- cypherpunks-legacy - lists.cpunks.org

Re: Trust
by rdew＠el.nec.com 17 Dec '03

17 Dec '03

James, That one is easy. NSA does not get all the recovery key info from stuff developed in Australia. On Fri, 13 Feb 1998 James Morris <jmorris(a)intercode.com.au> wrote: > > I'm just wondering if anyone else finds it particularly odd that while > Australia is prepared to commit SAS troops, medical and intelligence > personnel to a possible strike against Iraq, the US government still > doesn't seem to trust us enough to use US developed cryptography > products with adequate security. > > - James. > Bob De Witt, rdew(a)el.nec.com Opinions expressed are my own, and in no way relate to those of my employer, or anyone else. -rdew

1 0

Re: "This post is G-Rated"
by tcmay＠got.net 17 Dec '03

17 Dec '03

At 9:58 PM 1/25/96, Bill Frantz wrote: >goal. The big problem is how are the labels attached to the programs. I >agree with Tim that it is probably imposible for individual Usenet >postings. However is should be possible for TV programs, and whole >newsgroups. Coincidentally, after writing my post this morning, I saw a report on CNN or CNBC about how broadcasters, networks, and producers are strongly opposed to the mandatory V-chip. Not news, but some of there reasons were: not enough time. The movie business produces about 600 hours of product per year, and the ~weeks it takes for the MPAA to view the final product and rate it (and then negotiate with the director about cutting out scenes to make an R-rated film into a PG, etc.) is marginally tolerable. Such is not the case with television, with many shows in a final version only hours before broadcast (and obviously some with even less or no advance time available). And remember that the MPAA rating is most definitely NOT a "self-rating." I see self-ratings of Usenet and mailing list posts as possible, just nearly worthless. And the reall contoversial stuff, this kind of goddamned fucking shit, will not get screened out. After all, I voluntarily rated this thread "G," and look what got through! (And it's only the tit of the iceberg, so to speak.) A meaningful "parental filter" cannot be done on-the-fly with self-ratings. Some minor steps can be taken, but not all worth the expense and hassle of a mandatory system. --Tim Boycott espionage-enabled software! We got computers, we're tapping phone lines, we know that that ain't allowed. ---------:---------:---------:---------:---------:---------:---------:---- Timothy C. May | Crypto Anarchy: encryption, digital money, tcmay(a)got.net 408-728-0152 | anonymous networks, digital pseudonyms, zero W.A.S.T.E.: Corralitos, CA | knowledge, reputations, information markets, Higher Power: 2^756839 - 1 | black markets, collapse of governments. "National borders aren't even speed bumps on the information superhighway."

1 0

CAs and Digital Timestamping
by JMKELSEY＠delphi.com 17 Dec '03

17 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- [ To: sci.crypt, sci.crypt.research, cypherpunks ## Date: 01/25/96 10:08 am ## Subject: CAs and Digital Timestamping ] At the RSA conference last week, it occurred to me that there's a really neat application for digital timestamps (as done by Surety). Whenever a CA (Certification Authority) issues a public key certificate, it should also digitally timestamp it. This provides a relatively clean way to recover from top-level CA key compromises. First of all, let's talk about hash trees. A hash tree (I think the idea was originated by Merkle, but I could be wrong) is a binary tree. Each node in the tree eventually winds up with a hash value associated with it. The bottom nodes on the tree are hashes of individual messages. The other nodes are made up of the hash of all their children. This looks like this: (pardon the ASCII art) : H_0 = hash(H_1,H_2) : / \ : H_1 = hash(H_3, H_4) H_2 = hash(H_5,0) : / \ / \ : H_3 = hash(M_0) H_4 = hash(M_1) H_5= hash(M_2) 0 : : The neat thing about this is that, if I know H_0, then when someone wants to verify that M_0 appears in the hash, they only have to tell me H_4 and H_2, and the position of M_0 in the tree. I can then find H_3 = hash(M_0), H_1 = hash(H_3,H_4), and verify that the final H_0 = hash(H_1, H_2) gives the right output value. (Compare this with a hashing chain, and you can see that, for large trees, there's a big advantage here. The number of values needed to authenticate a given message in this tree is log_2(number of messages in the tree), while the number of values to verify a chain is is the whole number of messages in the chain. In the digital timestamping service offered by Surety, they use hash trees of this kind, because this allows efficient verification of a digital timestamp. The idea behind this is that a message is hashed into the hash tree, and that the final value of the hash tree each day or week is widely published, including in the New York Times. So long as it's not feasible to go back and change that value, and it's not possible to find collisions for the hash function, any hash value that appears in that tree must have been presented to the timestamping people at some point before that tree's final hash was calculated. Here's how we use this in having a CA sign certificates. CA Signs a Key: 1. At the beginning of the day, the CA generates a random value, R_0, and has it digitally timestamped. The resulting timestamp is used as one of the entries in today's hash tree. (If the CA is their own digital timestamp service, then they'll have to use the previous day's ending value, which is reasonable enough. They'll also have to work a lot harder to publish this value each day or week.) 2. For each certificate to be generated: a. The CA verifies the information in Certificate_u, then signs it with SK_{CA}. Certificate_u contains some indication of the time and date. b. The CA hashes Certificate_u into its daily hash tree. c. The CA sends Certificate_u to user u. 3. At the end of the day, the CA publishes its own final hash tree value, and gets this value digitally timestamped. (This amounts to having the CA act as its own "node." for the timestamping service.) Now, imagine that the CA's key has been compromised. (We have to assume that the CA's daily operations were OK--if not, there doesn't seem to be a clean way to recover cleanly.) The person who has the CA's key can issue false certificates. After a while, one of these false certificates are noticed. How do we recover? We use the hash trees and the digital timestamps which we've made in the past to verify each certificate presented for recertification. The digital timestamps allow us to immediately verify that this certificate was issued by this CA at this time. And we can be certain that this hash tree is correct because it's been digitally timestamped. What this does, essentially, is to allow us to quickly recover from key compromises. We still have problems if someone can take over the operations of our CA for a few days or weeks, though in that case, we probably know the likely dates. No compromise at the CA can put a different date's timestamp on a certificate. Now, I should note that I think Merkle talks about using hash trees to do public key certificates, in his thesis, and the idea may be patented. (It's been a couple of years since I looked at his thesis, so this is a little hazy.) However, we're not using them here to provide certificates--we're using them to authenticate the certificates in the event of a catastrophic key compromise. I don't think Merkle talked about this, but I could be mistaken. (At any rate, I've never seen any mention of it since then, though it's an obviously useful idea.) If the tree structure is patented, then we could still do this by using chains or some other structure. Does anyone know if this basic idea has been proposed before? I am pretty sure I haven't seen it, but it seems pretty obvious now that I think about it. We could even recover using this method from a break of the hash function supported (so long as the timestamps are done with multiple hash functions, and at least one isn't broken), or the public key algorithm used. --John Kelsey, jmkelsey(a)delphi.com / kelsey(a)counterpane.com PGP 2.6 fingerprint = 4FE2 F421 100F BB0A 03D1 FE06 A435 7E36 -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQCVAwUBMQfn70Hx57Ag8goBAQETUQQA3S7k8rrYDud5N3uVdUqSVHC3VH+Tpmuu wcRisf8PrYX/I9Q4vBTN7SS0H30Xl1bRTQyS1mQP+CuyBlESi1qn0OMKbgAYPndd 1qYOvrX/29fMbhrO7VQjAcSAHpCZOvoVfsq0fWwv4zzUIhJBZNFnMxdSm4eNrmEv U3/oo5CiM2o= =7TIl -----END PGP SIGNATURE-----

1 0

Are Cypherpunks Influential?
by Duncan Frissell 17 Dec '03

17 Dec '03

At 08:23 PM 1/24/96 +0700, Patiwat Panurach (akira rising) wrote: >Would you call cypherpunks (as a group and as a philosophy) to be >influential? Do you think governments listen to us much? Are they forced >to listen to us? Any stuff to support this? Please give me your comments. Much more influential than deserved. We've gotten a lot of press. I would say that we are an example of cultural entrepreneurship. An entrepreneur makes money by noticing that there is an unexploited difference between the price of a final product and the sum of the prices of the factors of its production (discounted to present value). We noticed that there is a difference between the social and technical capabilities of modern math and modern computing on the one hand and the public perception of those capabilities on the other. By simply pointing out those mistakes in public perception and doing some (as little as possible) coding to prove it, we've gained some publicity. An example of the gap between perception and reality that we can easily exploit, I would point to the regulation of broadcasting. There's a lot of talk with the telecoms bill about whether and how much broadcasting should be deregulated but it is obvious to us that it already has been. I was thinking of that while doing my mail and listening to the Leader of the Free World last night via RealAudio 2.0 and KLIF radio in Dallas via Audionet. Radio has been deregulated by technological change since RealAudio appeared last Summer. TV dereg will follow with the higher bandwidth. All cypherpunks have done is to point out simple facts like this. Pretty easy work if you can get it. DCF "My fellow Americans, we must free our nation from the tyranny of "Others Government" and turn instead to "Self Government". We must accomplish this great task not for ourselves, but for the *Children*." -- 10-second SOTU Address

1 0

John Doe
by jdoe-0007＠alpha.c2.org 17 Dec '03

17 Dec '03

The John Doe NYM/Remailer interface for Windows is a most excellent program that will allow even the most cypher-illiterate to make use of the technology that has been the exclusive domain of those in the "techno-know". It was a piece of cake to set up, obtain a NYM and select inbound and outbound remailers with options to chain as many as your paranoia deemed appropriate. $25.00 seemed a little steep but they will get my $$$. Nice Job. John Doe 0007

1 0

Problem building PGP on NT
by Max Domeika 17 Dec '03

17 Dec '03

Hi, First question: Is this the appropriate place to go with this kind of question. If not, I apologize. Anyways here goes: I'm trying to build PGP, the MIT version under windows NT using Microsoft C++ version 4.0. The program builds after a few modifications to the makefile but cannot generate a RSA key. Everytime, I type pgp -kg and answer the required questions, pgp attempts to generate some primes and apparently fails. Anyone have a clue as to what the problem may be??? I know there is an executable of PGP for NT already made, I've used it. So I at least know it's possible to build on NT. Anyone else attempt a port to NT??? Thanks ahead of time, Max

1 0

Re: "Gentlemen do not read each other's mail"
by jim bell 17 Dec '03

17 Dec '03

At 04:08 PM 1/25/96 -0500, hallam(a)w3.org wrote: > >>Why is he our patron saint? He was a government official coming out >>against invasion of privacy. Isn't that what we are all after, in the >>end? > >There is a considerable difference between running a government and being an >individual. It is not merely ethical for one government to read another's mail, >it is a duty. > >By not taking adequate steps to inform itself of the Japaneese intentions the US >suffered the loss of a substantial part of the US fleet at Pearl Harbour. Had >sufficient resources been avaliable the naval codes could have been cracked in >time. The closure of the Black chamber was a key reason why US espionage >efforts were inadequate at the start of WWII. While this may be based on the "classic" view of the start of the direct involvement in WWII, I agree with the opinion of an old college professor that the US KNEW that the Japanese were going to attack, SOMEWHERE and SOMEWHEN (but not exactly), and in fact WANTED the attack to occur to justify getting into a war that we "should" have entered. And in hindsight, I do not necessarily disagree with such a goal, within the limited context of the circumstances at the time. An extention of this interpretation is that much of the fleet was kept at Pearl to "lure" the Japanese into doing an attack that could be used to rally the public. (they needed to have enough "bait" to justify an attack.) Obviously, if that was the intention, then the big surprise was how EFFECTIVE the attack was going to be: Far from just "rallying the public" it smashed our defenses. The reason I mention this interpretation is that it entirely turns around your argument: Our criticism should not be how little we knew of Japanese intentions, but how we incompetently delayed entering a war that "needed to be fought." Naturally, however, considering the results of the attack, it would have been totally unthinkable to reveal to the public that the bigshots had actually DESIRED the attack; a far less incriminating version of the story is that some other people were merely negligent. Now, I was born in 1958 and thus can't claim personal knowledge of the time, but it's truly amazing how UNPERCEPTIVE the public must have been in the late 40's and early '50s about "intelligence" realities. Let me give you a specific example: The classic movie, "The Man who Never Was," relates the (true) story of a counter-intelligence mission done by the British to (I think) mislead the Germans into believing that the attack on Sicily would be substantially LATER than it actually was. The British took a man who had died of some natural disease (with the permission of his family, of course), dressed him up as if he were a courier and dumped his body (carried by submarine) off the coast of (then Fascist) Spain. With the body were (sealed) phony documents that described the FALSE date. (The idea was, the Germans would think that he was on an airplane that had crashed into the ocean...) ( He was given a false name, false address, and basically a false identity to complete the ruse.) By design, he was dumped at a point where ocean currents washed him ashore, where he was identified by papers. Naturally, the British were notified by Spain, and the British played along and INSISTED that none of the documents with the body be unsealed and given to the Germans. Naturally, however, the Spanish cooperated with the Germans, and allowed them to (secretly) unseal the documents undetectably. However, the documents (still apparently sealed and unopened) were returned to the British so that (to the Germans) the British wouldn't be aware that their secrets had been compromised. Follow me so far? Well, in the movie (whose accuracy I don't know) the Germans didn't totally believe that the courier was "real", so they sent an operative (probably by parachute airdrop, or whatever) to check out the particulars of the story they "learned" from the fake background. (Naturally, the British knew about this) At that point, the British had to plant operatives, and support the phony story on the spot, "verifying" the information. At that point, the German agent was able to transmit news that "the information is real!" to the Germans. Now, here was was (to me!) the "funny" part of the movie: The British let the German agent leave Britain. This sounds logical, right? Because if they were to PICK HIM UP as a spy, that would have alerted the Germans that his identity and mission were already known, which would only have been true if the story given the Germans initially had been a FAKE! Yet, in the movie, it was necessary to "explain" to the (early 1950's) audience why they "let that German agent leave Britain"!! I was laughing, practically falling off the edge of my seat, as I was watching that scene! Clearly, ordinary people of that era weren't very perceptive about such things. Had they (Americans) been told that, "We had to get into WWII to save the free world, so we let the Japanese sink half our Pacific fleet!" the public WOULD NOT have understood. This is why I tend to believe that professor's interpretation: While the exact time and location of the attack was not known, IT REALLY DIDN'T MATTER because they WANTED it to happen. The "we couldn't decrypt their traffic in time" (even if it was really true) was merely a convenient cover story for the truth.

2 1

NOISE Test Ignore
by jdoe-agamemnon＠alpha.c2.org 17 Dec '03

17 Dec '03

Test Test Test

1 0

Re: "This post is G-Rated"
by tcmay＠got.net 17 Dec '03

17 Dec '03

I typically enjoy Lucky's jaundiced analyses, but I think he is overly paranoid in this one: At 5:28 AM 1/26/96, Lucky Green wrote: >What you will see is something like this: all lists/newsgroups will have to >be moderated with the name of the moderator clearly stated just as it is in >print magazines today. If something that is potentially dangerous to First, printed material does *not* have to have an author or publisher specified. The usual case cited is the Supreme Court decision on "anonymous handbills." I know of some zines that have no identifiable origin point. And anonymous pieces are published (even where the publisher/editor does not know the true author.) Second, Usenet is a global thing. Which national laws will apply? If the U.S. somehow is dumb enough to demand that each of the 12,000+ current Usenet groups have a moderator in a U.S. jurisdiction....well, there are too many issues here to properly deal with in a short note. Third, if the moderator is liable for "allowing" inappropriate material to get through, I predict a rather severe shortage of moderators. It's already a thankless job, so why any sane person take on the liability of doing it? Fourth, the Soviets couldn't control the underground "samizdat" press, and this was when the sanctions were stronger and the channels of communicaton more available. Today, in the West, there are at least two orders of magnitude more channels, and several orders of magnitude more sites, bandwidth, etc. Fifth, major constitutional challenges would be mounted. The First Amendment says that a posting to a public discussion does not first have to be cleared by a censor/moderator! (I'm not saying a group cannot be moderated, I'm saying that making it a crime to have an unmoderated, uncensored group will not pass constitutional muster.) --Tim Boycott espionage-enabled software! We got computers, we're tapping phone lines, we know that that ain't allowed. ---------:---------:---------:---------:---------:---------:---------:---- Timothy C. May | Crypto Anarchy: encryption, digital money, tcmay(a)got.net 408-728-0152 | anonymous networks, digital pseudonyms, zero W.A.S.T.E.: Corralitos, CA | knowledge, reputations, information markets, Higher Power: 2^756839 - 1 | black markets, collapse of governments. "National borders aren't even speed bumps on the information superhighway."

1 0

Re: "This post is G-Rated"
by Greg Broiles 17 Dec '03

17 Dec '03

At 02:06 AM 1/26/96 GMT, Jeff Williams wrote: >Maybe nobody would use the flag, but I don't see how it could hurt. If I had >kids, I would appreciate having the option of sorting out all the stuff that >is "NIFK" by the author. My concern about such a flag is that if it was implemented widely, it would be a small step from being optional to being mandatory. As things stand, were some legislature to adopt a "mandatory labelling" statute tomorrow, we'd end up with a complex and pointless mess of conflicting and incompatible attempts at compliance, rendering the labelling scheme effectively useless; and, likely, "industry" (e.g., the big service providers + AT&T & MCI & Sprint etc) would oppose the legislation on the grounds that compliance would take several years and complex design to bring about. So we'd get some sort of grace period to argue against it, prepare good test case(s), establish offshore mailing lists/servers, and so forth. But if we let the market do an efficient job of developing and deploying a labelling standard, over the course of the next few years, we're an afternoon's sorry debate and a few bought-and-paid-for committee meetings away from a market standard turning into a statutory duty. This makes me think that if development of a labelling standard is imminent, it's time to find ways to subvert it, misuse it, avoid it, and otherwise treat it like an enemy. It's not necessarily an enemy until it's coopted into service by a coercive force (be it big brother^h^h^h^h^h^h^hgovernment or big brother^h^h^h^h^h^h^hcorporation) but it's a potential threat. Are you sure that your relatively moderate "less NIFK material is a good thing" stance is a good reflection of others' positions? My impression is that many people are dissatisfied with anything less than a "zero tolerance" position re kids & porn or kids & drugs or kids & sexual choice or whatever. Your view of a partial-cooperation world isn't especially draconian, but do you really think that others will be as tolerant of other positions re labelling? -- "The anchored mind screwed into me by the psycho- | Greg Broiles lubricious thrust of heaven is the one that thinks | gbroiles(a)netbox.com every temptation, every desire, every inhibition." | -- Antonin Artaud |

1 0