D. J. Bernstein
Fighting patents

US patent 4200770, Hellman Diffie Merkle, public-key cryptography

Text of the patent

This patent expired in 1997.

Most patented "inventions" are obvious from the prior art. The DH system was a rare exception: it revolutionized cryptography. Does this mean the patent was beneficial? No! Look at the timeline:

The patent was granted in 1980. Robert B. Fougner from "Public Key Partners" claimed that "all known methods of practicing the art of Public Key" were covered by this and followup patents. The patent itself claims the entire concept of public-key encryption (see claim 1), along with the DH modular-exponentiation key-exchange system in particular (see claim 6).

Under United States case law, a document has been published if it "has been disseminated or otherwise made available to the extent that persons interested and ordinarily skilled in the subject matter or art, exercising reasonable diligence, can locate it". How does a September 1977 patent on public-key encryption avoid admitting that public-key encryption had already been introduced in the "Multiuser cryptographic techniques" paper distributed in 1975? Answer: The patent description acknowledges the "Multiuser cryptographic techniques" paper as prior art but falsely claims that the paper "presents neither proof that public key cryptosystems exist, nor a demonstration system".

How does a September 1977 patent on modular-exponentiation key exchange avoid admitting that modular-exponentiation key exchange had already been introduced in the "New directions in cryptography" paper distributed in May/June 1976? Answer: The Stanford lawyers allowed the patent office to believe that the paper was first available in November 1976. Unlike most countries, the U.S. allows patents to be filed up to a year after public disclosure.

My notes say that Mike Matyas received a copy of "New directions in cryptography" in August 1976. Several further examples of public distribution of the DH results have been documented in detail. None of this was disclosed to the patent office.

The 1985 court case MIT v. Fortia invalidated another patent because the patent holder had handed out six copies of a preprint at a conference fourteen months before applying for the patent. It should have been straightforward to kill the DH patent in court: non-disclosure of prior art is "inequitable conduct" invalidating the whole patent, and, more to the point, the modular-exponentiation key-exchange system that people wanted to use was already stated in the prior art.

The fact that there was direct prior art for the DH patent appears to have been first pointed out as part of the complaint in a 1994 court case Schlafly v. Public Key Partners. This was 17 years after the patent was filed, too late to be of any use. This part of the case was privately settled by the parties in March 1997, presumably with the patent holder promising not to enforce the patent against Schlafly for the remaining 6 months before the patent expired.

To summarize, all available evidence is that Merkle, Diffie, and Hellman were distributing their work for free, driven by the usual academic publication incentives, and that they filed this patent only as an afterthought.