- cypherpunks-legacy - lists.cpunks.org

Re: How do I choose constants suitable for Diffe-Hellman?
by Matthew Ghio 17 Dec '03

17 Dec '03

0x7CF5048D wrote: > How do I choose constants suitable for Diffe-Hellman? > According to _Applied Cryptography_ n should be prime, > also (n-1)/2 should also be prime. g should be a primitive > root of unity mod n. n should be 512 or 1024 bits long. > Are there any other requirements? > > How can I choose such numbers? Are such numbers published > anywhere? Yes, Phil Karn posted a list of such numbers to the list last May, and the program used to generate them. Since some people have expressed their distaste for large files re-posted/forwarded to the list, I won't send it, but you can get it from ftp cs.cmu.edu: /afs/andrew.cmu.edu/usr12/mg5n/public/Karn.DH.generator

2 1

How to find a primitive root of unity, for Diffe-Hellman?
by 7CF5048D＠nowhere 17 Dec '03

17 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- > >How do I choose constants suitable for Diffe-Hellman? >According to _Applied Cryptography_ n should be prime, >also (n-1)/2 should also be prime. g should be a primitive >root of unity mod n. n should be 512 or 1024 bits long. >Are there any other requirements? > >How can I choose such numbers? Are such numbers published >anywhere? > Ok let me take a stab at finding g assuming n has been choosen to meet the above requirements. (I hope my math is still good.) Let Zn be the field defined by the prime n. Let G be the multiplicitive group defined in Zn. So |G| = n-1. Now n is large so 1 is not equal to -1 in Zn. Let N be { 1, -1} in G. It is a subgroup. Zn is abielian so it is Normal. We can consider the canoical map: G ---> G/N The order of G/N will be (n - 1)/2 which we are assuming to be prime. G/N is a cyclic group with no non trivial subgroups. Every element not = 1 is a generator. Pulling back to G we find that if g is not a root of unity, then the other member of its co-set = -g is! So take any g and raise to (n-1)/2 power. The result will be equal to 1 or -1. g raised to any lower power will not be equal to 1 or -1. Since (n-1)/2 is a large prime, it is odd. So if g to the (n-1)/2 is = to 1, then - -g to the (n-1)/2 = -1. So we can find a g which raised to the order (n-1)/2 power is = to -1. So g to the (n-1) power is =1 and g is a primitive root of unity. Have I made any errors? Did I get it right? -----BEGIN PGP SIGNATURE----- Version: 2.6 iQCVAgUBLmt8dw2Gnhl89QSNAQHzmAP9GUGAmFcbgMyWxKtrzEvQYJS33FXGoGmr w4rXblv14lkwJX32hpoRKmicm3bdND2OPGgmM4EefGYggj+iCI+NU+l6II+MxhjY C4Rk3Xjn59H81FhNdfcNqOU9AirjwMBSqKzYtNCfbedB6HuQDCTeLSU5pjI5PSEQ wvFP7F3i5rY= =0r8J -----END PGP SIGNATURE-----

2 1

Steganography by CAD
by John Young 17 Dec '03

17 Dec '03

Responding to msg by a.brown(a)nexor.co.uk (Andrew Brown) on Wed, 3 Aug 1:17 PM Andy: Features for concealment of plain or cyphertext by CAD programs: 1. Plain or cyphertext can be imported by each letter, by sentence, by block of text; then manipulated by the CAD program as one or multiple concealed CAD entities. 2. Text manipulation includes: Reduce to any scale. Place on hidden layers. Hide by other entities. Distort to a variety of shapes. And otherwise disfigure the text into non-recognizable forms. 3. In addition, there are a host of CAD-features for further, multiple, transformations of the text, including algorithmic permutations. 4. Subsequent manipulation of the CAD document may include exporting to other proprietary and generic CAD formats. 5. CAD document may then be transmitted plain (in the selected CAD format) or encrypted for transmission. 6. The recipient uses a CAD program (and decryption program if needed) to restore the concealed text to its original encrypted form. It is exported to the original encryption program for decryption. 7. Caveat 1: While this manipulation will be invisible or camouflaged in the CAD document the underlying code of the program will provide clues. However, it should be possible to encrypt these clues. 8. Caveat 2: Sender and recipient will need to exchange protocols, by encrypted transmission probably, for restoring the concealed text. 9. Caveat 3: This is not elegant and requires little or no crypto competency. But it can be easily done by a CAD operator to help conceal sophisticated crypto. Comments? John, an architect of the building codes domain.

1 0

REMAIL: scripts
by Karl Lui Barrus 17 Dec '03

17 Dec '03

Recently, somebody mentioned writing scripts to assist in using the remailers (I deleted the message and can't remember who it was). I used to keep a remailer list, and generated simple unix Bourne shell scripts and DOS batch files to do exactly this; however, I'm cleaning up my account and now defer to superior products: (Matt Ghio's and Raph Levien's remailer lists are better, and Hal Finney's chain program is more functional than the scripts). But, if they may be of some use to somebody, I just made available the various PERL scripts, makefiles, etc. I once used to generate the above mentioned scripts. It is available via anonymous ftp at chaos.bsu.edu in pub/cypherpunks/incoming as klb.scripts.tar.gz. The remailer data file included is fairly old and will need updating if somebody wants to toy with the scripts, etc. -- Karl L. Barrus: klbarrus(a)owlnet.rice.edu 2.3: 5AD633; D1 59 9D 48 72 E9 19 D5 3D F3 93 7E 81 B5 CC 32 2.6: 088C8F21; 97 73 9E 8B 98 3E DD B5 E8 97 64 7E 20 95 60 D9 "One man's mnemonic is another man's cryptography" - K. Cooper

1 0

Virtual Liability and Lethal Arguments
by blancw＠pylon.com 17 Dec '03

17 Dec '03

Responding to msg by Jim Hart: >What about liability of non designers for failing to >make changes? What about liability of people who use >fallacious moralistic arguments and use coercion to >stop changes? ................................................................. .......... You don't really have a problem with that, do you? :>) There are no expectations placed on non-designers to come up with anything of any practical use to society, therefore they have no responsibility or liability, as no one takes them very seriously. People who put up blocks to progress and improvement (not merely 'changes' per se) are suffering from a sense of loss of control over their existence, and should be sent to a hospital to recuperate from their psychic wounds. (Actually, I think that this wouldn't become such a big problem if, for one thing, everyone had a clear understanding & appreciation of the concept of "property", what that constitutes, and what that means in terms of the liberty to affect things in the world around us.) Blanc

1 0

Re: (fwd) "Will You Be a Terrorist?"
by frissell＠panix.com 17 Dec '03

17 Dec '03

At 01:53 PM 9/17/94 -0700, Timothy C. May wrote: >(In this interpretation, the remailers that many of us have talked >about, designed, written software, run on our systems, etc., could be >considered to be part of a conspiracy, should Bidzos, Mykotronx, >ClariNet, the Feds, etc., choose to focus on remailers as "a >continuing criminal enterprise.") But we designed and deployed the remailers not as part of a criminal enterprise but in an attempt to *generally* protect privacy. We may also have wanted to weaken some of the control measures inherent in the TCP/IP protocols. The TCP/IP protocols are *not* a government. It is legal to weaken them. Additionally, federal courts have explicitly held that anonymous communications (and anonymous associations) are protected by the First Amendment. No US Attorney is going to be interested in the sort of messy political case that would be involved in remailer prosecutions. What we are doing in any case is more like sedition than a normal criminal conspiracy. The Feds have only brought two sedition cases in this century (WWII isolationists and white supremicists in the 1980s) and didn't do too well. Sedition cases are real hard to win because in order to prove intent to overthrow the government you have to prove some realistic capabilities. Like Professor Crampton said when he saw my "Fuck the State" button in 1971 -- "That would be a rather large job, wouldn't it?" DCF

1 0

Patient Privacy at Risk (FWD)
by David Banisar 17 Dec '03

17 Dec '03

Fwd from the Coalition for Patient Rights (CPRMA(a)aol.com) Alert Patient Privacy at Risk Contact the ACLU The ACLU appears on a list of endorsers of the Wofford/Dodd amendment which amends one of the Senate health care reform bills. Major portions of W/D would have a severely adverse impact on the confidentiality of medical records. Although W/D has been rendered partly obsolete as newer health care reform bills are advanced under new names and new coalitions, many of its principal features remain intact in the new bills. It has become a reference point. It is for this reason that the signature of the ACLU on a list of endorsers of W/D (on a document entitled "Wofford/Dodd Fact Sheet") is so troubling and so damaging. The amendment creates federal standards for the disclosure of personally identifiable health care information and establishes a framework for a national health care data network. On the surface, the goals seem good. Who wouldn't be for establishing strict federal guidelines to ensure privacy where none existed before? For that matter, why not support a data network that would allow a treating physician to have immediate access to all pertinent medical information? Clearly we have to look beyond the advertisement and into the details of the bills for the answers to these questions. For example, in Sec. 508(a) of Mitchell 3 (the bill offered by the Senate majority leader), the "health information network service" is made the agent of the provider. This means that once a third party bureaucratic agency receives the information electronically, it is deemed the same as the health care provider in making decisions about the release of the information. Sensitive medical information, including intimate psychological information, would be available electronically to an increasing number of people legally--not to mention the well-documented risks of illegal access. Among those with enhanced access would be law enforcement officials and government agencies. Even researchers could access personally identifiable health information, if an institutional review board holds that the project is "of sufficient importance to outweigh the intrusion into the privacy of the person who is the subject of the information." The patient has no right to refuse such disclosure even though it includes his or her name. There are many examples of person-identified medical information, including sensitive personal information, that have been shared with health care providers with the expectation of privacy that would now be legally accessible to many third parties. The argument is made that this kind of access already exists, so why not codify it. The logic is faulty. It is true we already have serious problems protecting the privacy of medical records in this country. Legally sanctioning medical access to an ever enlarging list of third parties is not the solution. It will only compound an already serious problem. A compelling argument has been made that the establishment of a national health care data network that requires all providers to disclose information about every patient contact would violate the Fourth Amendment's prohibition of "unreasonable searches and seizures" of the person. Many organizations have raised serious concerns about Wofford/Dodd, including the American Psychiatric Association, the American Psychoanalytic Association, Coalition for Patient Rights, National Organization of Women, and the AIDS Action Council. We hope that the ACLU joins us in support of genuine privacy legislation. We hope that there was an error when it appeared on a short list of supporters of Wofford/Dodd (June 10, 1994) Call your state chapter of the ACLU. It is listed as Civil Liberties Union of (your state) in the white pages. Let them know of your concern.If possible, also fax Laura Murphy Lee at the ACLU in Washington (202-546-0738) and let her know your concern regarding the position of the ACLU in supporting W/D. This alert is provided by the Coalition for Patient Rights, Massachusetts (617, 433-0114).

1 0

Re: Unicorn??
by Jeff Barber 17 Dec '03

17 Dec '03

Black Unicorn writes: [ spelling errors, usage errors, grammatical errors, ill-reasoned arguments... ] This is not the same Black Unicorn we knew last spring. Has someone hijacked his account? Regardless of whether you agreed with his points, Unicorn always "spoke" eloquently and carefully, and with a distinctive style that has not been present in the posts emanating from this account in the last week or two. Uni's reputation is being rapidly squandered. Could this be Detweiler's revenge? -- Jeff

2 1

Re: On the crime bill and remailers
by John Young 17 Dec '03

17 Dec '03

Responding to msg by tcmay(a)netcom.com (Timothy C. May) on Tue, 20 Sep 10:40 AM >I agree with Hal Finney's point that current remailers >are far from the "ideal mix" described by Chaum and >are basically not very good at what they are supposed >to do. [Snip] >[Hal says] >> Claims here that remailers are designed to support sedition or to >> prevent government surveillance are both wrong and harmful. This kind >> of material could show up at some future prosecution of a remailer >> operator. [Snip] [Tim says] >Maybe. This is why I brought up the "supporting >terrorists" language of the Crime Act. Hal and Tim point to the vulnerability of the person running the remailer to being threatened through the equipment owner (like John Perry's recent experience) or squeezed by government agents into playing along with surveillors. This isolation and elimination (or co-optation) of a target has worked again and again to destroy networks once they become serious threats to law and order. Hal's warning is wisely heeded until all the remailer beneficiaries are bear an equal share of the risks of being charged a "supporting terrorist". John

1 0

RE: American Survival Guide
by Brian D Williams 17 Dec '03

17 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- Stanton McCandlish writes: >Oct. issue of _American_Survival_ has an article called "Clipper >Inside". >Dunno anything else about it, might be good, might suck rocks. CLIPPER INSIDE "CLinton's Orwellian plan to bug every compute,telephone,and fax... * Does a fair job on clipper. * mentions Matt Blaze's paper but gets it wrong. * does a side blurb on the EFF. * does another side blurb on PGP. * mentions Digital Telephony but doesn't mention EFF/FBI collaboration. * essentially very Anti-clipper * picture on page 46 alone worth the price of admission hehe... Typical issue, even better is the article "Liberalism as a form of codependency." Even better was last month's article "Make mine a Shotgun!" Brian Williams Extropian Cypherpatriot "Cryptocosmology: Sufficently advanced communication is indistinguishable from noise." --Steve Witham "Have you ever had your phones tapped by the government? YOU WILL and the company that'll bring it to you.... AT&T" --James Speth -----BEGIN PGP SIGNATURE----- Version: 2.6.1 iQCVAwUBLn79rlmp8FjpRfTVAQFgEAP+JMxaoOGcn9/nbCDwFCuSHuEmYf0ljI8I /HCkLmOOTOooVi0mLgQrSfFUA7DvwV/N9IwqkY6Th4l2Q1jw8MmOThShzL/RpRND sl/VJTbbhV/z0rdbHB2yc/DU10bBvrpQZyI0fVS5g/uouKsYEXoj0TB3ytXwo/aI dedfaVwGQOc= =+8W5 -----END PGP SIGNATURE-----

1 0