built when.


 * Hope this helps. See Bamford's Puzzle Palace and Body of Secrets, and Budiansky's Battle of Wits for additional info and pointers into the literature.


 * ww


 * Can you confirm what I have read somewhere, that it was actually named after the Elgar Enigma variations - Elgar's music was quite popular in Germany. And cf. Wotan, the navigational beam, after Wagner. --Straw Cat 01:31, 19 February 2007 (UTC)

Algebraic representation of Enigma action
I'm trying to figure out how to express the action of the Enigma machine algrebraically. As far as I can determine, the situation is as follows:

The wiring of each rotor is a fixed permutation pi, chosen from a set of 5 permutations labelled I, II, ..., IV. The rotational position of the rotor i is represented as a power ki of a cyclic permutation c of order 26 ("A"c = "B", "B"c = "C", ..., "Z"c = "A"). The action of a single rotor on a letter x is by the permutation x.(c-ki pi cki).

The three rotors in sequence form the composition p


 * $$p = (c^{-k_1} p_1 c^{k_1})(c^{-k_2} p_2 c^{k_2})(c^{-k_3} p_3 c^{k_3})$$

The reflector r is a permutation which is a derangement of order 2 (i.e., x.r &ne; x, and r2 = 1). The plug-board is another permutation b, also of order 2. To encode a letter x, we then apply the permutation


 * $$ x . (b p r p^{-1} b^{-1}) = x . (b p) r (b p)^{-1} $$

Note that we have that (b p)r(b p)-1 is also always a derangement of order 2, since it is conjugate to r.

An additional possible complication, not present in the military Enigma, is that the keys/bulbs might be initially wired to the procession of rotors in a number of ways, for example through a permutation g. This would result in a permutation (b g p)r(b g p)-1; but in the miltary Enigma, g is the identity.

The values of ki alter after each digit is encoded; this appears to be controlled by two initial settings: for each rotor, there is an offset ti between the letters written on the outer ring and the actual encoding ring; and there is the initial rotor position or message key given as a letter ai. If we comine these two values, zi = (ti - ai) mod 26, we obtain as base value z = z1 + z2*26 + z3*262. Then we can calculate ki above for the encoding pattern of the nth digit of the message, with each ki being a digit of the number (z+n) base 26.

Is this correct? Chas zzz brown 03:04 Mar 23, 2003 (UTC)


 * Almost correct (for a 3-rotor WWII German military machine, etc); the only problem is that the sequence of the ki is slighly different to how you describe it. The Enigma rotors do not step quite like an automobile odometer; sometimes the second rotor steps twice consecutively. &mdash; Matt 21:19, 7 Sep 2004 (UTC)


 * Chas:


 * Have also just noticed your request for comment. You are attempting an interesting problem, but an indeterminate one. The German government (Army, Gestapo, Abwehr, SD, Navy, ...) Enigmas were not one machine; the commercial one may have been. For instance, at least one Abwehr Enigma used 4 rotors -- but had no plugboard. The first Navy Enigmas had no plugboard but used 3 out of 5 rotors. The first Army Enigmas distributed only three rotors initially; moving to match the Navy's use of 3 out of 5 only later. The Army Enigmas either started out with plugboards or added them very early, I can't remember which. The Navy's 4th rotor wasn't actually a full rotor but a simplified half rotor (which may or may not have been the same as the Abwehr's 4th rotor, I have no info) and didn't appear until well into the War. And there were at least two 'sizes' of plugboards.


 * So you see that there can be no one algebraic discription of 'the' Enigma as there were several Enigmas. Perhaps the most interesting project would be a description of the early Army Enigma that Rejewski broke; it certainly has pride of place cryptanalytically.


 * ww

Rejewski and Enigma manuals
According to Simon Singh in the codebook the term enigma actually appeared in the title of both the manuals in agent asch's posession that the spy codenamed 'rex' photographed.

I am not sure that rejewski never found out about the manuals as the article currently states. it was my understanding from reading the codebook that the manuals asch had access to were instrumental to rejewski constructing his bombes. i cant find my copy of the book but could someone check this fact?

User:jy 21:18, 22 Dec 2003 (UTC)


 * jy,


 * My memory of the sequence is that Rejewski was not even told of the material the French had gotten from Asch and passed to the Poles until he had made the fundamental conceptual breakthrough in 32/33. Thereafter, I suspect that they were given whatever material was available to increase the efficiency of their efforts. But I can't recall reading that exactly.


 * Hope that helps.


 * ww


 * ok ive just flicked through the code book again. singh gives the impression that rejewski began the work without the manuals of Schmidt (Asch). He made substantial efforts toward cracking the cipher but he required the manuals in order to 'access replica enigma machines'. singh essentially says no messages were decoded until after the french handed over the manuals and these were instrumental to rejewski's progression


 * Jy

General thoughts
At Ww's suggestion, I (someone unfamiliar with cryptography, except for Singh's excllent book) gave this a look to see how it looks to us amateurs. :-) Looks good, actually.  Early on in Operations, though, several references are made to a picture that doesn't seem to be there.  If the picture referenced is the one at the top of the article, it should be moved.....but actually, it's hard to see the details in that picture that are being referred to in the Operations Section.  Is a large pic available?

In the "Basic Cryptanalysis" section, I think you take too much time trying to cover cryptanalysis in the abstract....I start to wonder what all this has to do with Enigma. I'd cut a lot of it, and focus on the cryptanalysis done on Enigma (I think it's right to note a couple of principles that people should consider when reading, but really if you think they'll need to know about cryptanalysis, I'd just drop a link to that article).

The description of Enigma's method of encoding is hard to follow....though I think it's well written. But it's such a complex scheme that I had trouble following it (and I've read Singh's book at least twice). I think a diagram would be useful -- I would encourage one of you crypto fanatics to do it in a drawing program somehow. Singh has great diagrams, as I recall, which are of course copyright, but I think if we could build something like that for illustrative purposes here, it would be great -- certainly it helped me understand his book.

Also, in a couple of places the text feels like it's been edited over a few too many times. For example, you make casual reference to "Ultra" before explaining what Ultra was, probably because that paragraph has been cut and added to many many times. I'd suggest someone go through very carefully and make sure that codenames and cryptographical terms are explained at their first mention, and that the style is generally fluid. I think the article generally reads well, though -- not a big deal.

Finally, in the middle I start to swim in the expanse of text. I think a few more carefully placed headers would help me see the structure of the article more clearly, and move through it more efficiently as a reader. I hope no one thinks my ideas are a criticism of some excellent work here -- this is a great article! But I'm trying to help out with a little peer review. If you need me to explain any of this, feel free to drop me a note, or leave a response here. Keep up the great work! Jwrosenzweig 20:39, 15 Apr 2004 (UTC)

Requested article: Bombe (Wikispam)
Hi, I'd love to read an article about the bombe used to break the Enigma, and not having the prerequisite knowledge have offered an exciting &psi;5 bounty for the author in the hope that I can entice someone else to do the work for me... &mdash; Matt 17:58, 16 Apr 2004 (UTC)


 * Matt, Just the thought of w 5 has me antsy. I'll add it to my list of pending work. Actually, you're right about interesting. The Poles (mostly Rejewski, Zygalski, and Różycki as I follow it) first developed something called a cyclometer, and only then the bomba proper. But they never had the resources (these things were being built by a little machine shop operation partially owned by one of the BS folks (Antoni Palluth?); and some of their personnel were captured by the Germans who seem to have failed to use their renowned methods on them, else they would have learned Enigma had been broken), and BP had larger and faster bombes built. Note name change! When the need for speed increased still more, the Brits were up against resource limits and so after much political wrnagling back and forth, the USA and USN both designed and had built much larger bombes (probably shortened in the inimitible American fashion in at least some instances to bombs) by, among others, the National Cash Register Corp in Dayton. They flew, though with persistent mechanical troubles (brushes fail to work properly at surface speeds in excess of some -- apparently troublesome) limit. It was the threat of these machines that seems to have settled the political wrangling twixt BP (and their masters in SIS/UK) and the assorted (also wrangling amonst themselves) American crypto outfits.


 * What I'm just drooling to know though, is whether bomba were actually named from a sort of ice cream dish in Poland? Or (say it ain't so!!) just the noise them made (a sort of ticking apparently) while they ran. Never have heard a credible answer to that one.


 * ww 19:21, 18 Apr 2004 (UTC)

Comments
Common name: the name used in the under the cover instruction refers to the machine as a "lamp box" which makes sense because of the function of the lights when the keys are held down. The instuctions refer to the long-term maintenance of the machine, the electrical contacts tend to oxidize with no use and using the machine, sometimes forcefully depressing the keys abrades the contacts to keep the machine functioning. The German and its English translated instruction are easily obtained from my files or from the Web. Do not disregard the oval Enigma logo, that was the company's.

--enm		14 May 2004, 11:00 PDT

P.S. http://frode.home.cern.ch/frode/crypto/simula/index.html has the product logo which shows its name (Enigma). 14 May 2004, 11:30 PDT

General thoughts
Generally a very nice page, well written, but it speaks more about cracking Engima than the machine itself. The naval historian Norman Polmar and others have written a few other interesting things.

The shear numbers are worth knowing. Something between 100,000 to 200,000 Engimas were manufactured (and variants were used until 1990).

Budiansky is good for the numbers of Turing Bombes and Colossi: about 300 and 1-2 dozen, and some visible (in the USA) and rebuilds of each at Bletchley Park, UK.

Polmar in his Spy Book, encyclopedia also noted that the US Army apparently purchased an Enigma for about $250 about 1928 or 1929 and decided against using them. If this is true, numerous questions arise like what the final disposition of this machine was? Did Herbert Yardley ever encounter this machine? And more.

The plugboard, the Steckerboard, was thought by the German Army to add an additional level of security (it didn't), so it was added and that won the contract and begat the mass manufacturing.
 * I just noticed this comment, sorry. I seem to remember that the stecker was between keys and rotor, but there is no essential difference if it was between rotors and lamps.
 * Actually, the plugboard did add some security, as it effectively (in modern crypto terms) increased the key length. It increased an attacker's difficulty in determining the correspondence between keys and the first rotor and as a practical matter this was significant. This made a simple approach infeasible and was one of the reasons Dilly Knox (later of BP) was able to break commercial Enigma (presumably model D) but not the German Army version. He was not using Rejewski's inspired mathematical approach, nor his felicitous guess that the Germans would be (rather dimly/dumbly) methodical in assigning keys to intial rotor positions. ww 14:39, 21 Jun 2004 (UTC)

The design of the machine made certain that a letter did not map to itself. This was one of the design flaws which enabled Rajewski to get a toe hold at reducing the combinatorics.

--enm 14 May 2004, 11:12 PDT

Triton
I have some doubts about "a U boat mistakenly transmitted a messsage using Triton before it was due to be implemented. Realising the error, they re-transmitted the same message using pre-Triton 3 rotor Enigma". The Triton system let the fourth rotor be operated in a fixed position so that it could communicate with three rotor Enigmas, notably those which continued to be used by the weather service. Did the submarine really use a different machine or did it just lock out the fourth rotor? Jamesday 02:00, 21 Jun 2004 (UTC)


 * My memory of this is that the new 4 (or 3.5 if you want to be picky) rotor machines were distributed as possible (ie, when a U-boat or other candidate got back in) and the conversion was to have happened on the morning of such and such a day. Someone on one of the boats had the new machine, but mistakenly used it in 4 rotor mode prior to the effective date and then repeated encryption of the same plaintext in the works_like_the_3_rotor machine compatibilty mode after they realized what they'd done. This was a large crib for BP. Hope this helps. ww 14:29, 21 Jun 2004 (UTC)

JN-25 Fabrications
Hi All:

I'm not really a Wiki contributor, but I have writings online and Wiki folk have ported some of them to the Wiki. I have a crypto document and parts of it ended up in Wiki, courtesy of Matt. I was checking them over and got to prowling around for more crypto data for the next revision of my document.

I notice on the Enigma page comments that some stories about British penetrations of Japanese codes were fabrications. This rang a bell. In an early release of my own document, I had commented that there were conspiracy theories that the British had cracked JN-25, knew the Japanese were about to attack Pearl Harbor, and withheld the information to bring the US into the war.

This sounded preposterous on the face of it and I pretty much said so, but then I got a really flaming message from a Britisher over it, who vectored me to a page (I think it was the Bletchley Park website) that said the British *had* cracked JN-25. No details. It had the sound of something Tony Sale might say -- I respect Sale's knowledge and find him a good source of information, but he's got a bad case of "flag fever".

"OK, now I'm confused." What, I ask myself, did the British really know about Pearl Harbor? Now the Wiki article on the Enigma suggests that stories about British penetration of Japanese codes were to an extent a fabrication. That would simplify life for me (and give me a bit of satisfaction relative to my Yankophobic friend from the UK).

Any details or references on this matter? Comments would be appreciated.

Greg Goebel
 * Greg, JN-25 was introduced near the end of the 30s. It was a superencyphered code, originally a one book code, using Latin letters (and, I therefore presume, Romaji), and served as the highest level JIN command and control channel. There had been some success in breaking pre-Pearl Harbor JIN codes/cyphers by the Americans and I suppose others (as for instance in association with the Panay incident in which intercepted/decrypted traffic made clear the attack (by JIN Naval aviators) was not, as publicly claimed, a mistake), but there was little Fleet level traffic since most Japanese military activity was on land, and in any case, in China. If you have courier or mail or blinker light communications, you don't need radio, much less encrypted radio. Some progress was made against pre 12.41 versions of JN-25 (some sources claim that as much as 10% of message traffic in mid 41), but lacking sufficient depth, progress was slow. In 40(?) sometime, there was an agreement amongst the Dutch, British, and US to cooperate on Japanese communication cryptanalysis. No credible source of which I'm aware (specifically not including Stinnett in Day of Deceit (2000) -- his grasp of cryptanalytic reality at the time is distinctly limited (see Talk:Attack on Pearl Harbor for some comments)) claims more progress than that against any JN-25 version before 7.12.41. In any case, there was issued a new version of JN-25 on 1.12.41 and everyone was forced back to the start line. With more traffic available starting on the 7th (JIN operations now encompassed major areas away from the home islands -- SE Asia, the Philippines, etc -- and radio was required) the combination of USN Stations Hypo (Hawaii see Joseph Rochefort), Cast (Philippines), and OP-20-G (Washington), together with Hong Kong, Singapore (probably with some support from Bletchley Park), and Batavia made progress against the new version of JN-25. By spring of 42 it was clear some major new JIN operation was in the works, and by late May (just in time!) enough had been figured out about it (mostly from decrypted JN-25 traffic) that setting up the ambush at Midway became possible.
 * See Battle of Wits by Budiansky (US author), Station X and the Emperor's Codes(?) by Smith (?) (UK author), and Combined Fleet Decoded by Prados (US author) for mostly sensible more complete, and reflecting relatively recently available, information. Note that Prados' account, though valuable, is somewhat scattered amongst a larger story. A perusal of Harry Hinsley's books also will turn up (if memory serves) some commentary on British crypto operations against the Japanese. Henry Clausen's book about his investigation near the end of the War is also relevant in that he talked to just about everyone, carried an astonishing authorization brief from Stimson, and reports no evidence of anyone reading anything other than diplomatic codes/cyphers. However, note that there are two structural problems (at least) with any argument that JN-25 was not readable prior to 7.12.41. First, information release (some material is still classified in the UK or US, though that has been easing in the last decade or so) making older references less than complete (and, sometimes innocently, misleading, eg some of the claims in Farago's work) and, second, it's not possible to prove a negative proposition. Believers can still claim that the information (ie, that JN-25 was fully read prior to 7.12.41) is still classified and that the lack of evidence means nothing.
 * There have been various claims made about the state of US/UK/Dutch information prior to 7.12.41 (from various sources). The claims began as early as 8.12.41 (in a recorded comment made by a Congressman, Guy Gillette), and at least initially were politically motivated. Later on was added to the politics in some cases, a little understanding that some Japanese codes or cyphers had been broken, but not much understanding of the context.
 * In particular, it is sometimes said that Purple decryptions foretold Pearl Harbor -- not so, as the Japanese Foreign Office was out of the militaristic loop in control of Japan in part because they were thought to be insufficiently hard nosed. And that JN-25 (or other systems) had been broken -- by the British, or Dutch, or someone. The most credible of these is probably Rusbridger's book, but his source (ie, R Nave's diaries of the time) doesn't support his claims, so its credibility is rather strained as a result. Stinnett claims that JN-25 (which he calls 5-num) was broken by OP-20-G in part because they had produced a manual for its cryptanalysis. This lacks credibility since much depth is required for such a cryptanalysis, the method to be used could be understood prior to being able to apply it, and not very much traffic was available prior to the end of 41 to apply such a method. And there are assorted claims that US crypto (or signals traffic analysis) had the attack force located and tracked across the Pacific after it left the Kuriles (eg, Toland in Infamy). Some of Toland's informants went public and disputed his account of their experience (eg, Ogg who was Seaman Z, if my memory of the pseudonyms is correct) and the two strains of they knew in Washington from the Dutch seem to have little behind them, as nearly as I can make out. Nonetheless, there have been History Channel documentaries citing such 'evidence' of foreknowledge. They get seen by far larger numbers than read any of the more plausible accounts, which helps keep the foreknowledge conspiracy pot a-boilin'.
 * I trust this helps some. If not, we probably should continue this (ie, support for your research) in email, as WP Talk pages aren't really meant for such things. My email address is usable from WP. ww 14:43, 29 Jun 2004 (UTC)

Clarification on JN-25 Fabrications
Sigh, as usual I failed to make the question specific enough.

To be a bit more specific: somebody put comments on the Enigma article about fabrications concerning British codebreaking operations. I am trying to find the person who made the comments and obtain a general idea of what statements were fabricated and who did the fabrications.

Do I believe that anybody had cracked JN-25 and knew about Pearl Harbor? The idea is silly on the face of it, and no persuasion is necessary on that count.
 * Greg, Oh well... The fabrications that were meant probably refer to Rusbridger and his book, and allegations that either derive from it or which have circulated independently in the hidden history of Pearl Harbor community. As nearly as I can make out, many (most?) of these are not easily traceable. You've heard about the 100mpg carburettor that GM (or Standard Oil, I can't remember which) bought up and put on the shelf, of course. Urban legends and their surburban and rural and seagoing counterparts are quite hard to trace. ww 14:05, 1 Jul 2004 (UTC)

comment on 28 July ver of todo list
This is a long article, though shorter than it was since much material was removed to Ultra. Several of the todo suggestions here have the effect of adding material which, though interesting, will not add to a reader's overall understanding. They are in significant respect a sort of inside baseball, of interest only to fanatics (in this case, cryptiacs, I suppose). Eg, Zygalski sheets (there was a different name used at BP, but I can't bring it up just now; Clifford?), Hervel tip, indicator systems, Banburismus, rodding, and so on. If these should be in WP, and I can certainly agree that they should be, they shouldn't be here, but rather in a separate article -- Breaking Engima or something similar. jwr's comment (above) is relevant, and definitive on this degree of detail, in this article, I think. He at least doesn't have a cryptiac's bias.

As for trimming the disclosure section, I think this is a different kettle of fish. (Not Lorenz fish either). There is a considerable degree of bumpfh in the available literature in and around Enigma, its nature, its operation, its cryptanalysis, and the uses and significance of its output. Some of this is from close participants (Winterbotham on the distribution side), (the story about the Polish underground's ambush of a truck carrying an Enigma -- in Gordon Welchman's Hut Six account and elsewhere), and so on. The reasons for this include governmental secrecy, staged release of information formerly secret, deliberate misinformation, ... Our Gentle Reader should be informed of this, as it is central to the subject, and knowledge of the subject. Just where and how this should happen is another question I think. This section is somewhat jarring as it is about Enigma from another perspective than the rest of the article, but how better to do this is not clear. We have a responsibilty to our Gentle Reader which requires something along these lines. We will not have a better example from contemporary crypto for, presumably, some years (decades) yet. ww 16:25, 28 Jul 2004 (UTC)


 * We need balance. There's something wrong if we have a page long section describing how the Enigma story was revealed and fail to mention the various methods used to break the Enigma by the allies; the story itself is more central than how the story was first told. I'm certainly not suggesting we have pages and pages explaining exactly how each individual method works, but an overview of the cryptanalysis is essential. We can't assume that the "Gentle Reader" is a layperson, or not interested, or anything of the sort. I suggest we add everything; sections which seem too detailed or minor will obviously suggest themselves as candidates for being summarised and split off into separate articles. &mdash; Matt 22:26, 28 Jul 2004 (UTC)