Philip Zimmermann

Author's preface to the book: "PGP Source Code and Internals"

Published by The MIT Press, 1995. ISBN 0-262-24039-4.
(no longer in print)

This book contains all of the C source code to a software package called PGP (Pretty Good Privacy). PGP is the most widely used software in the the world for the encryption of electronic mail. It uses public key cryptography to let you communicate securely with people you've never met, without the prior exchange of keys over secure channels.

Why publish an entire book (and a big one at that) comprised mainly of boring source code for a computer program? Well, there are some really good reasons. It concerns your civil liberties, and requires a bit of explaining, but it's actually quite an interesting story.

Cryptography is a surprisingly political technology. In recent years, it has become more so, with the controversy surrounding the Government's Clipper chip, the FBI wiretap legislation, export controls on cryptographic software, and the balance of power between a government and its people. Historically, cryptography has been used mainly by governments for diplomatic and military traffic. But with the coming of the information age, ubiquitous personal computers, modems, and fax machines, this is changing. With an emerging global economy depending more and more on digital communication, ordinary people and companies need cryptography to protect their everyday communications. Law enforcement and intelligence agencies want access to all of our communications, to catch people who break the law, and detect threats to National Security. Civil libertarians want to keep the Government out of our private communications, to protect our privacy and maintain a healthy democracy.

PGP is free software. Anyone may download it on the Internet, or from many Bulletin Board Systems. It has stirred up some controversy, because it has become a worldwide de facto standard for Email encryption, despite US export restrictions. Initially published in the US, this package has spread by the diffusion that is common to free software packages, with its "forbidden" flavor giving it an extra popularity kick. Oddly enough, the US Government may have inadvertently contributed to PGP's spread, by making it more popular because of my case. I am under criminal investigation because of PGP's spread overseas, which the Government holds is in violation of US export restrictions. My case has captured a lot of press attention, in part because journalists realize that if an American can be imprisoned for electronically publishing something in the USA, then journalists may themselves be at risk in tomorrow's world of electronic newspapers on the information highway.

Another reason why the press is so interested in my case is the Government's attempts to suppress public access to strong cryptography. The Clinton administration is trying to get the phone companies to put a special encryption device into every telephone. They expect it to take many years to accomplish this. When this "Clipper chip", as it's called, is manufactured by the Government, they place a unique encryption key in each chip, and keep a copy of the keys in a vast government database, for wiretap purposes. Your telephone will someday have Big Brother inside. The Government hopes that the American public will accept this government-controlled cryptography, and is trying to discourage other forms of cryptography that they do not control. One way that they discourage it is by the use of export restrictions on cryptographic software. This draws PGP into the press spotlight.

The US State Department has a list of items that may not be exported without a license. The Munitions List. Mostly weapons, but included in that is encryption software. Encryption software may not be exported without a license, and that license is hard to come by if the software uses advanced encryption techniques that the Government can't easily break. Software like PGP.

The State Department allows items on the munitions list to be exported if they grant a Commodities Jurisdiction (CJ) for it, allowing it to be handled under the jurisdiction of the Commerce Department instead of the State Department. A CJ allows the item to be legally exported from the US. It would be politically difficult for the Government to prohibit the export of a book that anyone may find in a public library or a bookstore. The State Department has already granted a CJ for another book containing cryptographic source code, Bruce Schneier's "Applied Cryptography". So, we're putting the PGP source code in a book, which may be scanned in with OCR (optical character recognition) software. And we are applying for a CJ. It will be interesting to see where this process leads.

PGP uses the best encryption algorithms available in the open academic literature -- the algorithms that have withstood the most peer review, and are rooted in the best design principles of modern cryptography. PGP version 2.6.2 uses RSA for key management and digital signatures, the IDEA cipher for bulk data encryption, and MD5 for a secure one-way hash for digital signatures. And it compresses the data before encrypting it, using the ZIP algorithm.

PGP was developed under difficult conditions, with no funding, in a race against time in 1991 to get it out before it became illegal to publish software of this type. Senate Bill 266, the 1991 anti- crime bill, had a measure buried in it that foretold the shape of things to come. It was a "sense of Congress" resolution, that said that communications system implementors should provide a way for the Government to obtain the plaintext contents of traffic. The resolution was defeated shortly after PGP was initially published, due to public outcry from civil liberties groups. PGP was published as freeware, in an effort to preempt the possibility that the Government could suppress this technology later.

There are other books in the computer science section of your bookstore that have source code examples of various algorithms. These books teach good programming style by providing good code examples. This book is mostly not like that. Although it does have some instructive examples of how to implement certain algorithms, one might observe that this body of source code is not the best example of clean modularity. The code is the way it is. The trying conditions of PGP's development led to expedient approaches to implementation, sometimes at a cost of elegance. As other software engineers all over the world contributed their own development efforts to PGP after version 1.0 was published in 1991, they inherited a misshapen crystal nucleus to accrete more code onto, and modified and remodified the code under pressures of expedience. Very little of my original code remains in PGP, mostly comprised of the multiprecision integer library, which I first wrote in late 1986. It is unfortunate that cleaner examples of code from my 20-year career as a software engineer could not have served as a basis for what would eventually become such a widely distributed body of source code that so many other programmers would look at.

Despite the inelegance of the code, the quality of the cryptography and key management is very good. Good enough to make a mark upon the world in such a way that no other cryptography software ever has, upsetting some elements of the US Government, as well as a number of foreign governments.

PGP would not be where it is today without the volunteer efforts of many talented programmers from all over the world. Peter Gutmann and Branko Lankester contributed much effort to PGP development.* Jean-loup Gailly contributed his ZIP compression routines. Many others contributed code, and they are named in the comments of the source code. Other people contributed their non- programming efforts to the PGP project, including my own legal defense team, including Phil Dubois, Eben Moglen, Ken Bass, and Curt Karnow. Other lawyers that provided legal advice include Tom Nolan and Chuck Marson.

I'm writing some of this preface in the air between Bucharest and Budapest, on a speaking tour about privacy in the information age. In Bucharest, I saw the terrible legacy of a system designed by men who craved certainty, not trusting the people with individual freedom. Those men would have loved the Clipper chip. The people there now are glad to have their freedom, and they understand my concern about the power of Government. They already get it -- and they don't understand why we Americans don't.

A book comprised entirely of thousands of lines of source code looks pretty dull. But then so does a nondescript fragment of concrete -- unless it happens to be a piece of the Berlin Wall, which many people display on their mantels as a symbol of freedom opening up for millions of people. Perhaps in the long run, this book will help open up the US borders to the free flow of information.

Philip R. Zimmermann
Boulder, Colorado
November 1994