________________________________________ ) PHILIP R. KARN, Jr. ) ) Plaintiff, ) ) Civ. A. No. 95-1812(CRR) v. ) ) (Judge Charles R. Richey) U.S. DEPARTMENT OF STATE, and ) THOMAS E. MCNAMARA, ) ) Defendants. ) ) ________________________________________)
PHILIP R. KARN, JR., declares as follows:
1. I make this Supplemental Declaration in response to the Defendant's Reply to our Opposition to their Motion for Dismissal. I have read the entire Reply, as well as all other filings in this matter. The statements in this declaration are based on my personal knowledge and experience.
2. The defendants state that export of the diskette "would present a far greater threat to national security than does the appearance of some source code on the Internet." (Defs. Reply at 3) That statement is simply wrong. The Internet is one of the primary means of facilitating exchange of information among cryptographers. Indeed it is far easier and faster to obtain a source code posted on a publicly-accessible Internet site than to obtain it by physical means. Physical shipment of the diskette is a far less efficient means of "spreading crypto" than the Internet. For example, I just used the searching features of the Internet World Wide Web to locate the entire contents of the Applied Cryptography Disk at issue in this case on a public Internet site in Italy. This includes the exact same "Triple DES" source code, with all errors corrected, that is the subject of the demonstrations in the Crowell and Karn declarations. This took about 10 minutes. I then entered a command to "download" the Triple DES code to my personal computer. The transfer from the computer in Italy direct to my computer in my San Diego home took 1.7 seconds. This is somewhat less time than it takes me to physically move to my computer, insert a floppy disk, and return to the keyboard. Anyone connected to the Internet anywhere in the world can obtain this code just as quickly and easily.
3. The defendants argue that the demonstration program described in the Declarations submitted by myself and Deputy Director Crowell illustrate that using the source code on the diskette to create a functioning program is "not a significant task." (Def. Reply at 5, n. 2) That statement ignores the fact that the demonstration programs we used were not fully-developed encryption programs that would actually be used in a commercial, or other non-experimental, environment. Practical programs require the design of sophisticated key-generation, key-management, user interface and input/output routines. These additional routines are far more substantial, in terms of programming time, than the encryption algorithm alone. For example, the printed source code for PGP is 895 pages long. The encryption code for that program, IDEA, takes only 16 pages. See PGP - Source Code and Internals 1-895 and 367-382 (MIT Press 1995)
4. The source codes on the diskette at issue in this case contain a number of "comments" which are marked off by "/*" and "*/" characters in Exhibits 3 and 4 to our Opposition. Source code comments consist of guidance to the human reader that are completely ignored by a computer compiling program. As such comments function solely to provide additional information to the reader to aid in understanding the source code and how it can be "called" by other program components needed to make a complete functioning program.
5. The defendants state that the "principal function" of the diskette is "to serve as a physical device that can be used to encrypt information." (Def. Reply at 7) That statement is again simply wrong. The diskette at issue here simply cannot be used to encrypt anything. It isn't a "device" that "functions" to do anything. It simply stores information in a form that can be "read" by a computer more readily than a printed page.
6. The defendants inaccurately state that "[e]xport of the diskette would provide foreign recipients with . . . a tool that would help shield their communications from national security surveillance by the United States." (Def. Reply at 9) The defendants' assertion ignores the indisputable fact that fully functioning encryption programs, as well as the entire contents of this disk, are already widely available in foreign countries. This availability makes the restrictions on this diskette immaterial with regard to the ability of foreigners to shield communications from surveillance.
7. The defendants' sophistic statement that "[m]any foreign users would be more likely to trust and use encryption software coming directly from reputable sources" than to use the programs available on the Internet, (Defs. Reply at 11), is identical to an argument advanced by Clinton Brooks, a high-ranking National Security Agency official, at a 1995 Washington, D.C. conference sponsored by the Cato Institute that he and I attended. In response, I explained that cryptographic "signature" techniques traceable to a specific originating individual are routinely used to certify the integrity of software such as the PGP program that is available over the Internet. Furthermore, fully documented security programs with published source code, like PGP, are often considered more "trustworthy" than commercial programs. The best way to assure potential users that an encryption program does not have a "hidden back door or "virus" is to make the source code available for public inspection and to allow users to recompile it for themselves. Not every user need be competent in reading source code to benefit from this, as anyone who does discover a flaw can spread the word almost instantly on the Internet. In contrast, the source code for nearly all commercial programs is a trade secret, not open to public inspection. Anyone who buys a commercial encryption program must necessarily trust the integrity and competence of its manufacturer. Indeed it is my understanding that it is US Government policy to obtain and inspect the source code of every program it buys for use in secure environments.
I swear under penalty of perjury that the foregoing is true and complete to the best of my knowledge and belief.
[signed] Philip R. Karn, Jr. December 21, 1995