Talk:Bombe

Missing image Key-crypto-sideways.png WikiProject on Cryptography

This article is part of WikiProject Cryptography, an attempt to build a comprehensive and detailed guide to cryptography in the Wikipedia. If you would like to participate, you can choose to edit the article attached to this page, or visit the project page, where you can join the project and see a list of open tasks.

Contents

1 Initial problems 2 A second US Bombe 3 Rejewski wiring solution 4 Principle of the bombe 5 Bombe revisions 6 Computer or not? 7 What a polished article!

Initial problems

A good start to this page. There are some problems, however. From memory then,

1). The bombe (in any Polish, British, US variation and whatever the spelling) was not a computing device in any signficant sense. No programmable operation, no decision elements (ie, branching on a condition), etc. It (they) were essentially reverse Enigma machines (in multiples per bombe), and served only to test possible keys (ie ground settings plus whatever message key (ie session key in modern terms) might be used. Saved much time and very valuable and so on, but not computing devices.

Well, there was STOP condition there. But even if not computing in the "calculation" sense, definitely it was an information processing device (input->processing->output) ?

Wojsyl 09:01, 30 Apr 2004 (UTC)

Well... Depends on your definition, of course. But by a sufficiently broad definition my cat is a computer when she decides there's not enough of her favorite food and protests. The general definition is now programability, memory, branching capability, and so on. See the article on the Colossus and its Talk page for some discussion on this point. Likewise, some of the history of computing hardware articles. ww 14:15, 30 Apr 2004 (UTC)

I also think "computer" is too strong a term here, but it was a device for mechanising deductions (in the case of the UK and US bombes). — Matt 02:47, 20 Aug 2004 (UTC)

2). The first widget used by the Polish BS was called 'cyclometer'. It became unusable as the Germans changed their procedures from '32 on and was replaced by the first bomba (s).

Cyclometer was useful until September 15th 1938, when German coding procedures were changed. It had to be replaced by "perforated sheets" and bombas.

Wojsyl 09:01, 30 Apr 2004 (UTC)

We should say so, and say why (briefly, if possible). ww 14:15, 30 Apr 2004 (UTC)

I believe the cyclometer was used only to generate the catalog of cycle characteristics corresponding to various rotor orders and positions. This catalog was then the tool which was used to break Enigma settings. The reason it was replaced was that, rather than using a common, global ground setting to encrypt the indicator, the operator chose his own setting at random, and sent it in the clear, along with the encrypted indicator. I'll try and work on this article soon. — Matt 02:47, 20 Aug 2004 (UTC)

3). When the German Army went from 3 possible rotors in some specified order to 3 out of 5 rotors in some specified order, the increased problem faced by the Poles was not a theoretical one, it was a practical engineering one. The bombas they had would take too long against the increased possibilites (the 3 of 5 rotors) and new bombas which would take a practical time would be far more complex and take too long to design and build then they thought they had either time or money to handle. Hence, much of the reason for the handoff to the British and French in mid '39.

I think the Poles preferred the perforated sheets to the bomby, especially when the number of plugboard connections were increased. But yes, they didn't have the resources to make 60 (instead of 6). — Matt 02:47, 20 Aug 2004 (UTC)

4). Bomba/bombe/bomb (s) continued to be used against Enigma traffic till the end of the War. They were not made obsolete by the acquisition of an Enigma plus auxiliary material from any source, including a U-boat.

5). The problem confronted by Turing et al in Hut 6 at BP had been different in kind and difficulty from the start. Naval Enigma had, since the Navy adopted the Army modifications to commercial Enigma in ('32?) (replacing their own version adopted in ('28?)) always used 3 of 5 rotors, and furthermore had always used a much more complex and harder to analyze method of setting session keys. In short, their "key scheduling" was much better than the Army/Luftwaffe/SD etc. Seizures of material (from U-boats in various places, from the Krebs in the Loftein Islands (sorry about the spelling if it's wrong), from the weather trawlers off the coast of Iceland, etc were helpful in getting past that key scheduling. Though they only got through it intermittently, and were blacked out for long periods.

6). On the 'political' side, it was apparently, the determination of the USA and USN to build their own bombes that finally compelled BP/Secret Service to allow US participation in the Enigma operation; full information on techniques and allowing US personnel to assist and so on. In fact, both the USA and USN eventually built ambitious (and much faster) bombe variations (very large physically and so fast there were serious brush making contact reliability problems with them) bombes. Some were made in Dayton at NCR.

ww 14:05, 28 Apr 2004 (UTC)

A second US Bombe

Source for the believed existence of another US bombe was Jack Ingram, Curator of the National Cryptologic Museum, during a tour of the museum on Thursday. He told of searching for it and not finding it whole, but noted that they can be broken down into many smaller pieces and thought that it may be buried in storage somewhere which just hasn't yet been discovered. I forget who he mentioned as his source for there being two of them. Jamesday 20:10, 18 Jun 2004 (UTC)

Rejewski wiring solution

I removed the following:

The system developed by Marian Rejewski in the 1930s could have worked out their wiring but because the Polish cryptanalysts had briefly been in occupied France it wasn't believed safe to trust them.

Reason: Rejewski's recovery of the wiring of rotors I, II and III was based on the doubled indicator system with a common ground setting, in use by the German Army and Air Force until 15 September 1938 and the Navy until 1 May 1937; after that, his original attack couldn't be applied. (As an aside, wheels IV and V were introduced on 15 December 1938; Rejewski broke these because the SD Enigma continued using a common ground setting for a few months after the other services changed.) They got hold of the wiring for the Navy's rotors VI and VII captured from the crew of U-33 on 12 February 1940.

Principle of the bombe

I removed the following:

A cryptological bombe is essentially a large number of cypher machines in parallel, each decyphering a given message using different keys. By using dozens or hundreds of machines at once, prospective keys can be attempted in bulk, and the product from each of the decipherments can be screened for plaintext phrases which would indicate that the correct key was used. The British were concerned with the radio traffic of the German Navy, who used relatively strict procedures for encryption; the technique of loooking for a repeated three-letter wheel setting would not work (as it did for the Army Enigma), and naval Enigmas changed their daily settings based on predetermined schedules, transmitted by courier to ships while in port.

This is where the known plaintext attack became useful. No human could screen the tens of thousands of keys attempted by the bombes on a daily basis, but the machines could be configured to automatically detect common paintext phrases, most often "EINS" (German for one), or "HEILHITLER". Upon decrypting such a phrase, the machines were configured to halt and make the operator aware of the potential hit. </blockquote> I'm afraid this has a number of problems. 1) A single bombe examined the keys in series, not in parallel. Although there might be a dozen Enigma-equivalents used at each point, each rotor setting was examined in sequence. 2) The bombes were used for most Enigma traffic, not just that of the Navy, since the double encipherment indicator procedure came to an end in May 1940. 3) The bombes did not recognize plaintext. The bombe performed logical deductions on a known plaintext and flagged those which did not lead to a logical contradiction. — Matt 09:17, 14 Oct 2004 (UTC)

Bombe revisions

We should talk about the revisions to bombe. I was trying to give an explaination without the hand waving so people could get the gist of the idea. Admittedly, it was oversimplified. I also don't understand why you excised my description of the Stecker board. The sentence you left has no context. Perhaps we could negotiate a joint text. My e-mail is areinhold "at" alum.mit.edu. --agr 16:22, 11 Nov 2004 (UTC)

Sure; the problem was that your description wasn't just oversimplified, but (I'm afraid to say) it was actually incorrect. But I do think hand-waving gists are good, and I think it's a good idea to start off with such a summary before lurching into a description of the nitty-gritty detail of how it worked. I'm very happy (in fact, prefer!) to work collaboratively (after all, this is Wikipedia ;-) ). — Matt

I thought the model i described was a good introduction to the concept. Maybe it would help if you could say what you thought was incorrect about it. --agr 19:08, 11 Nov 2004 (UTC)

Sure; your description said If one had a guess, say that the letter G as input should produce a letter K as output, one could connect a voltage to the G input of the spinning set of rotors and use the K output to stop the motor. The stack would then spin until there was a match. — if there was no plugboard, such a machine would have worked, but this is not what the bombe did; the plugboard stops you from knowing the input and output to the rotors. So Turing had to come up with something a little more subtle. — Matt 19:27, 11 Nov 2004 (UTC)

I thought I tried to say that, but i wasn't as emphatic. I'd be up for reworking the section to make that clear. Would you, like to see it first? I can put it here or the discussion page for bombe. --agr 15:47, 12 Nov 2004 (UTC)

Sure, I'm going to be away for the weekend, so please do go ahead and work on the page as you think is necessary, and I'll have a look on Monday and make some more changes and additions myself. — Matt 16:04, 12 Nov 2004 (UTC)

Computer or not?

It would be nice to know whether or not the Bombe is considered by experts to fit into the definition of a computer? If so, we should replace the "Category:Computer hardware" with "Early computers". If not, we should arguably rm the article from the former category. --Wernher 14:35, 4 Dec 2004 (UTC)

I don't know of any expert comment either way — it does fit into the first description given at computer: A computer is a device for making calculations or controlling operations that are expressible in numerical or logical terms. The bombe makes a calculation that is expressible in logical terms. However, I wouldn't call the bombe a computer, for whatever that's worth. It's certainly common in books about WWII cryptography to label Colossus as an early / the first computer, but not to describe the bombe in that way (even though the bombe was built earlier). — Matt 20:33, 6 Dec 2004 (UTC)

What a polished article!

Ground 01:22, 9 Feb 2005 (UTC)

---

ok Matt, let me rephrase: (pls edit and state ambiguity) I think this is a good verbal explanation !!

Assume you find a triple loop abc, ideally in plain: abc, cipher: bca.

plaintext : a big cool

cipher  : b cxy aqwe

a-b b-c c-a = a closed triplet (A-B-C -->A) OK ?

So with a crib you find plaintext letter a mapped to cipher b, plain b to c, and plain c to cipher a again within short distance .

Now assemble the rotor mechanisms of three enigmas serial-in-line (abovementioned triplets) and set it to the original rotor positions,

with their offset (here 1 step each) accordingly. Then you get a corresponding physical wire closed loop from the input of the first triplet to the output of the 3rd triplet. {Now this in fact is turings cool idea ! try to understand this and ure a genius too ! no really, the letter loops corrspond to physical wire loops which can be detected by a el.circuit quickly }

You can detect this with lamps connected to the rotor contacts. {wire 'em up to + and - pole}

The lamp in the wire loop will stay dark. {meaning triplet fit}

Now you turn the rotor systems synchronously. {this is called decryption, to find da rotor position}

If only one lamp stays dark because of the one wire loop, you can quickly calculate the Steckerfeld, and reject those positions with all lamps lit. As mentioned, this typically happens several times in 17000 permutations.

It's not clear where the lamps are connected, nor is it clear why the lamps stay dark. Moreover, the Bombe didn't use lamps. — Matt Crypto 10:30, 11 Feb 2005 (UTC)