December 2003 - cypherpunks-legacy

Noise source processing
by mgraffam＠mhv.net 17 Dec '03

17 Dec '03

Hello all.. After a little bit of work I have put together a white noise source (hacked FM reciever, heavily shielded to reduce local EMR bias) and have it jacked into a mic port on one of my machines. I would like to develop of bit of code (oh, horror :) to take bits from the sound card, and bits from /dev/random, and mix them together to get a random number stream. The noise coming off of the sound card is more beige than white though.. Does anyone know of any papers, articles or whatever on good techniques to remove bias from the audio source? My initial thoughts are along the lines of just hashing everything, but this will be slow, and I'd like to see what other ideas are out there. Michael J. Graffam (mgraffam(a)mhv.net) http://www.mhv.net/~mgraffam -- Philosophy, Religion, Computers, Crypto, etc "Let your life be a counter-friction to stop the machine." Henry David Thoreau "Civil Disobedience"

5 9

Re: text analysis
by jim＠mentat.com 17 Dec '03

17 Dec '03

CyberPsychotic writes: > So I was playing with some crypto-analysis algorythms, and some steps of > them involve such thing as finding out the frequency of some character or > set of characters. So I wonder what would be the optimal (speed, resources > etc)way of coding this. While playing with single character frequency, I > guess the best way would be having unsigned int 256 elements array (I > refer to C coding) and each time, I find certain character, i just > increase the element of the array with this char offset. This seems very > neat to me (except the thing that some chars could be never found in > text). Anyways, when things come to 2 characters set, i have to get 1024 > character set, and so on, which looks quite unreasonable to me to allocate > memory for elements, which probably will be never found in text... I was > thinking of other solution and came to two way connected lists (correct > term?) things, i.e. : i have some structure like: The term is "linked list". If you want to collect all the arbitrarily long repeats you may have to resign yourself to a two-pass algorithm. For small cryptograms like the ones I usually deal with, it's sufficient to use an N^2 algorithm, looking at each starting point and each possible offset to see whether a string of at least k characters is duplicated at that offset. For a longer file, off the top of my head, you might start with a set of pointers for where each possible digraph starts; you can use a linked list, storing all the starting locations for each digraph (a two-dimensional array of pointers, total size 64K). On your second pass you can do an N^2 search within those reduced sets to see which repeated digraphs can be extended. It will help a lot if your target ciphertext fits in memory. I make no claim that this is optimal. Jim Gillogly

1 0

CTIA Daily News from WOW-COM - August 7, 1998
by CTIA Daily News 17 Dec '03

17 Dec '03

========================================== Welcome to today's edition of the CTIA Daily News from WOW-COM. Please click on the icon / attachment for the most important news in wireless communications today. If you cannot open the icon/attachment, please visit http://www.wow-com.com/professional/ for the wireless daily news. Don't miss the major conference on implementing Phase 2 of the E-911 requirement--August 26-27 in San Francisco. For more information, click http://www.wow-com.com/pdf/locntech.pdf ===========================================

1 0

text analysis
by CyberPsychotic 17 Dec '03

17 Dec '03

Hello folks, I have abit practical question, since I know this list is partly devoted to cryptography matters, I guess this is the right place to ask: So I was playing with some crypto-analysis algorythms, and some steps of them involve such thing as finding out the frequency of some character or set of characters. So I wonder what would be the optimal (speed, resources etc)way of coding this. While playing with single character frequency, I guess the best way would be having unsigned int 256 elements array (I refer to C coding) and each time, I find certain character, i just increase the element of the array with this char offset. This seems very neat to me (except the thing that some chars could be never found in text). Anyways, when things come to 2 characters set, i have to get 1024 character set, and so on, which looks quite unreasonable to me to allocate memory for elements, which probably will be never found in text... I was thinking of other solution and came to two way connected lists (correct term?) things, i.e. : i have some structure like: struct element { char value[ELEMENT_LENGTH]; unsigned int frequency; struct element *previous; struct element *next; } and could dinamically allocate memory for each new found element, but this would slow down whole code by the time list of new elements grow up. (b/c i will have to look thro.. whole list in order to find whether such element has already been found or not), so I wonder maybe there another neat way to complete tasks like this? I would appreciate any ideas, hints, code examples. (would be very helpful having a look on some code performing such a task, I surfed the web but didn't find many)... Thanks beforehands, Fyodor

3 3

Follow-up Letter
by wcts＠globalvisionexpo.com 17 Dec '03

17 Dec '03

From: Charles H. Borden President, GlobalVision Sarasota, Florida Please kindly direct to the Sales and Marketing Decision Maker Please accept this important reminder as follow up to my letter of last week about GlobalVision's potential effective marketing of your firm on our upcoming world tour. If you will recall, we promote "global-minded" companies at our exclusive overseas 'Direct from the USA' catalog expos. Once you've registered, we will invite key overseas distributors, joint-venture partners, government buyers & licensees in your industry. Remember, you stay home while we do the legwork. We have space available for Mexico, Brazil, Argentina, Venezuela, Australia, Canada, Singapore, S. Africa, Chile & Colombia. Plus we have World Catalog Trade Shows in Europe and Japan. These special events are joint ventures with World Trade Centers Association. Fees range from just $375-$695 per expo with a money back guarantee. If you are considering doing business in any or all of the above markets please respond by E-mail today. Simply hit your reply button and send us a short note. Please be sure to include your NAME, COMPANY, AREA CODE and PHONE NUMBER. Susan Khrystal, our V.P., will contact you by PHONE during business hours. If you have received this message in error, please accept our apology. If you wish to be excluded from future mailings, simply press your reply button and send us the word "remove" in the subject line. Sincerely, Charles Borden President, GlobalVision

1 0

Re: SecDrv and Win95
by Anonymous 17 Dec '03

17 Dec '03

Ian F. Silver wrote: > I use Blowfish and GOST (depending on the size of the container I'm > working with), and avoided DES, even before Deep Crack was invented. > :-) > "Deep Crack" was invented in 1993 by John Gilmore. (based on a paper at Eurocrypt '93). It was implemented more recently. BTW this is all in the archives in 1993. Everything was discussed in the first 12 months of cypherpunks.

1 0

Half a mil for a spill:Clinton's expensive cock-up/Robber shoots own knob off
by STUFFED/EUREKA! 17 Dec '03

17 Dec '03

Welcome to today's issue of Stuffed. To read it you should click on the URL below. If it is not made clickable by your email program you will need to use your mouse to highlight the line, copy it and then paste it into your browser (then press Return). http://stuffed.net/98/8/6/ If you have an email program that can display web pages you will find today's front cover should have already been disp- layed and can be clicked on to take you straight to Stuffed. This email is never sent unsolicited. Stuffed is the supplement for the Eureka newsletter you subscribed to. Full instructions on unsubscribing are in every issue of Eureka!

1 0

RE: compression vs encryption
by Brown, R Ken 17 Dec '03

17 Dec '03

> Bernardo B. Terrado[SMTP:bbt@mudspring.uplb.edu.ph] seemed to write: > > What's the difference between compression and encryption? Eh? Encryption means keeping things secret. Compression is keeping them small. Not the same thing at all. OK, they might use some of the same tools - but you can use a screwdriver to clean your nails. > Correct me if I'm wrong, in compression we lose something, like in > image compression and audio compression, but in encryption, we don't. Be corrected. You are wrong. There are ways of compressing that don't lose anything (like PKzip that you might have heard of...). There are ways of encrypting that don't compress. (in fact you might not want totally efficient compression in encryption because that might give an eavesdropper clues as to the real length of the plaintext which could be useful to them) > > One sure protection againts evil hackers is encryption (?), > but what if that evil hacker deletes the ciphertext file and the > plain text file (that is assuming the evil hacker just deletes > every file he wants to delete). I go to the backup, that's what. If I care enough about a file to encrypt it I care enough about it to back it up. Your hacker is doing what they call a "denial of service" attack. Encrypting your files doesn't make any difference is someone deletes them. You need different defences against different kinds of attack. Or was all this some sort of troll... > ====================================================================== > ========= > You give the words you have spoken, > it is not lended and are not taken back. > It is like what The Corrs sung "...you're forgiven not forgotten...." > > bbt(a)mudspring.uplb.edu.ph > bbt(a)peak-two.uplb.edu.ph > metaphone(a)altavista.net >

1 0

Re: text analysis
by Anonymous 17 Dec '03

17 Dec '03

A good data structure is a tree: -----> 2nd letter / ------> 3rd letter / / 1st letter -------> 2nd letter -------> 3rd letter \ \ \ ------> 3rd letter -----> 2nd letter \ -----> 3rd letter Start with an array indexed by the 26 letters. Each array entry contains a count of that letter's frequence and a pointer to a subsidary array. The pointer starts out null. When you find a new digraph (like the first time you find a letter after 'g') you allocate a new structure to represent the letters following 'g'. This will again consist of 26 entries like the first one. Likewise when you find a new trigraph, you allocate a new third-level array of 26 entries. This will be a compromise between space and speed. Most digraphs don't exist so you will save considerable space over the brute force way of allocating an array of 26^n spaces to count n-graphs. However it is a bit wasteful in that it allocates a full 26 entries each time even though only a small fraction will be used. The last layer does not need to have the pointers allocated per letter since you won't be looking beyond that (e.g. if you are only interested in counting trigraphs, the third layer above can consist of just an array of counts). struct entry { int count; struct entry *next; } top; Handle letter, given array of previous letters. prev[0] is current letter, prev[1] is 1st previous letter, prev[2] is 2nd previous letter, etc. Size of prev is n entries. Letters are normalized to range 0-25. Untested code, hopefully it will give you the idea. handle (int prev[], int n) { struct entry *cur; cur = top; for (i=0; i<n; ++i) { cur[prev[i]].count += 1; if (i==n-1) break; if (cur[prev[i]].next == NULL) cur[prev[i]].next = calloc (26, sizeof(struct entry)); cur = cur[prev[i]].next; } } This actually counts the entries backwards, so that "the" is entered under top['e'-'a'+1]->next['h'-'a'+1]->next['t'-'a'+1]. You can correct for this when you print them out. Or you can store the letters in the prev array in the opposite order, putting the current letter in prev[n-1], then they will be in the correct order.

1 0

Re: Noise source processing
by Enzo Michelangeli 17 Dec '03

17 Dec '03

-----Original Message----- From: mgraffam(a)mhv.net <mgraffam(a)mhv.net> Date: Thursday, August 06, 1998 10:17 AM [[...]>The noise coming off of the sound card is more beige than white though.. > >Does anyone know of any papers, articles or whatever on good techniques to >remove bias from the audio source? > >My initial thoughts are along the lines of just hashing everything, but >this will be slow, and I'd like to see what other ideas are out there. What about an analog filter placed before the digitizer? Nowadays there should be inexpensive single-chip implementations of multiple-tap equalizers made for consumer audio. You just have to straighten up the amplitude ignoring the phase rotation, because the power spectrum is a real-codomain function (being the Fourier Transform of the autocorrelation, which is an even function). It won't be perfect, but it'll help. Enzo

2 1