December 2003 - cypherpunks-legacy

Re: Master Key: A Clipper Story
by George A. Gleason 17 Dec '03

17 Dec '03

This is odd,... I never got the original posting; if anyone can email it to me I'd be thrilled (gg(a)well.sf.ca.us) thanks- -gg

1 0

New York Times: Federal Inquiry on Software...
by gtoal＠an-teallach.com 17 Dec '03

17 Dec '03

In article <199309220117.AA11227(a)bongo.cc.utexas.edu> zeek(a)bongo.cc.utexas.edu writes: > ---------------------------- > The New York Times > Tuesday, September 21, 1993 > Business Day > ---------------------------- > > Federal Inquiry on Software Examines Privacy Programs > > By John Markoff John Markoff regularly writes good and informed articles that he has clearly researched (if that's the right word) by reading usenet and possibly lists like this one. Yet I've *never* seen the guy post. John, if you're reading this, drop me a line to say hello just so I have your email address, in case I ever want to mail you something. Graham -- Personal mail to gtoal(a)gtoal.com (I read it in the evenings) Business mail to gtoal(a)an-teallach.com (Be careful with the spelling!) Faxes to An Teallach Limited: +44 31 662 4678 Voice: +44 31 668 1550 x212

1 0

Re: a horrible conspiracy revealed
by Perry E. Metzger 17 Dec '03

17 Dec '03

Timothy Newsham says: > > > > > > "L. Detweiler" says: > > > I was thinking. What is the significance that PRZ was not actually > > > cited in any subpoena so far? This is very puzzling. It seems to > > > > I wonder what the significance of L. Detweiler's name is? Obviously > > its a Germanic name. Perhaps he's a Nazi... in fact, he might be the > [...] > > What was the point of this letter? Is this example output of a > new stegonagrphy program you just wrote? where can I get it? > Do you know mr. subliminal from saturday night live? Can I meet > him? Can I have a porsche ? If I told you what the point was, there would be no point in having written it. Perry

1 0

PKZ and NON-RSA
by David Colston 17 Dec '03

17 Dec '03

When I posted the offer for source code to a non-RSA public key yesterday, little did I know that I would be beseiged by requests. I will therefore send the paper on the matter to cypherpunks and ask that anyone wanting a copy of the source send me a floppy. Any size or density is okay. As a preamble to this paper, I'd like to note why and how it was developed. After the first release of PGP, Phil told me he was worried about using RSA. I offered to come up with another scheme and originally designed a version of Rabin. I asked Phil to post a copy of that paper on sci.crypt since I do not have internet access. There were few responces, none of which attacked the safety of the method, but mainly stated that it was a version of Rabin. Nothing was ever done with that proposal. A year latter, 2.0 PGP was due to be released and I came up with the current public key system. Again, Phil ignored it. Posts on sci.crypt were totally ignored by anyone with the ability to judge the math. Phil was then approached by the customs folks and I recieved a call from him in December of last year. By that time I was tired of waiting around for Phil and had coded QPK (Quick Public Key) in PDS 7.1. At any rate, I suggested that Phil foward a copy of the program to RSA and see if they would challenge that program in court. He declined. It became apparant that Phil wanted to stick with the original RSA concept and would refuse to do anything with another approach, although he did stick an bite in 2.1 to flag an encryption as being non-RSA. At any rate, I believe that RSA would not attempt to defend the original public key patent in court, since it smacks of over kill. They attempt in the RSA pattent to claim any encryption method involving a polynomial modulo another number. This is clearly prior art, since Ceaser used m+1 mod n as an encrytion method. The program I generated has been posted on Compuserve, Bix, and GEnie for over a year with no feed back from anyone, even though I am resonably sure that RSA is aware of its existance. They have not challenged it because there has been no publicity. I would welcome a court challenge on the whole idea of crypto patents. The customs thing is meerly a way for RSA to challenge Phil and scare off others while avoiding a test of their patents. The "paper" follows. FOR THE MATHEMATICALLY ORIENTED - HOW A QUICK PUBLIC KEY WORKS. Math notation: + plus - minus +- plus or minus * multiplication / division ^ exponent <> unequal = equals == congruent < less than > greater than INT truncated integer round SQR square root ( open expression ) close expression x^-1 modulo p the multiplicative inverse of x in the field of p x... y the range of values XOR exclusive or N a number equal to P times Q, where P and Q are prime Variables in capital letters are permanent and those in small letters are temporary. BACKGROUND Because the "secret key" function of many public keys methods are so slow, most cryptographers use these functions only to "boot strap" into a conventional key system, which is faster to send the actual message. Most of the conventional key systems use comparatively small numbers in relation to the size of the public N as a "random seed number". The holder of the secret key may actually have a larger amount of computer time to decipher the starting point of the conventional algorithm than to decipher the actual message. It would seem to be a good idea, if a public key function could take advantage of the actual message size required to speed up the public key process. The range of message sizes is described below, but generally speaking we a discussing messages less than SQR(Q) in actual size. Imagine a series of related equations modulo a prime, P. These equations have the formula ((e * e + e)/2 * L + C) modulo P. The value, C, is a constant determined by the rule (L - 1) * (1 / (2 * L)) modulo P == C. For L = 1, C = 0. Therefore, if L is known, the expression ((e * e + e)/2) modulo P may be determined, even if e is unknown. Each value of L has an area or series of areas, in which the value of e becomes discoverable, WITHOUT resorting to a modular square root. ie. Let r == ((e * e + e)/2 * L + C) modulo P. If e is in the correct range relative to L, then (r * L * 8 + (L - 2) * (L - 2)) will have an integer square root and the value, e, may be determined with ease. The range of values of e, for any value of L, which have this property and the location of those values vary greatly. The following illustrates a public key approach for L = 12, but other values of L may also be used. Perhaps I should also note that the particular L, which a secret key holder uses need not be public knowledge, but is not all that sensitive. ESTABLISHING A QUICK PUBLIC KEY (Q.P.K.) BASED ON L = 12 A person wishing to receive public messages, which he/she alone can decrypt calculates N = P * Q. Where P and Q are a randomly selected prime numbers, Q being the larger. A == (11 * 24^-1 ) modulo Q B == (2 * A) modulo Q D = Q - B If D > (Q - 1)/2 then set D = Q - D - 1. F = (Q - 1)/2 - D NOTE: We may chose to use the F for all of the following calculations instead of D. This applies to no value of L lower than 12 and F is not valid for most sequences higher than 12. F may be an even safer value to use, for reasons which are too long to discuss here. L is NOT super-sensitive information. Let Y1 ... Y2 be a range of numbers with in the limits: (D - k) and (D + k), where k = INT(SQR(2 * Q / 12)). Y1 may be randomly selected from any point in this range, but Y2 may not be larger than (D + k), and Y2 - Y1 is the maximum message size. A message range for N, public information, is then created by using Chinese remainder theorem to find the modular intersection Q == Y1 and P == x, x being a random number in the range x > 0, x < P. This intersection is called S. A check is made to verify the following: A' = (11 * 24^-1) modulo N B' == (2 * A') modulo N D' = N - B' If D' > (N - 1)/2 then Set D' = N - D' - 1 NOTE: P, Q, D, F, Y1, Y2, k, A, AND B are secret values. Q.P.K. ENCRYPTION A public key for short messages consists of S and N. To send a Message the sender calculates: e = (S + Message) ((e * e + e)/2) modulo N == Cipher Q.P.K. DECRYPTION t == Cipher modulo Q f == (t * 12 + A) modulo Q g = SQR(f * 8 * 12 + 100) NOTE: If g is NOT an integer value, the message is rejected as invalid. If g modulo (2 * L) <> (L - 2) then Q is repetitively added to g until g modulo (2 * L) == (L - 2). z = (g - 10)/24 e == ((B - 1) + z) modulo Q If e > (Q - 1)/2 then set e = (Q - 1) - e. Message = e - Y1 For other values of L: A == ((L - 1) * 2^-1* L^-1)) modulo Q k = INT(SQR(2 * Q / L)) NOTE: If L = 1 then D = 0 and the message range is 1... k. If L = 2 then the message range is D... (Q - 1)/2 and these values modulo Q are already perfect squares < Q. f == (t * L + A) modulo Q g = SQR(f * 8 * L + (L - 2) ^ 2) z = (g - (L - 2)/(L * 2)) If anyone wants the source code for this drop a disk to: Colston & Associates 5111 Rogers Ave. Suite 507 Fort Smith, AR 72903

1 0

a horrible conspiracy revealed!
by gtoal＠an-teallach.com 17 Dec '03

17 Dec '03

In article <9309220341.AA23923(a)snark.lehman.com> pmetzger(a)lehman.com writes: > I wonder what the significance of L. Detweiler's name is? Obviously > its a Germanic name. Perhaps he's a Nazi... in fact, he might be the I sincerely *hope* this article was a forgery, otherwise Perry you're digging yourself a big credibility hole. Drop it, please. G -- Personal mail to gtoal(a)gtoal.com (I read it in the evenings) Business mail to gtoal(a)an-teallach.com (Be careful with the spelling!) Faxes to An Teallach Limited: +44 31 662 4678 Voice: +44 31 668 1550 x212

2 1

Former S.O.F. Editor Supports PGP in Newspaper
by James Still 17 Dec '03

17 Dec '03

Local boy makes good at home. The Federal Grand Jury's PGP/RSA test case, as reported in the NYT has hit full circle here in Boulder. Boulder's own Paul Danish (former editor and writer for Soldier of Fortune magazine, based in Boulder) has come out in support of PGP/privacy issues with an excellent op-ed piece in today's Colorado Daily. The Colorado Daily is read by over 50,000 students, fac/staff and residents of Boulder and the Univ. of Colorado. Danish's piece does an amazing job of backgrounding the issues, presenting them in layman's terms, and he even goes into the Clipper chip proposal. Could it be that the NSA/U.S. Customs investigation is causing a huge backlash of public support in favor of PGP?

1 0

Re: Why RSA?
by wcs＠anchor.ho.att.com 17 Dec '03

17 Dec '03

Derek writes: > First, the ViaCrypt version: I realize that it is legal. > It's also very expensive when compared to the price of If you buy it now, it's $100; even $199 is within the realm of what corporate customers consider non-annoying, especially since volume discounts kick in very rapidly. (Q5 = $120 each, Q20 = $83 ea, Q50+ = negotiable.) $100 is within my pain threshhold, and that of most serious non-student cypherpunks, though I'd be much happier at $49.95, and I'd guess they'd get more up-front cypherpunk revenue. I'm not sure the legalities of buying ViaCrypt PGP and then using the latest off-the-net version with trustable source and fewer bugs, but I'd feel better about that than just using PGP without a license. I don't know if ViaCrypt includes any distinguishing features that let you tell a ViaCrypt PGP message from a Real PGP message - they could do subtle stuff like make the session keys all include some checksum (e.g. be a multiple of N mod M), or crude stuff like put "Version 2.3ViaCrypt" in the headers, which would let them detect non-ViaCrypt PGP users. I assume not, but I haven't seen it. > * is there a reference I can read that covers the scope of > public key crypto patents? Basically, there are the patents themselves. Check your favorite FTP sites, including rsa.com, and ask your favorite mathematically-trained patent lawyer how realistic the claims are and how much broad the interesting ones are. > * in broad terms, what would I have to do to develop an > algorithm that works from a user's perspective like > p.k.c. (ie public/private keys, the central functional > point of all the wonderful schemes based on pkc) but > doesn't violate patents? To avoid patent problems, either get a license from PKP, or do any implementations outside the US (it's ok to use the math as math, you just can't apply it for encrypting stuff), or work for the U.S. government which gets use of at least RSA free. If you develop a new public-key algorithm that's any good, PKP may be willing to make a deal with you. # Bill Stewart wcs(a)anchor.ho.att.com +1-908-949-0705 Fax-4876 # AT&T Bell Labs, Room 4M-312, Crawfords Corner Rd, Holmdel, NJ 07733-3030 # # goin' where the climate suits my clothes ....

1 0

News You Can Use
by Duncan Frissell 17 Dec '03

17 Dec '03

AP 09/21 0844 Encryption Software Copyright, 1993. The Associated Press. All rights reserved. SAN JOSE, Calif. (AP) -- A federal grand jury is investigating exports of a controversial computer program in a case that could affect how software is distributed worldwide. U.S. Customs officials asked for an investigation into ViaCrypt of Phoenix and Austin Code Works of Austin, Texas, and the companies' plans for foreign distribution of software, including PGP, a program that turns data into an indecipherable code using encryption technology. William Keane, an assistant U.S. attorney, confirmed that an investigation is continuing, but declined to comment on the case. The PGP program has been distributed worldwide over computer networks by some computer enthusiasts who oppose the U.S. government's trade regulations on encryption. The National Security Agency, which monitors international communications, has supported strict encryption technology export regulations, arguing that it would be difficult to keep tabs on hostile governments and foreign terrorists. But opponents say the restrictions hurt sales and violate the First Amendment that protects the right to publish information about encryption. "I wrote PGP to make democracy healthier. I didn't do it to make money," said Philip Zimmermann, a computer consultant who developed PGP. "We believe everything we are doing is above board and well within the law," said ViaCrypt president Leonard Mikus. He said the company had no intentions of violating export regulations. WP (Washington Post) 09/18 Encryption Program Stirs Security Debate By John Mintz and John Schwartz Washington Post Staff Writers Computer industry officials and civil-liberties activists are launching new attacks on the Clinton administration's plan to make the so-called clipper computer chip the national standard for encrypting, or scrambling, data and voice communications. Under the clipper plan announced this year by the Clinton White House, police agencies that receive court authorization for a wiretap to intercept encrypted communications would then need the technological cooperation of two independent "escrow" agents to crack the code. Earlier this week administration officials told congressional staff members that the two escrow agents will be officials of two government agencies: the Commerce Department's National Institute of Standards and Technology (NIST), and a non-law enforcement section of the Treasury Department that has not been selected. Yesterday industry and civil-liberties groups criticized that selection because they said NIST and Treasury are not independent, but arms of the same federal government that could some day be called upon to listen in on their communications. Douglas Miller, government affairs representative of the Software Publishers Association, made up of U.S. software firms, said his group has "grave doubts" that foreign corporations will encrypt their communications with the clipper chip because "the U.S. government holds the key." A main reason the administration is promoting clipper is that the U.S. National Security Agency, the super-secret code-breaking agency, wants to discourage use of highly capable, non-clipper encryption programs that are becoming increasingly popular but that the NSA can't pierce. Industry officials for years have regarded NIST as a stalking horse for the NSA. Jerry Berman, director of the Washington office of the Electronic Frontier Foundation, which promotes public-interest causes in technology-policy areas, said NIST is "so close to the NSA that it can't give the public comfort that this is a true escrow system." John Podesta, assistant to the president and a key White House staff member on this issue, said such objections are "a phony issue." "We clearly are looking for procedures and escrow agents that would maintain privacy and confidentiality and security of the keys," Podesta said. "Cryptography lends itself to a certain degree of paranoia." Kate Martin, director of the Center for National Security Studies of the American Civil Liberties Union, mocked use of the term "escrow" in this case. An escrow agent is someone who is independent of two parties potentially in conflict, like a settlement attorney at a real estate closing, she said. "As long as the escrow agents are government agencies, it's misleading to call them that," she said. "The government doesn't have a fiduciary obligation to the people whose (communications) keys it holds," but only to the government. "The whole idea continues to be structurally flawed," said Bruce Heiman, attorney for the Business Software Alliance, a group of top U.S. software firms, such as Microsoft, Novell, Lotus and Apple. Companies and individuals who transmit secure information "will have serious doubts about the integrity of the system." Since the government currently prevents the export of many powerful U.S.-made encryption techniques, the administration's attempts to promote its clipper chip "will discourage use of encryption, period, or hand over the market for encryption to foreigners." When one listens to an encrypted conversation, it sounds like a crackle or buzz. Under the plan, every law-enforcement agency will have a special personal computer or "black box" to descramble that crackle, but the device will work only when they have been given a special key from the escrow agents. When police get a judge's permission to intercept an encrypted conversation or stream of computerized data, they would use the box to determine the special encryption identifier or label assigned to that particular encryption device. A detective would notify NIST and Treasury that he or she has permission to listen in on the party. NIST and Treasury would have a list of the secret encryption key numbers - extremely long lists of 0s and 1s - for every encryption device sold in the United States. NIST and Treasury would find the appropriate one on the list, and then they would send the needed key number to the police over telephone lines. The police would then insert that decoder number into the black box to tap the phone line in question. The ACLU's Martin said the government, given lists of secret encryption numbers, "has an enormously greater ability to eavesdrop than it's ever had." Government officials deny that. Duncan Frissell The $1 Trillion/year Health Security Act of 1993, the most expensive government program in the hisotry of mankind. --- WinQwk 2.0b#0

1 0

Why no publication?
by David Colston 17 Dec '03

17 Dec '03

Carl, I have not published for several reasons. The main one is that I have no academic creditials. My masters is in Rehab Counseling. I haven't had a math course since high school in 1964. I am mainly self taught in the area of math and crypto, but have had a lot of assistance from Charlie Merritt and Joel Benston. They were two of the 3 authors of Dedicate 32, the first public key crypto program for the PC, which was released in the early 80's. As I have noted, the paper was sent to sci.crypt by PKZ. /s

1 0

Re: NYT article - WOW
by Matthew B. Landry 17 Dec '03

17 Dec '03

"L. Detweiler" writes: >write! he almost appears to be alluding to *digital cash*! Obviously, >Mr. President forgot to clear his remarks with the NSA first! Saying He doesn't have to clear it with the spooks first. _They_ work for _him_! Maybe we'll all get really lucky and have someone in charge who realizes that. (Personally, I'm not holding my breath, but there's always some hope.) -- Matthew B. Landry ml7694a(a)american.edu (Finally!) mbl(a)ml7694a.leonard.american.edu

1 0