What a tome! I feel like I just finished the Bible. It's a good fraction as long and covers an even bigger fraction of human history. Took me six weeks of pretty steady reading (and getting an upper body workout just carrying around the hardback library copy). It's pretty much four books in one -- classical secrets; codes in the era of telegraphs and WWI wireless; WWII and the Cold War; and miscellanea including rum runners and the like. (I gotta say, I wasn't expecting the decoding of Linear B script, talking to dolphins and communicating with aliens to arise in a book on cryptography, but they fit and I enjoyed them, even if modern thinking has evolved a bit (and the Mayan script decoded) since the book was written.)
The original edition is 1966, and the updated 1996 edition consists essentially of only a single, short chapter tacked onto the end, so it inhabits almost entirely the pre-electronic, pre-binary age. This colors Kahn's entire way of viewing the importance of various topics. At my age, we tend to think almost entirely in terms of encryption of digital data, whereas of course much of the work prior to 1960 centered on manual encoding and the application of tremendous brainpower to solving codes.
I learned a lot about the science and its evolution, of course, including the distinction between codes and ciphers (though I'm still not sure I could distinguish a nomenclator from a two-part code if you handed them to me). Basically, codes are secret, often shorthand numbers for names or phrases ("1121" = "invade Egypt on Thursday"), whereas ciphers ignore structure and operate on letters (or, later, bits).
But the book goes much farther than the evolution of ideas, focusing on people, historical context, and events. We learn much about how the Union read Confederate telegraph messages but not vice-versa, about the political repercussions of Yardley's disclosures of the codes the Americans read during WWI, how "practical cryptanalysis" (stealing code books) worked, how hard creating, distributing and protecting code is, how the U.S. confirmed Midway as the Japanese target, and how the Allies worked so hard to keep secret what they knew, to prevent the codes from changing. (This last item forms much of the basis for the plot of Cryptonomicon, which certainly rivals this book in length.)
The bits on secret inks and Cold War microfilms are.interesting. The latter stands in stark contrast to Tolkachev, the CIA's billion dollar Soviet spy, just handing over bags of 35mm film.
Importantly, it was originally written before the disclosure of Ultra, so the extensive Enigma history in the middle part could really use to be revised more extensively than it was. Turing and the Polish contribution to cracking Enigma deserve a little more, for sure. But it's also pre-DES, pre-public key, and pre-Internet, except for a few pages at the end. Of course, books covering the modern era abound, as well, so it's easy to pick up the missing bits.
As a book now more than fifty years old, it also shows its age and how much it was a product of its era. It comes across often as patronizing toward the Japanese, all women are "girls", and homosexuality was scandalous (Kahn uses pejoratives I won't repeat here). So, read it, but be prepared for a shock or two, and use them as teaching moments, if you're sharing the book with someone.
Overall, this was *well* worth the six weeks I invested in reading it, during which I read almost nothing else for pleasure and not even many other things for work. (For me, this sits right at the boundary of the two.) Highly recommended.