December 2003 - cypherpunks-legacy

The TRILATERAL COMMISSION -was: [REBUTTAL] Censorship on...
by George A. Stathis 17 Dec '03

17 Dec '03

At 12:34 ðì 14/11/1996 -0500, Mark M. wrote: (in response to me saying): >> Man you're nutts. There are very few and quite vast Media Companies >> in the world, and they're on the verge of becoming monopolies. >> >> Even your American President is in reality a puppet of the Trilateral >> Commission, who effectively also control CNN, the Washington Post, >> and many many many other things all over the world. >> >> And you are saying that control of the media by ownership is >> impossible? You're far out maaan! :-) >> >> Only in America such a naive opinion could actually be _believed_. >> (here I go again... aga! :-) ) > >Is this a troll? Nope; It's the truth. Expressed like a sarcastic joke as well, but it's _true_. And not intended to be a troll at all. As regards the "gullibility", it's plainly dangerous; And cultivated by the Media, apparently stronger in the U.S. than in many other places. But: How many times do I have to state explicitly that my grumbles about America are not "anti-American" (troll) propaganda? Nope; They're grumbles about _freedoms_, that if you Americans lose, the rest of the world will _also_ lose, (even quicker, I think)... Here is a REAL example: (written by E. Francis, on the "Trilateral Commission", and distributed to all members of his own -private- mailing list, of which I'm a part). >THE LAND OF SHADOWS > > Early this month is the last presidential >"election" before the year 2000. As of this writing and >long before, the election was decided in favor of Bill >Clinton. It works out astrologically, but the real >reason is because Clinton's bosses control the >corporations that control the nation, and they also >happen to own the major media which, shall we say, help >the public make up its mind about who to vote for -- >media like CNN, NBC, The Washington Post, The New York >Times, Tribune Co., Time-Warner and others. > This is not speculation, and if you would allow me >to pre-empt this astrological column with a brief >exposé, I will explain. All of these news organizations >are members of an international institution called the >Trilateral Commission (which is listed in the Manhattan >phone book, by the way). Bill Clinton and more than 20 >of his cabinet ministers are also on the list. Bob Dole >isn't a member, and neither is Jack Kemp, so in a very >real sense, they are the outsider candidates. > Since the organization's creation in 1973, every >White House has had Trilateral Commission >representation, starting with former Vice President >Nelson D. Rockefeller, the brother of the Commission's >co-founder, David Rockefeller. In 1976, Jimmy Carter, a >member of the Commission and close associate of the >Rockefellers, was "elected" president, and so on down >the line in an unbroken chain through Bill Clinton. In >one presidential election in the 1980s, all three >candidates were Trilateralists! And in the election >right before the current one, it was Trilateralist >George Bush versus Trilateralist Bill Clinton. It's a >little like both candidates in the old Soviet elections >coming from the Communist party. > You may wonder, "What's the big deal?" But look at >it this way. What if every president or vice president >for the past 22 years had been an executive of Grumman >Aerospace or a graduate of Harvard University? You >might suspect that Grumman or Harvard had a little >extra influence and was getting its people into >positions of power. But since the nation's most >powerful media are also represented on the Commission, >its existence rarely gets reported, and its true nature >is never accurately reported except in the alternative >media. Check the National Newspaper Index in any >library and you'll find, at most, one or two references >to the Commission, and both seek to belittle its >importance. > And people who are involved in the organization >don't like to talk about it. I recently caught up with >Secretary of the Interior (and Trilateral Commission >member) Bruce Babbit while he was on a speaking >engagement at the posh Mohonk Mountain House in my >area. A number of newspaper reporters were gathered >around him asking him the usual trivia. When I asked >him directly about his involvement on the commission, >and the president's, and that of some two dozen other >cabinet members, with the red light of a video camera >coming from behind me, Bruce cracked a joke about how I >would be a popular guy "out west" (ostensibly where >everybody has weird political theories), then he turned >tail and ran away from the news conference. > Commission membership also includes key federal >legislators, people on the incredibly powerful Federal >Reserve Board, major bankers and the CEOs of numerous >corporations of the General Electric and General Foods >ilk. And it's an international institution as well, >_majority-dominated by business and political interests >in other lands, particularly Germany and Japan_. >Consider the implications. It is arguable that the >Trilateral White House is dominated by government and >business interests of foreign nations, which is >clearly, blatantly unconstitutional. > The above-mentioned national media directly >involved with this organization, which set the national >news agenda here in the U.S. and for much of the >western world, are the same ones which are manipulating >the current election by sanitizing all the news that >should be coming out of Washington and elsewhere, by >not tallying the impressive pile of corpses collecting >around the President, and most of all, by bombarding us >with PR-generated trivia and commercials 24 hours a day >when there are somewhat more significant things to >concern ourselves with. > And of course, the economy keeps getting better. > But just in case you ever wondered just how it >happened that some unknown governor of Arkansas ended >up tooling around the globe aboard Air Force One, now >you have a clue. And in case you thought people were >stupid, well, this makes a good case for people just >being manipulated. But of course, you'd have no way of >knowing this if the only place you got your information >was a television set or newspaper. Perhaps this is why >big government is so interested in censoring the >Internet. > I don't believe that the Trilateral Commission is >a conspiracy, I believe it's just one fact of how the >ruling class of the Western world organizes itself. > There are other levels of aristocratic shadow- >government organization, and they're quite worth while >to study, but this one sure points to some interesting >possibilities, and boy, the fact that you never hear it >mentioned should make really you wonder. > Not that there isn't enough to wonder about >already, and not that any of it matters. > After all, nothing really happens nowadays. Flight >800 got shot down, but nobody did it. It just sort of >exploded and fell out of the sky. Not only wasn't there >a perpetrator, there wasn't even a cause determined for >the explosion! > We have wars -- safe ones, where nobody comes back >with their arms and legs blown off. The federal budget >deficit increased by a few billion dollars in the time >it took me to write this article, and yet it's >meaningless. There is no public mention of the >Constitutional amendment, the famous 14th Amendment, >which explicitly states, "The validity of the public >debt of the United States...shall not be questioned." >It's the law of the land! > So, lest I blow a few too many of my brain cells >thinking about this, I will now go back to being an >astrologer. > End of exposé >

2 1

Re: FCPUNX:Small 'Hard' Problems Wanted
by mpd＠netcom.com 17 Dec '03

17 Dec '03

Eli Brandt <eli(a)gs160.sp.cs.cmu.edu> writes: > (I'm getting this this through fcpunks, so my reply is > going to you. If you reply, you can cc cypherpunks if you > want.) Ok. [snip] >> ... I plan to test it on reduced round DES, full 16 round >> DES, and some RSA problems of various sizes. > These are all in NP (though not necessarily NP-C), so you > can certainly solve them with SAT, but I'd be *very* > surprised if SAT were an efficient way to do it. If > SAT-reduction beats the number field sieve at factoring, for > example, you've got a major research result. Actually, I would expect any AI algorithm which has some notion of convergence to beat all of the "combination of congruences" factoring methods, which are essentially exhaustive searches, albeit it under the image of some useful transformations and homomorphisms. SAT (boolean satisfiability) does not seem to illuminate factoring or other cyptographic problems in any obvious way. Even a prime implicant covering of a typical strong cryptographic function would be mondo huge in terms of memory. Crypto problems map pretty nicely into circuit problems, however, and circuit satisfiability seems to be a nice wedge with which to attack a number of the most popular ones. Of course, you can go back and forth between circuit and SAT problems, but thinking in terms of circuits allows you to add the correct additional variables to a problem which would be gigantic if you did the output directly in terms of the input bits. I will wait to see if this thing munches through DES of various rounds before claiming any "research results." >> I already have C code to map RSA and n-round DES problems >> into an appropriate circuit satisfiability problem > How big a SAT problem does an n-bit RSA or n-round DES > problem turn into? Well - the typical DES S-Box requires between 130-155 2-input logical operations to express its 4 bit output in terms of its 6 bit input if you basically construct it in the obvious no-brainer way without attempting any clever optimization. Throw in a few XORs, and you can wire 2 round DES in about 2,500 one and two input logical operations, and the full 16 round DES takes about 25,000. I'm sure this can be improved upon greatly, if someone wanted to tweek the wiring diagram. What I've done to convert DES to a circuit satisfiability problem is to make a logical network with (56+64+64) input bits and 1 output bit. The input consists of (key,plaintext,ciphertext). I pass both the plaintext and ciphertext through the IP, and do half the rounds on each with the appropriate subkeys. Then I XOR opposite 32 bit sections between them, OR it all together, and complement the output. This yields something which lights up if the key maps the provided plaintext into the provided ciphertext and outputs zero otherwise. When it is instantated with a particular plaintext/ciphertext pair, this diagram can be munched by the previously mentioned algorithm to yield a possibly non-empty set of keys. RSA is a little bit more messy. Given a modulus between 2^k < N < 2^(k-1), you can construct a circuit which when instantated with the modulus bits will light up if and only if the larger of the two distinct primes is input. The way I do this requires only two multiplicative operations, one being a modular reciprocal of an odd number modulo 2^k, and the other being a multiplication. Both of these are wired as N^2 algorithms, using successive approximation, although near-linear algorithms probably exist. As a data point, a circuit which lights up when RSA-140 is input can be done in about a million NANDs. Someone has offered me some time on a 64 meg ultraSparc to try some RSA problems, but I am going to debug the C version on DES first. -- Mike Duvos $ PGP 2.6 Public Key available $ mpd(a)netcom.com $ via Finger. $

1 0

Re: FCPUNX:Small "Hard" Problems Wanted
by mpd＠netcom.com 17 Dec '03

17 Dec '03

Eli Brandt <eli(a)gs160.sp.cs.cmu.edu> writes: > (I'm getting this this through fcpunks, so my reply is > going to you. If you reply, you can cc cypherpunks if you > want.) Ok. [snip] >> ... I plan to test it on reduced round DES, full 16 round >> DES, and some RSA problems of various sizes. > These are all in NP (though not necessarily NP-C), so you > can certainly solve them with SAT, but I'd be *very* > surprised if SAT were an efficient way to do it. If > SAT-reduction beats the number field sieve at factoring, for > example, you've got a major research result. Actually, I would expect any AI algorithm which has some notion of convergence to beat all of the "combination of congruences" factoring methods, which are essentially exhaustive searches, albeit it under the image of some useful transformations and homomorphisms. SAT (boolean satisfiability) does not seem to illuminate factoring or other cyptographic problems in any obvious way. Even a prime implicant covering of a typical strong cryptographic function would be mondo huge in terms of memory. Crypto problems map pretty nicely into circuit problems, however, and circuit satisfiability seems to be a nice wedge with which to attack a number of the most popular ones. Of course, you can go back and forth between circuit and SAT problems, but thinking in terms of circuits allows you to add the correct additional variables to a problem which would be gigantic if you did the output directly in terms of the input bits. I will wait to see if this thing munches through DES of various rounds before claiming any "research results." >> I already have C code to map RSA and n-round DES problems >> into an appropriate circuit satisfiability problem > How big a SAT problem does an n-bit RSA or n-round DES > problem turn into? Well - the typical DES S-Box requires between 130-155 2-input logical operations to express its 4 bit output in terms of its 6 bit input if you basically construct it in the obvious no-brainer way without attempting any clever optimization. Throw in a few XORs, and you can wire 2 round DES in about 2,500 one and two input logical operations, and the full 16 round DES takes about 25,000. I'm sure this can be improved upon greatly, if someone wanted to tweek the wiring diagram. What I've done to convert DES to a circuit satisfiability problem is to make a logical network with (56+64+64) input bits and 1 output bit. The input consists of (key,plaintext,ciphertext). I pass both the plaintext and ciphertext through the IP, and do half the rounds on each with the appropriate subkeys. Then I XOR opposite 32 bit sections between them, OR it all together, and complement the output. This yields something which lights up if the key maps the provided plaintext into the provided ciphertext and outputs zero otherwise. When it is instantated with a particular plaintext/ciphertext pair, this diagram can be munched by the previously mentioned algorithm to yield a possibly non-empty set of keys. RSA is a little bit more messy. Given a modulus between 2^k < N < 2^(k-1), you can construct a circuit which when instantated with the modulus bits will light up if and only if the larger of the two distinct primes is input. The way I do this requires only two multiplicative operations, one being a modular reciprocal of an odd number modulo 2^k, and the other being a multiplication. Both of these are wired as N^2 algorithms, using successive approximation, although near-linear algorithms probably exist. As a data point, a circuit which lights up when RSA-140 is input can be done in about a million NANDs. Someone has offered me some time on a 64 meg ultraSparc to try some RSA problems, but I am going to debug the C version on DES first. -- Mike Duvos $ PGP 2.6 Public Key available $ mpd(a)netcom.com $ via Finger. $

1 0

RE: Remailer Abuse Solutions
by ph＠netcom.com 17 Dec '03

17 Dec '03

At 3:12 PM 11/14/1996, Mullen Patrick wrote: >From: Peter Hendrickson on Thu, Nov 14, 1996 13:22 >> This assumes that spam has gotten so bad that everybody filters their >> mail and only accepts mail on the "accept" list. People sending mail >> directly to your account would get a message back saying that they had >> to get on the "free" list or send their mail through one of the approved >> remailers. > So our ideas on implementation are similar, except you have made the > distinction that anyone on the "free" list can have direct access. > Unfortunately, your idea pivots on the idea spam has exploded to unbearable > proportions. My complaint on this isn't your idea, it's the projection > such an event may occur. People complain all the time about advertising spam. People have been maliciously spamming other people since the beginning of time. (The epoch here is the birth of the Net.) The problem has arrived. Right now it is merely irritating, but we have already seen calls for more control and more legislation. In my view, that would be a disaster. Let's head it off at the pass. > While I hope this plan won't ever be necessary, at least not on such a > global scale, the application of such techniques toward a mailing list > sounds decent. This may sound like a quibble, but it really isn't: I am not talking about a global scale at all. A very small group of people can make use of this technology successfully. It doesn't matter whether everybody in the world uses it or just people who are tired of spam - it still works. You only need one remailer operator to do this and it will be fully available. > I'm still thinking about how I would go about charging WRT mailing lists; > anonymous postings are puzzling me at the moment. I was probably a little quick on the keys. Charging to send to a mailing list does not solve the spam problem. Why? Because it is worth a dollar to send an advertisement to 3000 people. What it does do is cut down on "me too" posts and encourage people to make the best use of the bandwidth they are paying for. It would be easy for a remailer operator to do this if mail were accepted only from the "paying" remailer. However, if you restrict postings to an approved group of people, perhaps everybody on the mailing list, you can eliminate spam. How, then, do we allow anonymous postings to come through? Individual people on the list can receive the proposed post and forward it to the list if it is appropriate. They could even charge a fee for doing it. That's easy to do if there is a "paying" remailer which will handle the money for them. > Which brings up another topic: How would an anonymous remailer operate? > It's hard to eliminate an audit trail when there is some monetary tie back > to you, whether it be credit card, ecash (assuming they never quite figure > out anonymizing it), ... I assumed untraceable cash transactions for small amounts were available. E-cash, the product licensed by Digicash, offers full payee anonymity and would be an ideal candidate. Peter Hendrickson ph(a)netcom.com

4 3

Playing Cards
by ph＠netcom.com 17 Dec '03

17 Dec '03

A number of us have been concerned about how PGP generates entropy. Striking the keyboard beats using time as a source of random numbers, but the degree of entropy is not well understood. Are there machines where - for some reason - the keyboard strikes fall into some sort of pattern? And that's just when you are generating your public/private key pair. What happens when you are just generating 128-bit keys for individual messages? Where is the entropy coming from? I don't understand completely, but somehow PGP collects entropy from the system and then runs it into IDEA and then uses the numbers from the output. When the program is not running, a pool of this data is kept in randseed.bin. We already know it's a problem because it is hard to understand what it is even doing, much less determine if it is consistent with sound engineering practice. This problem is certainly not confined to PGP. A lot of people have worked hard to make computers deterministic. What do you do when you need entropy? Yes, you can buy a chip which generates random numbers at nearly any bandwidth you like. But, right there you have a problem. How do you know the output is really random? The answer is that you do not. Playing cards are a nice source of randomness because they are widely available and their behavior has been under study for a long time by people with strong financial reasons for finding flaws. I slightly prefer cards to dice because dice may be slightly predictable or even loaded. It would be nice if cryptography software would allow you to enter randomly selected playing cards from time to time to increase the entropy of keys. Careful people (Black Unicorn?) would enter the cards prior to sending each message. A well shuffled deck of 54 cards has about 237 bits of entropy. This is easy to use: the program asks the order of the cards, converts this to a string, and runs it through a one-way hash. (Entering the cards is a bit of a nuisance. Is there an easy way to have them read automatically?) The Economist reports that seven riffle shuffles bring a deck "very close to being random". ("Science and Technology: How to win at poker and other science lessons" The Economist, October 12, 1996, pp:88-89.) Cards have other uses. You can assign a (large) number to each configuration of the deck. This allows you to use a deck of cards to represent numbers. That is certainly convenient for generating large random numbers. I have included some Lisp code below which converts both ways between shuffled decks of cards and unique numbers. This could be quite useful for strong steganography. Let's say you have 230 bits you want to deliver across a border. It is unlikely that anybody will study a deck of cards very carefully. Of course, you need not limit yourself to cards. Any set of objects which can be ordered will do the trick. For instance, a shelf with 100 books may be used to store 524 bits. One thousand books may store 8529 bits. How many bits can you store in that database which is in apparently random order? You can hide your data in two decks. There is a natural order to cards: pick your favorite ordering of the suits and then do the rest numerically. However, if you shuffle one deck of cards and lay them in a row, you could use that ordering to define the mapping of the second deck of cards to numbers. The Enemy may seize one deck or the other, but without both, the number will not be revealed. While I hesitate to use the term, this is a one-time pad. The first deck can be thought of as the key and the second deck is the message. How the Lisp Code Works ----------------------- Let's use as an example a deck of five cards numbered from 0 to 4. There are 5! = 120 combinations of these cards. We can think of each card as a "digit" in a slightly odd numbering system. In decimal arithmetic, we have columns whose value increments by 10^0, 10^1, 10^2, etc. In our numbering system of five cards, columns increment by 0, 1, 2!, 3!, and 4!. Consider the card sequence of (4 3 2 1 0). We wish to compute the value stored in the leftmost column. Since there are five choices of cards, this divides the number of possible combinations into five pieces of 4! combinations each. So, in this example, the value in the first column will be 4*4! = 96. This leaves us with cards (3 2 1 0). We can look at this as a smaller number represented by a deck of four cards. This smaller number is added later to the larger number to get the final total number for our original deck. The first column of our smaller deck is 3*3! = 18. The next one is 2*2! = 4, the one after that is 1*1! = 1, and the last card never matters because there is only one card to choose (0*0!). The total is 96 + 18 + 4 + 1 = 119. This is the highest number we can represent with a deck of five cards. Consider the card sequence (0 1 2 3 4). This is the smallest value we can represent, 0. The value of the first column is 0*4! = 0. This leaves us with (1 2 3 4). But what has happened here? We can't treat this as a normal smaller deck because it is missing a card. Besides, 1*1! = 1, which is not the zero we expected. We must renumber the cards so that it is a reasonable deck again. The sequence of cards we should really be looking at is (0 1 2 3). This means the value of the next column should be 0*3!. Then we renumber the deck and continue until we run out of cards. The answer is 0. For an exercise, what is the value of (4 2 1 3 0)? (Answer at end.) Now, how do we construct the deck given only the number defining its combination? The code below is recursive. It computes the first digit and then calls itself to recursively find the remaining cards. The cards returned are always an internally consistent deck of cards; that is, if sorted they will be consecutively numbered starting from 0. When a new card is inserted, the cards have to be renumbered to accomodate the new one by incrementing every card which has the same or greater face value as the card being inserted. (Insertion is slightly misleading here because you place the card at the head of the list. Insertion refers to the conceptual process of inserting the card into the pre-existing order and adjusting the other values to allow it.) In other languages, this might be a little more complicated to implement. Lisp provides a nice bignum package and also (I believe) a compiler which knows how to convert the recursive routines below into iterative routines. This code performs quite well even on a small slow machine. ;;;-*- Mode: Lisp; Package: COMMON-LISP-USER -*- (defun compute-combination-number (cards) "Converts list of numbered cards into a unique number representing their order." (cond ((> (length cards) 1) (+ (* (car cards) (factorial (1- (length cards)))) (compute-combination-number (renumber-cards cards)))) (t 0))) (defun renumber-cards (cards) "Removes lead card from deck, decrements higher numbered cards by one so there are no gaps." (let ((renumbered-cards-reversed) (lead-card (car cards))) (dolist (card (cdr cards)) (cond ((> card lead-card) (push (1- card) renumbered-cards-reversed)) (t (push card renumbered-cards-reversed)))) (reverse renumbered-cards-reversed))) (defun reconstruct-deck (combination-number deck-size) "Converts unique number representing order of a deck of cards and returns a list of numbers representing the deck." (cond ((not (<= deck-size 1)) (multiple-value-bind (digit remaining-combination-number) (floor combination-number (factorial (1- deck-size))) (insert-card digit (reconstruct-deck remaining-combination-number (1- deck-size))))) (t (list 0)))) (defun insert-card (new-card card-list) "Inserts a card into a deck, increasing by one every card which is of higher or equal number so there are no duplicate cards." (let ((new-deck-reversed)) (push new-card new-deck-reversed) (dolist (card card-list) (cond ((>= card new-card) (push (1+ card) new-deck-reversed)) (t (push card new-deck-reversed)))) (reverse new-deck-reversed))) (defun factorial (number) (cond ((or (= number 1) (= number 0)) 1) (t (* number (factorial (1- number)))))) ;;; Testing routines (defun shuffle-deck (deck-size) (randomize-list (build-card-list deck-size))) (defun randomize-list (some-list) (do ((randomized-list nil)) ((null some-list) randomized-list) (let ((item-number (random (length some-list)))) (push (elt some-list item-number) randomized-list) (setf some-list (remove (elt some-list item-number) some-list))))) (defun build-card-list (deck-size) "Returns a list of consecutive numbers representing a deck of cards." (do ((card-list nil) (count 0 (1+ count))) ((= deck-size count) card-list) (push count card-list))) (defun identicalp (list-one list-two) "Returns non-nil if the two lists are identical." (eval (cons 'and (map 'list #'equal list-one list-two)))) (defun exhaustively-test-card-combinations (deck-size) "Verifies that the correct deck is reconstructed from the unique order number for every combination of a small deck of cards." (let ((combinations (factorial deck-size))) (do ((combo-number 0 (1+ combo-number))) ((= combinations combo-number) t) (cond ((not (= (compute-combination-number (reconstruct-deck combo-number deck-size)) combo-number)) (warn "Test failed for ~D" combo-number)))))) (defun test-one-deck (deck-size) "Shuffles a deck of cards and verifies that it may be reconstructed from the unique number representing its order." (let* ((original-deck (shuffle-deck deck-size)) (reconstructed-deck (reconstruct-deck (compute-combination-number original-deck) deck-size))) (cond ((not (identicalp original-deck reconstructed-deck)) (warn "Test failed for ~A" original-deck))))) (defun test-many-decks (trials max-deck-size) "Shuffles trials decks of maximum size max-deck-size and verifies that their order may be reconstructed from the unique number we compute." (do ((trial-number 0 (1+ trial-number))) ((= trials trial-number) t) (test-one-deck (1+ (random max-deck-size))))) (defun complete-combination-test () "Good test of all the card combination routines." (format t "Performing exhaustive test.~%") (exhaustively-test-card-combinations 6) (format t "Performing random test on large decks.~%") (test-many-decks 10 100)) ;; (Exercise Answer: (4 2 1 3 0) = 111)

1 0

[Fwd: [Fwd: Send this out to as many people as possible!!]]
by Thug 17 Dec '03

17 Dec '03

1 0

Remailer Pricing
by ph＠netcom.com 17 Dec '03

17 Dec '03

At 6:02 PM 11/14/1996, Mullen Patrick wrote: >Peter Hendrickson switched the transistors to say: >> E-cash, the product licensed by Digicash, offers full payee anonymity and >> would be an ideal candidate. > Oops. I forgot. :-) I guess this idea would work. But, there would be a > very large price for anonymity - two mailers holding their hands out for a > piece of the pie. > Well, I must admit you seem to have this idea pretty well thought out, and > it just may work. The hardest part of the plan would probably be gaining > acceptance. True, everyone loves getting a check, especially if all they > had to do was open and close email to have it register as being read, but > the hard part is getting people to accept having to pay to send someone a > message. Of course, there's the argument that currently we pay for stamps... > :-) It looks like I wasn't perfectly clear again. Sorry about that. The user does not tell anybody whether or not the mail has been read. In fact, it's nobody's business but their own. This is the advertiser's problem, and I don't really care whether they solve it. In fact, the user does not have to put the remailer on the accept list. He or she just has to tell the remailer operator that they would like to receive a dollar (or whatever) if the remailer operator sends them any mail. Perhaps we've been thinking about anonymous mail the wrong way. Is it like the U.S. Post Office where you have to physically go someplace, buy a stamp, physically write your message, put it in a physical envelope, carry it to a box someplace, and then wait (maybe four days) for it to arrive, all for "only" 32 cents? Or is it more like Federal Express where you pay 20 bucks and it arrives the next day, for sure, every time? Earlier today somebody sent a message about his scary former employers and (apparently) how they just kill people. Would that person pay, say, $5 to have the message delivered reliably and very anonymously? My judgement is that it would be worth every penny, and probably more. Right now the remailer network is a mess. There just aren't that many remailers operating in a timely and reliable manner. I am not knocking the remailer operators for this, it's just clear that "free" doesn't make it worth their while to keep the remailers operating perfectly at all times. It is especially important that anonymous mail be delivered in a timely and reliable manner because you probably will not have a good way to verify that it arrived. And reliability is absolutely required if you are chaining remailers because errors multiply. The top five remailers in Raph's latest report had these reliabilities: 99.77%, 99.28%, 99.23%, 99.18%, and 98.78%. If you chain those remailers, your failure rate is 3.71%, which is just too high. And, the long delays in sending messages through the remailers make it hard for people to get up and running because it takes hours to determine whether it worked, if it worked at all. Furthermore, many remailers don't use 2048-bit keys. Why not? Because they don't want to spend money on the cycles. That's okay with me - it's charity. But, if I pay a dollar for a remailer, I can expect to be able to use a very strong key. A good pricing strategy for remailers would be to charge, say, $1 for instant delivery, $.50 for 30 minute delivery, etc. To generate interest, 4 hour delays could be imposed for free remailing, if the resources are available. Peter Hendrickson ph(a)netcom.com

2 1

RE: Remailer Abuse Solutions
by Mullen Patrick 17 Dec '03

17 Dec '03

Peter Hendrickson switched the transistors to say: >> While I hope this plan won't ever be necessary, at least not on such a >> global scale, the application of such techniques toward a mailing list >> sounds decent. >This may sound like a quibble, but it really isn't: I am not talking about >a global scale at all. A very small group of people can make use of >this technology successfully. It doesn't matter whether everybody in >the world uses it or just people who are tired of spam - it still works. Yeah, I knew I used a bad choice of words. I meant global as in "all- encompassing" rather than the geographical definition. >However, if you restrict postings to an approved group of people, perhaps >everybody on the mailing list, you can eliminate spam. How, then, do we >allow anonymous postings to come through? Individual people on the list >can receive the proposed post and forward it to the list if it is >appropriate. They could even charge a fee for doing it. That's easy to >do if there is a "paying" remailer which will handle the money for them. >> Which brings up another topic: How would an anonymous remailer operate? >> It's hard to eliminate an audit trail when there is some monetary tie back >> to you, whether it be credit card, ecash (assuming they never quite figure >> out anonymizing it), ... >I assumed untraceable cash transactions for small amounts were available. >E-cash, the product licensed by Digicash, offers full payee anonymity and >would be an ideal candidate. Oops. I forgot. :-) I guess this idea would work. But, there would be a very large price for anonymity - two mailers holding their hands out for a piece of the pie. Well, I must admit you seem to have this idea pretty well thought out, and it just may work. The hardest part of the plan would probably be gaining acceptance. True, everyone loves getting a check, especially if all they had to do was open and close email to have it register as being read, but the hard part is getting people to accept having to pay to send someone a message. Of course, there's the argument that currently we pay for stamps... :-) PM >Peter Hendrickson >ph(a)netcom.com ------------------ RFC822 Header Follows ------------------ Received: by mail.ndhm.gtegsc.com with SMTP;14 Nov 1996 17:46:22 -0400 Received: from netcom18.netcom.com by delphi.ndhm.gtegsc.com with SMTP; Thu, 14 Nov 1996 22:42:05 GMT Received: from [192.0.2.1] (ph(a)netcom3.netcom.com [192.100.81.103]) by netcom18.netcom.com (8.6.13/Netcom) id OAA05098; Thu, 14 Nov 1996 14:40:19 -0800 X-Sender: ph(a)netcom15.netcom.com Message-Id: <v02140b05aeb13a70acf4(a)[192.0.2.1]> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Date: Thu, 14 Nov 1996 14:40:29 -0800 To: "Mullen Patrick" <Mullen.Patrick(a)mail.ndhm.gtegsc.com> From: ph(a)netcom.com (Peter Hendrickson) Subject: RE: Remailer Abuse Solutions Cc: "Cypherpunks" <cypherpunks(a)toad.com>

1 0

Re: [REBUTTAL] Censorship on cypherpunks?, from The Netly News
by Sean Roach 17 Dec '03

17 Dec '03

>On Wed, 13 Nov 1996, Mark M. wrote: >> Mark >> - -- >> finger -l for PGP key >> PGP encrypted mail prefered. To which, at 08:42 AM 11/14/96 -0500, aga wrote: >Why? Are you a criminal? >What are you hiding behind your PGP? Okay, I'll bite. Where is it said that a person who wants h[is,er] privacy is a criminal? Charlie McCarthy might have said that.

3 2

Re: Censorship on cypherpunks?, from The Netly News
by nobody＠replay.com 17 Dec '03

17 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- Declan McCullagh wrote: > > The mere fact that a privately-owned discussion group becomes popular > does not mean that it becomes a public forum. This is true. On the other hand, privately-owned public forums do exist. Chartered FreeNets, for example, and the facilities of private Universities like Carnegie Mellon. There may be a fine line somewhere, but I don't believe we need to draw it, because what is at issue here is the character of the moderator, not the legality of his action. We're just spinning wheels here talking in the abstract. I actually agree with Vulis's reasoning -- you need to look at the facts of the case in order to determine whether an act of censorship, or heckler control, or editing, or whatever, was proper. But apply that reasoning to Vulis's case, and you conclude that he's completely in the wrong, by any standard. > Say I start a poetry mailing list to discuss Blake's writings. I have > three people on it. One becomes obnoxious and emailbombs the list > since he disagrees with my interpretation of "A Memorable Fancy." Do I > have the right to kick him off? Clearly. And if your statement of the facts is correct, then you would be right in exercising your right. If there is more to the story, though, or if your story is wrong... well, then you still have the right to be a jerk, but you are, nonetheless, a jerk. In the cypherpunks case, though, the only jerk here is Vulis. It is worthless and IMO dangerous to prove that by arguing from first principles. The "Freedom Knight" kooks are correct in pointing out that every "philosophical" argument you've made can be and has been misused. Try a little empirical evidence; there is plenty to go around. "Truth is far more fragile than fiction... reason alone cannot protect it." Proof is syntactic; truth is semantic. > How is this different from a private poetry reading in my home? It differs in one small but important way: it all happens in the digital domain, so it can all be archived in its entirety. It is more difficult to lie about what happens online (but not impossible, as Vulis and company demonstrate). In an online forum, you have access to *all* the empirical evidence. There's no reason to rely on pure reason. So enough already with the hypothetical cases. You've got a specific case of your own that, oddly, you still haven't mentioned. - -rich -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQBVAwUBMovDK5NcNyVVy0jxAQGjIAH8DTYuUzkwt7+9i4RNCPkspNBEj7MBoH0Z CluHlmULFdamQ1HPDAXRct/DoqPzsXR+IzlMOr0y4bPFtMq1y+kEZw== =TA0F -----END PGP SIGNATURE-----

1 0