🚀 go-pugleaf

[RMS asked me to post this]

Hon. Tim Valentine
Chairman, subcommittee on Technology and Competitiveness
House Committee on Space, Science, and Technology
U.S. House of Representatives


Dear Mr. Valentine:

In a letter sent to your committee on September 20, 1991, James Bidzos,
President of RSA Data Security Inc., states that the RSA encryption
algorithm should be used in the NIST's proposed DSS (Digital Signature
Standard) because it is the de-facto standard for public key
encryption.

We must inform you that this is not true.  The RSA algorithm has not
become a standard for public key encryption in the United States
because RSA Inc. has used the various patents covering the algorithm
to prevent its large-scale use.

[Note added subsequently: RSA responds to this by citing a list of
companies that have signed agreements to resell RSA software.  But
this is not an indication of widespread use; many dealers is not the
same thing as many users.]

To make a protocol for communication standard and generally accepted
requires surmounting a large obstacle: the fact that there is (at
first) no one else to talk to who speaks the same language.  Consider,
for example, the case of UHF television; at first, no one would buy a
UHF receiver because there was no broadcasting, and no one would set
up broadcasting because no one was watching.  This obstacle was
overcome only by explicitly requiring all new TV sets to support UHF
channels.

Widespread adoption of public key encryption depends on the universal
availability of software to use it.  The most effective way to do this
is to write a public domain program which everyone can use.  Over the
last decade, several such programs have been written.  RSA Inc. has
squashed each one with accusations of patent infringement.

Bidzos raises concerns about the undue haste in the public comment on
DSS, and the size of the key it permits.  These concerns may be
valid.  Surely they can be addressed by modifying the DSS or its
adoption process, and are not an argument for switching to RSA.  The
suggestion that NIST might use patents to prevent improvement of DSS
is also alarming, and NIST [should] take positive steps to allay this
concern, but such a patent problem is not as bad as the one we know we
will have using RSA.

The computer industry deserves a public key standard which software
developers are free to implement.  We hope you will make sure the
eventual standard permits this, by supporting NIST in its rejection of
the RSA algorithm as long as patents continue to cover it.


			Sincerely,



			Richard Stallman
			  (MacArthur Fellow and ACM award winner)
			545 Technology Square, room 430
			Cambridge, MA  02139
			(617) 253-8830

			and

			Patrick H. Winston
			  (Director of MIT Artificial Intelligence Lab)
			Marvin Minsky
			  (Professor, MIT Media Lab; founder of
			   Artificial Intelligence Lab) 
			Berthold K.P. Horn
			  (Professor, MIT Artificial Intelligence Lab)
			David MacAllester
			  (Professor, MIT Artificial Intelligence Lab)
			Hal Abelson
			  (Professor, MIT Artificial Intelligence Lab)
			Gerald J. Sussman
			  (Professor, MIT Artificial Intelligence Lab)
			Christopher Hanson
			  (Research Staff, MIT Math and Computation Project)
			Leonard H. Tower, Jr.
			  (Boston University computer services)
			Jerry Cohen
			  (Patent Attorney, Perkins, Smith and Cohen)

cc:	Members, Subcommittee on Technology and Competitiveness
	Hon. Jack Brooks, Chairman, House Committee on the Judiciary
	Hon. Robert Mosbacher, U.S. Secretary of Commerce
	Dr. Willis H. Ware, RAND Corporation