From: Mike Ingle <MIKEINGLE@delphi.com>
Rich Lethin <lethin@toast.ai.mit.edu> wrote:
I was chatting today with someone moderately well-informed about the clipper controversy (unlike me). He pointed out the following work to me by Prof. Silvio Micali at LCS on a technical scheme which can serve as a compromise between the needs of society for legitimate wiretaps and the need of individuals for strong privacy. Basically, it seems to be a protocol for extending a public key algorithm into a k-escrow system. This apparently differs from the Clipper chip in that algorithmic details are well publicized.
The work was in Crypto '92 apparently, and an MIT lab for CS tech report numbered TR-579b.
Even worse, Micali is claiming that his patent on fair cryptosystems (#5,276,737) covers Clipper as well. In the Wall Street Journal (May 31, 1994, p. B6):
Mr Micali, whose patent was issued in January, says his patent covers the concept of breaking an encryption key into multiple parts that are guaranteed to work, and are held by escrow agents.
It seems to me that Clipper does not guarantee that the multiple parts will work in anywhere near the same way as his scheme does (see my book for details); Clipper is simply a secret splitting scheme. On the other hand, Micali filed his patent application in Apr 92, a full year before Clipper became public.
Bruce (Schneier)
One thing of note from the book on Crypto 92, is that the conference occurred in August. The paper on fair crypto systems contains references to President Clinton and Clipper, having been written or revised between May and August of 93. Hardly seems fair for something supposedly presented (and reviewed in 92). Is it just me or does this seem questionable? (Possibly being an attempt to show prior art, or perhaps being simply revised by someone close to the publication process.)