David Kahn, Leading Historian of Code and Code Breaking, Dies At 93 (nytimes.com) 5
Clay Risen reports via the New York Times: David Kahn, whose 1967 book, "The Codebreakers," established him as the world's pre-eminent authority on cryptology -- the science of making and breaking secret codes -- died on Jan. 24 in the Bronx. He was 93. His son Michael said the death, at a senior-living facility, was from the long-term effects of a stroke in 2015.
Before Mr. Kahn's book, cryptology itself was something of a secret. Despite an explosion in cryptological technology and techniques during the 20th century and the central role they played during World War II, the subject was typically overlooked by historians, if only because their possible sources were still highly classified. "Codebreaking is the most important form of secret intelligence in the world today," Mr. Kahn wrote in his book's preface. "Yet it has never had a chronicler."
Over the course of more than 1,000 pages, along with some 150 pages of notes, Mr. Kahn laid out cryptology's long history, starting with ancient Egypt 4,000 years ago and proceeding through the French and American revolutions, the innovations wrought by the advent of the telegraph and telephone to the mid-20th century and the dawn of computer-assisted code breaking.
Before Mr. Kahn's book, cryptology itself was something of a secret. Despite an explosion in cryptological technology and techniques during the 20th century and the central role they played during World War II, the subject was typically overlooked by historians, if only because their possible sources were still highly classified. "Codebreaking is the most important form of secret intelligence in the world today," Mr. Kahn wrote in his book's preface. "Yet it has never had a chronicler."
Over the course of more than 1,000 pages, along with some 150 pages of notes, Mr. Kahn laid out cryptology's long history, starting with ancient Egypt 4,000 years ago and proceeding through the French and American revolutions, the innovations wrought by the advent of the telegraph and telephone to the mid-20th century and the dawn of computer-assisted code breaking.
If you haven't read Codebreakers (Score:2)
The History of Codebreaking Began... And Ended (Score:4, Informative)
The Codebreakers was a ground-breaking, deeply informative and fascinating work when it was finished and published in 1967, although parts of it were removed and never published due to government pressure so it was not entirely up-to-date to the public state of knowledge even for 1967.
So the enormous Second World War codebreaking operations, the work of Turing and Bletchley Park, Engima, and knowledge of crypto-machines generally is absent as this information did not start coming out until several years later. And the vast changes driven by computers and algorithmic advances are nowhere to be seen. So it would have been wonderful if Kahn had updated his work, which remains very much historical -- more or less the story of cryptography and codebreaking before the modern era.
In fact he may have done just this. Only we can't read it or even know of its existence because it is so highly classified. After the publication of The Codebreakers the NSA found a more effective way to suppress the publication of code-knowledge by this gifted historian -- they hired him. Not right away, but when we might have hoped he would produce a The Codebreakers II to cover the dramatic changes of the preceding 30 years in 1995 he became a scholar-in-residence at the NSA. When his original work was republished a year later it had a cursory add on chapter that skimmed the developments since, but did not rise to the level of an actual history of them and provided no knew insights to the subject. He maintained his NSA ties for the next 15 years.
Re:The History of Codebreaking Began... And Ended (Score:5, Interesting)
So it would have been wonderful if Kahn had updated his work, which remains very much historical -- more or less the story of cryptography and codebreaking before the modern era... In fact he may have done just this. Only we can't read it or even know of its existence because it is so highly classified.
I doubt it.
Starting in the 70s and culminating in the 90s, cryptography moved almost entirely out of government hands and into academia and industry. Obviously it's possible that the NSA, etc., knows some things the public industry doesn't, but it's really not very likely. We had some evidence in the 70s that the NSA was still ahead, when they changed the S boxes in an IBM algorithm that became DES. It wasn't until 1990 that the technique of differential cryptanalysis was discovered, and it was demonstrated that IBM's original design was vulnerable to it but that the NSA's changes had fixed it. But the NSA had also asked to have the key size reduced, probably on the theory that they had the resources to brute force the smaller key size, and hinting that they didn't know of any other weaknesses.
But in the decades since, it seems pretty clear that the NSA leans pretty heavily on the work of academic cryptographers, and academics have found flaws in NSA-designed systems. The best example of the latter is the flaws in the design of the Law Enforcement Access Field of the protocol used in the Clipper chip design. It would make no sense for the NSA to try to inject a hidden backdoor in the LEAF, because the LEAF was a backdoor, for use by government officials. Related, the Skipjack algorithm used in Clipper has been sufficiently broken by academics that it doesn't meet modern security standards. We don't have a practical break of Skipjack, but the attacks we do have get close enough to make it unwise to use the algorithm. And the US government actually did use it for a long time so, again, the idea that the weaknesses were an intentional backdoor makes no sense. It seems vastly more likely that academics found attacks the NSA didn't know about. And then there's Dual_EC_DRBG, where the NSA did try to foist a backdoored algorithm off on us, and was immediately caught.
And that's not even touching on the real revolution in modern cryptography, which is the introduction of security proofs. Security proofs are the kind of thing that applied cryptographers are unlikely ever to have invested much into, it took pure academics to investigate what could be done... but the results turn out to be far more useful in practice than would have appeared likely. The NSA et al have clearly adopted the modern, proof-based paradigm, but there's no hint that they had ever thought of it on their own.
Of course, it is always possible that all of the joint projects, flawed designs, etc., are just a smokescreen to hide the fact that the NSA is way ahead, but if so they do an astounding job of maintaining this cover consistently. It's far more likely that things are exactly as they appear, and what basic logic would suggest: The NSA's cloistered mathematicians cannot, in fact, outcompete the academic and industry researchers of the whole world. They could in the early 20th century because there basically weren't any non-government researchers in this space (and very few government researchers). Now, there are, and people with talent and interest can make more money applying their talents in industry or if they prefer have more notoriety and prestige by applying them in public in an academic setting.
The point I'm trying to make here is that Kahn couldn't write the secret story of modern cryptography not because it was classified, but because it wasn't secret! Basically all novel cryptologic work is done in the open, which makes its history about as exciting as the history of most academic fields. That is, it's somewhat interesting to people who work in the space, and pretty boring to everyone else.
Re: (Score:3)
None of that matters.
If he wrote a history under NSA auspices it would be classified regardless of much of the material being about public developments. Recall that they successfully suppressed material in 1967 all of which was about publicly available information. Recall also the large amount of formerly classified stuff that emerged in the 1970s and 1980s that Kahn never wrote about - the electro-mechanical crypto machines which were the backbone of spy work into the 1990s. It is a fact that Kahn never pu
Re: (Score:2)
Also, although the NSA does naturally rely on the open literature on cryptography they have in addition their own math geniuses doing highly classified work on top of it.
But they're following, not leading. The point is that there's just not that much history to write that isn't public.
I recall that when the NSA recommended new public crypto-standards, which they had deliberately undermined, academic researchers found evidence that the weakening was so specific that it hinted that the NSA had a theory of the algorithms that was unknown to the outside world.
You're referring to DUAL_EC_DRBG, which I mentioned as an example of how we know the NSA isn't secretly far ahead of public cryptographers. If they really were, the weakening wouldn't have been noticed.