Make or break for code warriors

Buy this book online from Amazon.co.uk

Codemaking and codebreaking were intimately involved in the birth of the digital age.

If it hadn't been for the need to crack the Enigma code during the war, Alan Turing would not have turned his vision of a universal computing machine from a set of philosophical principles into an urgent, practical design.

But for a long time cryptography was not of much value to people who actually used computers, unless they belonged to the specialised world of the intelligence community.

The discipline of cryptography had been built around the idea of securing communications between a Person A and a Person B who knew each other, who had already worked out their code and wanted to exchange messages in a way that revealed as little as possible to an eavesdropper about how the code worked. It didn't seem to offer a lot to a computing world which was a network of strangers, loosely connected by modems.

This book is about how that changed: how a few pioneering thinkers began to re-invent crypto in the USA from the 1970s on, almost as soon as it became clear to the far-sighted that networked computers were going to form a new medium of communication, with its own unique demands. Steven Levy's first set of protagonists were basically mathematical hippies, who took a positive pleasure in investigating whatever the military-industrial complex said they shouldn't. Diffie and Hellman, the duo on the West Coast, and Rivest, Shamir and Adleman, the trio on the East Coast, set off a gradual nightmare for America's official cryptographers, the National Security Agency.

As Levy tells it, Whitfield Diffie, aged 30, his beard as long as Buffalo Bill's, was musing in the kitchen of a borrowed house in Palo Alto, when he was suddenly arrested by the idea that seeded the whole of the modern technology of encryption. He realised that you could have secure communication between strangers, and dispense with the orthodox secretiveness of official cryptography, if everyone was equipped with not one but two personal "keys", a public one and a private one. You could look up someone's public key in a computer-age equivalent of a phone directory, or they could send it to you; then you'd use it to encrypt a message to them that would only be decodable with their private key, which they never revealed to anyone. Diffie and Hellman worked out the theory, while Rivest and Shamir and Adleman came up with the first algorithms that could implement it. "Public key cryptography" was born, and, to their horror, the securocrats who had controlled America's codes for 30 years found they were facing a system that it might be mathematically impossible to break.

Of course, the success of public key, on which all current e-commerce operates, depends on the detailed maths by which Diffie's two keys can be related to each other. But though Steven Levy usefully sketches out the terrain of large prime numbers and "one-way functions" on which the solution was found, he isn't really interested in leading his readers into the mathematical thick of the problem, for this isn't a science book, exactly; it's more like a technically grounded business book, or a business-oriented social history. At any rate, he's most comfortable focusing on the human to-and-fro of developments over the past 25 years, as crypto has become a battle, not between hippies and buzz-cut generals, but between Silicon Valley entrepreneurs and "cypherpunks", on one side, and on the other officials who want to stop them exporting codes the NSA would have trouble with.

Levy's own sympathies are broadly clear. He's on the side of the innovative private-sector codemakers, rather than the codebreakers hidden away behind triple barbed-wire in the NSA's Fort Meade HQ. But he also subscribes to a different kind of code, the code of American journalism, which insists that reporting shouldn't make editorial judgments. So the book is scrupulously even in tone. Levy faithfully lays out the NSA opinion that "the techies ... are clueless about the very real dangers in the world", alongside the wild libertarian prediction that if the populace can't have the most super-powerful codes imaginable, it will "end freedom in America".

And he certainly leaves the reader room to wonder what psychological forces may be in play when people are so insistent that it shall take the minions of the evil guvmint a million billion trillion years to read their e-mails.