Privacy on the Line, by Whitfield Diffie and Susan Landau, asks who can hear an electronic conversation and who decides when they may listen. First published in 1998 amid the fights over Clipper key escrow and cryptographic export controls, it returned in 2007 as an updated and expanded MIT Press edition, reviewed here. Its new preface adds network-built wiretapping and the eroding wall between foreign intelligence and domestic policing. The book is worth reading because, across its technical and legal chapters, it keeps four questions apart: who can observe a channel, who holds keys and can prove identity, what interest the law protects, and who authorizes and records interception. Its ciphers, key lengths, forecasts and legal doctrine describe 2007 and read best as history.
A telephone call crosses miles of wire and radio links. The speakers control only a few feet, and the carriers handling the rest can listen or record. An opened letter might show tampering, a risk that deterred spies. Copying a call leaves the line unchanged, so sender and receiver have no way to tell who else holds a recording. Registered carriers have also replaced scattered couriers, telling interceptors where to look. The authors conclude that a search is usually obvious to its target, while a wiretap's invisibility weakens accountability.
In chapter 2, Diffie-Hellman key exchange lets two parties derive a shared secret from messages an eavesdropper can read. An intruder who runs a separate exchange with each party can still sit between them, translating and recording. A certificate is a signed message from a central key authority binding a subscriber's identity to a public key, and anyone can copy one. Verifying the authority's signature shows that the credential is authentic. A challenge-and-response step then tests whether the party on the line controls the matching private key. Using the public key-exchange value as the challenge, the authors add, ties that party to the encryption. Placement of keys changes what an observer learns. Link keying decrypts at each switch and can hide addresses from outsiders. End-to-end keying shields content from the network, yet traffic analysis still shows who talks to whom, and dummy traffic that would mask the pattern is costly end to end.
Chapter 6 supplies a reason to care beyond crime. History records speeches and pamphlets, and the authors argue that the finished positions in them take shape in private conversations. Confidential relations with lawyers, they say, rest on the same condition. They also describe an unequal contest, in which national security and law enforcement have organized institutions behind them and privacy lacks comparable backing. The case is normative.
Chapter 7 traces a legal test that shifted with technology. The Olmstead majority found no search because agents tapped lines without trespass and conversation was intangible. Brandeis replied in dissent that tapping one line exposes everyone who calls or is called on it. The Nardone decisions excluded wiretap evidence under the Federal Communications Act. They rested on the statute and left Olmstead's constitutional holding untouched, and the Justice Department read them as barring only interception combined with disclosure, which permitted taps whose contents stayed inside government. Katz later protected a conversation with no intrusion on property.
As the book reports it, a 2002 FISA Court opinion found misstatements or omissions in more than 75 applications and unauthorized sharing with criminal investigators, then restricted prosecutors' role in FISA cases. On the government's first such appeal, the Court of Review sided with the administration, holding that the lower court had misapplied minimization procedures. The conclusion adds a hypothetical: had millions of escrowed devices sold, the authors expect administration lawyers would have argued for putting the key database at NSA's disposal. Surveillance machinery takes years to build, while decisions to use it in new ways can come quickly.
The authors grant that interception can expose conspiracies and supply evidence without endangering officers. Their 2007 reform is administrative: digitally signed electronic warrants could cover only calls between two suspects, answering Brandeis's concern about everyone on a tapped line. They add, conditionally, that if wiretaps help as little as public evidence suggests, the cost may be unjustified.
For newcomers, chapter 2 builds from simple substitution ciphers toward key exchange, and some later material is technical. Its public-key and communication-security sections suit readers tracing who can read a message and prove identity. Readers of privacy law can go to the constitutional history in chapter 7, and the closing sections of chapter 12 connect installed surveillance capacity to political choice. This review draws on those sections and other selected passages. The Library of Congress hosts a complete PDF of the 2007 edition.