December 2003 - cypherpunks-legacy

Re: Intel, Microsoft doing Internet Phone Software
by jim bell 17 Dec '03

17 Dec '03

At 10:47 AM 7/23/96 -0400, W Lee Nussbaum wrote: > > >On Mon, 22 Jul 1996, jim bell wrote: > >> If these people REALLY wanted to promote the use of Internet telephoning, >> what they'd do is implement a system where an Internet ISP could be "called" >> over the Internet by a person wanting to place an LD telephone call to that >> area, and (presumably using A/D and D/A techniques) rather than generating >> and receiving modem tones, woudl generate and transmit the audio over the >> telephone line. That way, the target of the call would simply need to pick >> up the telephone and talk, as he would ordinarily do: He wouldn't even need >> a computer. He might not even know the call was going over the Internet. > >...see IDT's Net2Phone product, at http://www.net2phone.com/; it does >what you describe. Two notes: (1) I haven't used it yet; (2: disclosure) >I'm now employed by IDT, though in a different area. Excellent! It seems that the one remaining piece in the puzzle is financial motivation. Many ISP's might hesitate to install a "feature" on their systems which has the prospect of tying up some of their lines, with no payback. As much as it pains me to face it, at this point what is needed is some system to pay the ISP's for maintaining and even upgrading their systems to make this whole thing practical. Currently, Sprint is proud to announce that they charge 10 cents a minute for off-peak calls. Could an ISP make money charging, say, 2 cents a minute for a similar service? The main marginal cost is that of an extra phone line, which is probably about $30 per month or $1 per day. Assuming the line is occupied 8 hours per day, that's 480 minutes, which means that the phone line costs 0.2 cents per minute. Even if other costs increase this by a factor of 5 or so, there should be nothing to prevent an ISP from making money off a 2 cent per minute charge or even less. Jim Bell jimbell(a)pacifier.com

1 0

Re: Flaws of Thinkers (Jefferson, Rand, Nietzsche, Voltaire, etc.)
by tcmay＠got.net 17 Dec '03

17 Dec '03

At 11:40 AM 7/23/96, bryce(a)digicash.com wrote: >I know a lot about Rand, and about her deep flaws as an >individual, especially in terms of treatment of her supporters, >but this is the first I've heard of this one. Perhaps Tim is >thinking of a play by Murray Rothbard called "Mozart Was a Red" >in which the Ayn Rand caricature insists that her followers >smoke. > >As for smoking being "proof" of man's dominance over nature, >Rand _did_ believe that in the sense of "demonstration" or >"symbol" but she did not believe that in the rigorous sense of >"proof". This was well-known to a bunch of us in the early 70s who were interested in Rand (and her extreme followers, known widely as "Randroids"). I think some of her essays in her Objectivist Newsletter had explained why smoking was essentially de rigeur. (By the way, at the time Rand was writing this stuff, doctors were recommending smoking as a digestive and health aid, and nearly everyone smoked. I don't condemn Rand for smoking, or for falling into the all-too-common practice of using "logic" to justify one's beliefs and practices.) Rothbard was a source for the smoking example, but not in a play (although he may have also used his knowledge in a play...I wouldn't know). Rothbard wrote an article for "Liberty," circa 1986-8, which is where I read the details. Also, I believe Barbara Branden's biography of Rand dealth with this, but I haven't read it in many a year. And there is this comment, from an admittedly off-beat source (http://www.zonpower.com/zonpower/book/chapters/chapter29.html) "Whatever the root of his irrationality, Dr. Peikoff's persona shrinks with his advocating force-backed intolerance as he expressed during his 1995 Ford Hall Forum lecture. Recall how Ayn Rand's life wastragically diminished by her irrational, deadly, "dot-of-light" glamorization of smoking. Her emotional, irrational denials of the narcotically addictive, physically destructive nature of tobacco smoking led her and some of her "caped" followers to the grave." --Tim May Boycott "Big Brother Inside" 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, Licensed Ontologist | black markets, collapse of governments. "National borders aren't even speed bumps on the information superhighway."

2 1

Re: Digital Watermarks (long, getting off-topic)
by Alex F 17 Dec '03

17 Dec '03

> Easy enough. > > - Unless somebody reversed-engineered it, filtered it, and re-stamped it. Never said that it would be fool proof :) > - The entertainment industry has a reputation of being paranoid Sometimes with good reason, just like software producers are paraniod about piracy. Though, the Ent. Ind. does tend to overreact. IMO the copyright laws that are currently in place are enough to protect against the forms of piracy that they are trying to protect themselves against. I really don't think that there is need for new legislation or potentially privacy invading practices at this point. The forgers of the copyright laws (at least as they relate to music) had incredible foresight. Basically, from the laws that were originally drafted (30's maybe? Then revised in the early '70's at least as far as public domain goes) both videos and CDs are protected. These were written when there were no CDs or videos. > Some of you may recall > the flap over DAT, which significantly reduced the consumer market > penetration (the industry itself uses them all over the place). The Ent. Ind. got what they wanted though. There are taxes, etc. (some sort of import restrictions anyway) that keep DAT player/recorder devices at around $700 per unit. By this time normally the prices *should* be down to like $200 (using the CD industry as a guide) DATs are used all over the industry because they are cheap (see below), and because going from analog tape to DAT for CD mastering is a million times easier then sending off reels of analog tape, even if the tape is a properly mixed down duplicate of the master. There are still track times, numbers, etc., etc. A HUGE pain in the ass for CD manufacturers, but easy to do on a one-off basis in the studio where the producer & artist can sit there and help mix, fix tracks, select times, indexes, etc., etc. It can all be done to a single DAT )(which would then be copied for safety) and set along with a time code sheet. > (you still > have the even-less-effective argument of the associated cover art not > being included or being scanned and duplicated with reduced signal > quality, unless the distribution is all on-line). Cover art is pretty easy to duplicate if you have access to a color laser printer. Just scan the original in at 300dpi, and print it out at the same resolution/size and you have it. Just don't scan it in as a .GIF :) (too few colors) > Probably not done that way. My guess is that the disk ID is assigned > to the disk at the time of manufacturing. At the point of purchase > the customer is forced to give name, address, ID, whatever. This is > then stored in a database > > - Would YOU want to be responsible for maintaining that database? It's > like maintaining a hardware store trying to maintain an ID on every > single screw and nail in inventory. You would run into the same problems if it were done by CC. Hopefully the industry will do some sort of a cost-analysis (an accurate one) and realize that they would spend more on this than they lose (esp. since they still wouldn't eliminate piracy, just make it a little more difficult). > - Nobody's going to try and do a higher-frequency encoding (I HOPE). While > the human ear cannot hear those frequencies directly, we have found out > that those higher-frequencies interact in such a way to influence the > sound waves that influence what the user can hear. Yes, that's true. Anyone ever hear of HAARP? :) Certain frequencies can affect the brain in certain ways (a guy by the name of Robert A Monroe, while maybe a little eccentric, has been using this method since the 50's to do things like keep people awake when they are sleepy, vice versa, etc.). Also the body. Your body parts resonate a certain frequencies. For example, there is a very low note (I believe that it is a B) that vibrates at the same frequency as your bowels. Play that note, and you loose control... :) (If anyone knows this frequency, PLEASE let me know. I'm serious :) ). This is the reason > there's still a debate between digital and analog recordings, and is > still a big reason a lot of artists still record on analog equipment > (in musical "fuzzy" terms, it's equated with the warmth of the sound, > sort of like the tube-amp vs. solid-state amp debate among some guitar > players, etc.) If somebody deliberately played with such frequencies, > the journalistic media would probably have a field day. Yes, there are > audio cancelling and other tricks that could be deployed, but no matter > what, you're still deliberately introducing signal noise I touched on that in my other posting. The real difference between analog vs. digital is actually 2 things; static and musical "overtones" (used to produce various distortion effects and feedback, for example. ANyone who has listened to Robin Trower, Hendrix, Van Halen, etc. knows). People *are* playing with these frequencies. It's known as COSM or Composite Object Sound Modeling, and apparently is fuzzier (as in fuzzy logic, not fuzzy sound) than cold sampling is. Companies like Roland and Line6 are playing with such things. Roland is really doing some amazing things with this technology. > If I remember correctly, there is plenty of room in the design of the > audio CD protocal to embed such information, just like you can embed > the timing and track number information. Yeah, that's something else too. I'm not sure exactly how that works, but I *think* it's like a 1Khz or 1 hz signal that signals this. At least it is for the start of the first track on a cd. In the manufacturing process, at least > - Well, the MASS market piraters are exactly the point. Well, let's face > it, if the industry controllers got their way, there would be no > second-hand market like garage sales - there IS money involved here > (witness the bizarre dealings with CD-rental stores that have shown up The thing is, there is no money lost, really. Think about it. In order for one CD to be bought at a garage sale, someone else had to buy it at a retail store. If the record companies were in the used CD business then there may be money lost, but otherwise. The place where real money is lost is sale of promo CDs (many say "Promotional copy. Not for sale" on them). Here the record company loses nothing. The artist loses big time. With the exception of Sony records, most record companies will only pay artists royalties on 85% of records sold. The other 15% is said to be "promotional material" which is a huge scam run by the recording industry to take advantage of the artists. These 15% are still paid for (manufacturing, etc.) by the artist, and are given away to radio stations, etc. There is where the real money is lost. The rest is lieing with numbers. Getting off topic, Alex F =-=-=-=-=-=-=-=-=-=-=-=-=- Alex F alexf(a)iss.net Marketing Specialist Internet Security Systems =-=-=-=-=-=-=-=-=-=-=-=-=-

2 1

Re: DES Optimization (Brute Force DES)
by Peter Trei 17 Dec '03

17 Dec '03

> "Peter Trei" writes: > > The fastest general purpose, freely available des implementation I'm > > aware of is libdes. by Eric Young. With this, I can do a set_key in > > 15.8 us, and an ecb_encrypt in 95 us/block. That adds up to > > about 9,000 keytests/sec (this is on a 90 MHz P5, running NT). > > I'll point out that like most DES implementations, Eric's tries to > spend a lot of time in key setup to save time later on in > encryption/decryption. This tradeoff would probably be very different > if you didn't plan on trying more than one or two blocks of decryption > after getting a key. > > Perry Yep - with good optimization, the keygen and the des rounds get very close to each other in processing cost. Let's look at the steps involved (I'm refering to the DES description in Schneier, 2nd ed, p 270-278). I've not yet coded these optimizations, so this may be subject to revision. I'm assuming a known plaintext attack on a single 64 bit block, ECB mode. What we want to obtain is the key (which was presumably also used to encode interesting data which we can't read) First of all, we can move the initial and final permutes (tables 12.1 and 12.8) completely out of the testing loop. These have to done only once per run, and thus are effectively zero cost. Similarly, we can eliminate the key permutation (table 12.2), and iterate the permuted 56 bit key. If we get at hit, we invert the key permutation and add back the parity bits to get the original key. In the DES round The S-box and P-box steps can be combined into a single 48->32 bit permutation. So, per round (there are 16), we now have: 1. copy 32 bit sub-block to use as 'other' half in next round. 2. expansion permutation (32 -> 48) 3. xor with appropriate subkey. 4. perform the s-p permutation. (48->32) 5. xor with 'other' half from previous round. The key scheduling can be done in parallel with the rounds, since we're only planning on using each key once. However, on the Pentium it may be more efficient to do it before the des rounds, due to register starvation. Generating the subkeys is actually quite painful. You have to rotate the 56 bit key as two 28 bit halves, by one or two bits depending on the round, and then do a 56-> 48 bit permutation to generate the subkey. This step needs to be done 16 times. In a regular DES implementation, you generate the key schedules once at the start, and reuse them for each block, so there is little to be gained by optimzing this step. In a keysearch situation it's a different matter. Optimizing the permutations: DES was originally designed for hardware implementation, where permutation is a simple matter of braiding the wires between the input and output. It's a lot harder in software. I'm aware of two basic approaches: 1. Algorithmic: Analyse the permutation table to find bits which get shifted in the same direction, by the same number of bits, and arrange a series of SHIFTs, ANDs, and ORs to generate the desired permutation. This is essentially a geometry problem. 2. Lookup: Create tables for the permutation. While a permutation with n bits of input requires 2^n entries (each the size of the output data) if done as a single table, it's possible to break the permutation into several smaller tables, at increased processing time. This is a classic speed vs space tradeoff, with a big step if your tables are too large to fit into cache (and even bigger if they go to virtual memory) Example: A straight 32 -> 32 bit permutation: If done as one table, it would have 2^32 entries, and take about 16 Gbytes. If broken up into 4 tables, each of which dealt with 8 bits of the input, it would take 4096 bytes total. However, the calculation would require extracting the four eight bit subkeys from the input, doing four lookups, and ORing the four results together to get the final output. A 32 bit perm could also be done with two tables, but they would occupy half a Mb. The S-P step and the compression step of the key schedule are probably faster by lookup than by algorithm. I'm not sure about the expansion permutation, which is very regular. The size and number of tables used is going to depend a lot on cache size and available memory - for example, the rotates in the key scheduling can be eliminated if I'm willing to maintain 16 key compression tables (one for each round). I strongly expect that the whole key testing loop can fit into the 8k L1 code cache. The lookup tables *may* fit into the L2 cache. If anyone has any other optimizations, I'd like to hear about them. Peter Trei trei(a)process.ocm Peter Trei Senior Software Engineer Purveyor Development Team Process Software Corporation http://www.process.com trei(a)process.com

1 0

Kids and Computer Privacy Was Re: No more stupid gun thread ...
by Scott Schryvers 17 Dec '03

17 Dec '03

At 06:24 PM 7/22/96 -0700, you wrote: >Ok. This thread has gone on long enough and covered just about every >point except the one which I originally made in my first response, >which is that I abhor the idea that kids should carry weapons (of any >sort) to school as standard equipment. In fact, I abhor the idea >that kids should carry weapons at school for any reason. > >Enough said. > >I do not care to discuss: > >1. Should kids have any weapons at any time? > >2. Should kids have guns (specifically guns)? > >3. Should kids know how to operate weapons of any sort? > >If any of you really really have to discuss this issue, let's spare >the rest of the list and send me E-Mail directly. > >Thanks! > >Ern > Should kids have crypto? [weapons] Should parents teach their kids crypto at an early age? Should kids have a right to privacy? If yes, should there be a limit to the privacy they hold? Under Itar crypto is a weapon. If a kid were to bring a disk containing pgp to school could they be expelled for carrying a weapon? Should children's legal rights be taught at school at an early age?

1 0

emscrypt 0.01 ALPHA
by Matt Carpenter 17 Dec '03

17 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- I finally managed to find some time to do a little testing of emscrypt and make some of the changes that were suggested here earlier. It is still rather ugly and I haven't added many of the features/capabilities I hope to eventually implement. But I'm making it available so that people can play with it if they want. It looks like I'm going to be very busy for at least the next month trying to finish up the work for my master's degree, so I probably won't be making any major changes for a while. emscrypt's purpose is to automatically run PGP signed scripts received by e-mail and return the results to the submitter. emscrypt is a heavily mutated version of morepgp, originally by Jason Steiner and modified by Greg Spencer. Please realize that this program has NOT undergone extensive testing, so you may encounter strange behaviour. Make sure you read the documentation that exists. Using this program may make your system insecure, especially if you don't follow the installation procedure carefully. Use at your own risk. You should be able to find emscrypt and some documentation at: http://www.bmen.tulane.edu/~carpente Look under the "Other random projects or possibly useful stuff" heading. In order to try out emscrypt you need the following: * Perl -- emscrypt is written in Perl. * PGP 2.6.2 -- Other versions may work, but I haven't tested them and emscrypt relies strongly on knowing the format of the PGP output messages. * Procmail -- For passing incoming mail to emscrypt. This can be accomplished in other ways, but at this point I haven't tried any of them. If you don't have procmail you can still play with emscrypt by piping messages to it manually. I strongly suggest you do this anyway, to make sure you trust emscrypt to answer incoming mail. Here are the major things that have changed with emscrypt since I described it here a while ago: There are now two required headers which must be included in the body of the signed message: 'Reply-To:' and 'PGP-Key-Fingerprint:'. Both must appear before the beginning of the script. The beginning of the script is considered to be anything other than the above headers, blank lines, or lines that begin with '::' but that are otherwise empty. Duplicate headers will generate an error and prevent the script from executing. Case within the headers is not important. (You are free to use eLiTE d00dz capitalization techniques, just don't try to use kRe8yv sP3lliNg5.) 'Reply-To:' must be followed by a valid e-mail address, otherwise you won't get your results. (emscrypt will also recognize a 'Request-Remailing-To:' header in place of 'Reply-To:'. I added this so that emscrypt could be treated as a "remailer". The idea is that you could use software such as premail to automate the generation of messages going to emscrypt, so that you don't have to do the signing/encrypting manually. Then you could just send a script to yourself, "chaining" it through your emscrypt "remailer". But I don't think this will work without slight modifications to the remailer message generating programs, since emscrypt requires the 'Request-Remailing-To:' header within the signed body of a message (it would also be nice if the 'PGP-Key-Fingerprint' header could be automagically included)). The 'PGP-Key-Fingerprint:' header gives the fingerprint of the key used to sign the message and which will also be used to encrypt the results. It must also be included within the signed body of the message. Most of the other changes involved minor debugging and general clean-up (not that I consider it clean now). I also improved the error handling a little, but it needs more work, too. Right now emscrypt generally tries to generate, encrypt, and mail an error report if things go wrong. If that doesn't work, then it saves an error message to a log file. If you find any problems, please let me know. Comments, suggestions, etc. are also welcome. Thanks, - --Matt - -- mcarpent(a)mailhost.tcs.tulane.edu -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQEVAwUBMfXlaijtJAMyBnp9AQHAdAf+MQ/ZroKoLeYyQDYabVrIq1eLSQB6vpr+ 2tXu63wDbcUeFoeSFNx6Sar7DNtAJyJUlwVcVKlb5SOuYR/8aFDvAnIYuQLPfdXd xXjC4iv+Hh3hNx4ibeyAB4xbFmDYAMB19zEf6nhmJdxR03oFXP+Qfx2m/aN/LDKZ zVSjtOs/ujTa6ltP6r/9x1vdiqmSNuSCNLvL/f4YulfdzR8frF0uLyLmiDH6mUpm etKxSpIg4ZI+iy1YvvSd+FtA0F3XSziaLEepx4X8gYjZKP0YXPTEMGhTeWj1o6bP yl06KUIRTL5k56P0xnW2MezGN5c0Cz2W9TEa9NBbHxY1DgSTDWa3sg== =DTpY -----END PGP SIGNATURE-----

1 0

Re: DES brute force? (was: Re: Borders *are* transparent)
by Dan Bailey 17 Dec '03

17 Dec '03

If we choose a plaintext/ciphertext pair carefully, we can easily save ourselves some work (50%) while still making the attack a credible demonstration. The idea is to use each trial encryption twice. In _Differential Cryptanalysis of the Data Encryption Standard_, Biham and Shamir note the following, known as the complementation property of DES: if T = DES(P, K) where T is ciphertext, P plaintext, and k key, then: T' = DES(P', K') where T', P', and K' are the bitwise complement of the above. Now the interesting part. If two pairs (P1,T1) and (P2,T2) are available with P1 = P2' or T1 = T2', then an attacker can restrict his search to only the keys with LSB = 0. The attacker runs through the remaining 2^55 keys (with LSB = 0) and tests the results against both T1 and T2'. Since testing for equality is much faster than performing the actual encryption, time savings is on the order of 50%. Just a thought on how to save some cycles. Dan

1 0

Re: Digital Watermarks (long, getting off-topic)
by tcmay＠got.net 17 Dec '03

17 Dec '03

(My comments are really minor quibbles, based on my longtime use of DAT machines, which I now have three of.) At 2:19 PM 7/23/96, Alex F wrote: >The Ent. Ind. got what they wanted though. There are taxes, etc. >(some sort of import restrictions anyway) that keep DAT player/recorder >devices at around $700 per unit. By this time normally the prices >*should* be down to like $200 (using the CD industry as a guide) As I said, I have three DAT machines. They are complicated machines, having lots of moving parts and precise tolerances. (And they are prone to break!) CD players are vastly simpler. I'm not at all surprised that prices have remained at about the $400 level for DAT decks, and about the same for DAT portables. After all, camcorders, which use much the same technology, have also remained at about the same price. And I don't think the SCMS code had too much to do with mass-acceptance. Most comsumers, according to available figures, *buy* C-90 cassette tapes, and do not make their own. (That _you_, the CP list reader, may use your cassette deck to make tapes has little to do with the vast numbers of cassette users out there do...most don't know how to record with their cassette decks.) Pre-recorded DAT tapes were available for a while...they did not sell. I believe this was because DAT machine purchasers were sophisticated and new how to make CD-to-DAT copies, with or without SCMS. Thus, the failure of DAT as a consumer medium (not to mention the much-hyped MD and DCC formats) probably is due to other reasons, including the mechanical issues, the lack of a real need for consumer DAT, and the confusion over new emerging formats. --Tim May Boycott "Big Brother Inside" 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, Licensed Ontologist | black markets, collapse of governments. "National borders aren't even speed bumps on the information superhighway."

2 1

Re: A Snake-Oil FAQ
by David E. Smith 17 Dec '03

17 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- To: cypherpunks(a)toad.com Date: Wed Jul 24 00:10:54 1996 > Honestly, this boycott campaign looks out of place on Cypherpunks, at > least to me. I mean, we are for freedom of speech, aren't we? > Sternlight is talking about on-topic things. How come that renowned > defenders of freedom of speech resorted to name calling and attempts to > push their opponent out of the public forum? > (cheer) I don't agree with David Sternlight; frankly, I doubt very many of us agree with David Sternlight. But if we're going to start plonking just on the basis of a periodic rant, after he goes Perry goes, and Tim goes, and before long there's no list. Now to keep things interesting, somebody wanna subscribe Dorothy Denning to the list? Might be fun :) dave - ---- David E. Smith POB 324 Cape Girardeau MO USA 63702 dsmith(a)prairienet.org http://www.prairienet.org/~dsmith send mail of 'send pgp-key' subject for my PGP public key "I'm only a social smoker, just a few packs a day really" -----BEGIN PGP SIGNATURE----- Version: 2.6.2 Comment: Automagically signed with Pronto Secure for Windows. iQEVAwUBMfWwaTVTwUKWHSsJAQHmAwf/XkwZZtAbymiSni4L8N6D6qW75W2B/Hgv yuPX0uCoXnBmvUXzq6RS7zNN0A1vtgXhCgzgyYY+1/1PRAAoQAd8q7evX1K+rTVe 1MmlOfJWFHJG9q0tHc45LONagW1m0eAp9Z9fkO4NRhGpybfWDCZKwtClbaLiAdoA 86lJnNRacjEHf1dAAHc/1dibJz+6617nfgLkea0OVbLICIsSpkc3PJLkFE9jjK2a g+5CXJ8oXg4bKlEw7QoefagXYsM9o6PF+MHIDVT/pZnfoVmGoI9+3BYHkByXh/vQ 8M+w6hZu4X50zzxCx/6lDvzfd1zccmHvJAsx6GXf4VGUp6lIw+lfCg== =KXCp -----END PGP SIGNATURE-----

1 0

Re: Bare fibers
by Bill Stewart 17 Dec '03

17 Dec '03

h > > Doesn't that make it vulnerable (detectable) to Tempest attacks? h > > Harka rw> No. rw> Transmitting light via fiber doesn't emit EM. Light is electromagnetic :-) rw> Anyway, the original post, as I recall, was about keeping sensitive data rw> on a second hard drive, connected via (very thin, therefore harder to rw> notice) fiber. Tempest monitoring was not a factor. rw> -r.w. I assume that if you've got a fiber hanging from your PC, and thugs come in to steal/confiscate/forfeitize your PC, that they're bright enough to notice it and maybe to follow it. Thugs who want your information will follow it to find information; thugs who just want to resell your hardware will follow it to find more hardware. I had initially assumed that the mention of fiber optics was in the context of "Infrared transmitter on the PC, fiber optic sticking out of the wall to receive the IR and transmit it to a hidden detector." or something silly like that. If you want to hide a small PC in your attic/wall/etc and use the newer faster IR stuff for clandestine backups, and have a Real Operating System so you can run it in the background (since your disk drive is presumably much faster than IR), I suppose you could do that. The main use of TEMPEST here is to detect backup systems hidden in the attic that they hadn't noticed. Hiding it in your stereo system doesn't protect it from honest thieves, who might also want to resell your stereo, and info thieves have been known to seize anything even resembling computer equipment, such as Mozart CDs. You could also hide the computer in your kitchen cabinets like one Famous Cypherpunk, but you'd have to use a thinner bundle of cables than he does to connect the monitors and keyboards :-) # Thanks; Bill # Bill Stewart +1-415-442-2215 stewarts(a)ix.netcom.com # http://www.idiom.com/~wcs # Confuse Authority!

1 0