December 2003 - cypherpunks-legacy

Re: Moving from 1024-bit -> 2048-bit key.
by an169306＠anon.penet.fi 17 Dec '03

17 Dec '03

I am wondering if going to 2048 bits is worth it or not. After a certain point, it doesn't matter whether it takes 200 quadrillion years or 200 decillion years to factor a key, it will be a long time before its factored ;-). Perhaps keep your 1024 bit key, unless you want to expire it and generate a new one. (I did that, expired two keys, replaced them with a 2048 bit key, but realize that a 2048 bit key takes somewhat more time and processing power to do the usual PGP hijinks with. I probably would be better off just using a 1500 bit key, but too late now...) Just remember (so your web of trust stays intact) to sign your new key and use your new key to sign your old one before revoking the old key. This confirms that it IS a new key, rather than someone trying to spoof you. Also, keep a backup of your public and secret keyrings before revoking the old key. This is just common sense. Make sure your backup is secure, though. (I PKZIP my PGP stuff, encrypt it several times with convential key encryption, wnstorm it, THEN lastly use Norton Diskreet with a password that is different from all the rest. If Diskreet is not secure, PGP and wnstorm are, so I don't really care if Diskreet has holes in it or not. Its just there to 'spice things up' for anyone trying to crack the backup, which is stored in a place any party I authorize or I can easily get at it. If anyone sees something weak about this, E-mail me... ;-) The main thing is not to go keysize crazy which I did, as you do pay a performance penalty for those huge keys, and make sure that your other avenues of attack are defended against. (As soon as I can, I am buying a portable 486 or Pentium, since my main weakness is a 'black bag' attack using a good keystroke monitor or a hardware bug, then a simple copy of the secring file when I mount my SecureDrive partion and the Stacker partition inside of that to access my PGP stuff.) ------------------------------------------------------------------------- To find out more about the anon service, send mail to help(a)anon.penet.fi. Due to the double-blind, any mail replies to this message will be anonymized, and an anonymous id will be allocated automatically. You have been warned. Please report any problems, inappropriate use etc. to admin(a)anon.penet.fi.

1 0

voice encryption for portable PCs
by an109574＠anon.penet.fi 17 Dec '03

17 Dec '03

One of the problems with Philip Zimmermann's very worthy voice encryption project is the lack of support for PC laptops, notebooks and palmtops that lack soundcards. If anyone involved in this project is listening, please consider creating a version that will work well with the PCMCIA "Wavejammer" card or some other peripheral soundcard substitute. Cogon Electronics Inc., as has been discussed in this group, markets a voice encryption program called Aquafone that employs Diffie-Hellman+triple DES. The current version uses a push-to-talk feature like a CB radio, but the company plans a fully duplexed version in March. Like the Zimmermann project, the Aquafone uses a standard soundcard. The company reports mixed results using Wavejammer cards. The audio quality afforded by the Wavejammer is noticeably inferior to that of a soundcard. Seems the audio digitization algorithms would have to be tweaked to optimize the fidelity using the Wavejammer. BTW, please don't misconstrue this as an advertisement for Wavejammer or Aquafone. I fully expect Phil's voice encryption software to become the de facto telephonic standard just as PGP became the Internet standard. Cogon's Aquafone, while commendable, has the disadvantage of employing DES. I presume Phil will select a more resilient conventional algorithm. Cogon also lacks Phil's cachet as certified by his fans at Justice and Customs. And, of course, the Zimmermann software will be freeware. Likewise, I mention Wavejammer because that is the only PCMCIA substitute for a soundcard with which I am familiar. If someone can suggest other solutions to the sound card deficiency in PC laptops and notebooks -- short of buying a Mac or restricting one's selection to the few PC portables that have internal soundcards -- please do. I don't regularly monitor this group, so I would greatly appreciate any replies by e-mail. This is an important problem to solve. After cryptanalysis, traffic analysis is the biggest threat to secure telephony. A readily portable PC armed with Phil's voice encryption program represents a real dilemma for traffic analysts since one or both targeted parties are no longer tied to specific telephones. I am neither a telecommunications expert nor a programmer. I am a civil liberties activist. Periodically, from the safety of the United States, I must communicate with people whose courage and idealism puts them in jeopardy of hostile parties. PGP is quite popular in such circles. As you might expect, Phil's voice encryption software is awaited with equal enthusiasm. ------------------------------------------------------------------------- To find out more about the anon service, send mail to help(a)anon.penet.fi. Due to the double-blind, any mail replies to this message will be anonymized, and an anonymous id will be allocated automatically. You have been warned. Please report any problems, inappropriate use etc. to admin(a)anon.penet.fi.

1 0

Re: Moving from 1024-bit -> 2048-bit key.
by bryce＠cybernet.co.nz 17 Dec '03

17 Dec '03

>Just remember (so your web of trust stays intact) to sign your new key >and use your new key to sign your old one before revoking the old key. >This confirms that it IS a new key, rather than someone trying to >spoof you. You would want to sign your new key with your old key wouldn't you? Anyone can add a signature to your old key, but only you can add your (old key's) signature to a new key. I agree that signing your old key with your new key is probably a good idea, as it shows that the owners of the key (the same person) trust that the keys belong to who they say they are. ------------------------------------------------------------------------------- Bryce Boland <bryce(a)cybernet.co.nz> Finger for PGP Public Key http://cybernet.co.nz/home/bryce/www/home.html

1 0

PGP at MacWorld Expo?
by crawford＠scruznet.com 17 Dec '03

17 Dec '03

The MacWorld Expo is coming up soon in San Francisco, I believe from the 4th through the 7th of January. There must be _someone_ on this list who works for a company or user group that will be exhibiting (I know for sure the Berkeley Macintosh User's Group always exhibits, as does the Boston Computer Society). Do the world a favor. Distribute PGP from your booth. This being the MacWorld Expo, you'll need mostly Mac disks, but you would want to have some Unix and DOS disks on hand as well. Have flyers about the Digital Telephony Bill, to clue in the legislatively impaired. Imagine what a crowd you'll draw to your booth with a sign reading "Unbreakable encryption software - just $3 a disk". You might even make headlines in the trade rags or the SF newspapers. Certainly BMUG and BCS already have the "infrastructure" to sell freeware disks. But any of your companies out there would do well to sell freeware disks too. I understand, from my own experience of working at MacWorld (no, I won't be exhibiting...), that one of the biggest objections to this will be simply dealing with getting the disks duped and labeled, getting them to the show floor, and dealing with the large number of people who will be purchasing the disks and needing you to make change. This could be no small amount of impact on your booth operations. Another option is to offer PGP free with any purchase of your regular product, or to just distribute flyers about the DT bill with instructions for obtaining PGP from the Internet and online services. Regards, Michael D. Crawford crawford(a)scruznet.com <- Please note change of address. crawford(a)maxwell.ucsc.edu <- Finger me here for PGP Public Key.

1 0

Milton on CSPAN
by Adam Shostack 17 Dec '03

17 Dec '03

Got this, thought there might be interest. | if you have cable, Milton Friedman will be discussing Hayek on CSPAN | (I'm not sure whether I or II, I think I) this Wednesday at 8.

1 0

Re: Breaking into girlfriend's files
by dmandl＠bear.com 17 Dec '03

17 Dec '03

Black Unicorn <unicorn%access.digex.net(a)bear-lock.bear.com> writes: > I guess I suffered from the silly idea that as a whole, the members of > the list would not put short term morality before the long term goal. Do me a favor, Uni: I want to assassinate you and all the other members of your family. Could you send me some powerful handguns and silencers? Oh, and while you're at it, give me all of your addresses and precise travel schedules for the next two weeks. As an "amoralist," I'm sure you won't put short term morality before my long-term goal. Thanks. > It seems there are those who disagree with me. I would offer the > following: While it may be that Joe Break-Into-Girlfriend's-Files may or > may not be justified, that is not ours to judge. I'll judge what I want to, OK? What are you trying to do, _coerce_ me into giving information to someone that I don't want to? > It is simply for the > crypto non-challenged to comment on the security or lack thereof of a > given system. Hmm, sounds like you're telling us what to do. And defining what "we" are, to boot. I'm many things besides "crypto non-challenged." I'm an individual and an individualist, not a cog with a particular static role in society, which is apparently the way you'd like it. If you ask me what the color of the sky is, I can say "39" if I want. > As soon as this list turns into a pile of bleeding heart > liberals, anxious to embroil themselves in the personal matters of > others, it has failed. Yeah, well it seems you were just as quick as the "bleeding heart liberals" to scold others and impose your personal (a-)morality. Typical. > I cannot believe that people on this list, those who claim to be > interested in the preservation of privacy, would support the proposition > that knowledge about the strength or weakness of a given system should be > surpressed. No one was suppressing anything or in any way attempting to stop the flow of information. They were just expressing their opinions. Just because _you_ were unhappy with those opinions, that's no reason to hurl accusations of censorship around. I can refuse to give help to anyone I don't feel like helping, and if I feel like giving him a piece of my mind in the process, that's my business and his. > I submit that the response should have been the same regardless of the > nature of the material. > > Which is it going to be? > > > 1> > > Q: "How do you attack X?" > A: "Y" > > or > > > 2> > > Q: "How do you attack X?" > A: "Realistically X should not be attacked, because to allow the > widespead lack of confidence in X will destroy society as we know it, and > anyhow it's nasty." Great, I've always wanted to be a simple input-output machine. Pseudo-individualist Republican rubbish... --Dave.

2 1

credit card charges
by A5113643667＠attpls.net 17 Dec '03

17 Dec '03

Peace @ bix.com has posted a Windows program mpj/public/pgcharge.zip to ftp.csn.org that will create an order with a credit card using X12 formats that will be signed and encrypted with PGP and then sent to a mailer. The posting asks for volunteers to serve as alpha testers for the program. Tom

1 0

Re: voice encryption for portable PCs
by nobody＠jpunix.com 17 Dec '03

17 Dec '03

an109574(a)anon.penet.fi wrote: > I fully expect Phil's voice encryption software to become the de facto > telephonic standard just as PGP became the Internet standard. Cogon's > Aquafone, while commendable, has the disadvantage of employing DES. > I presume Phil will select a more resilient conventional algorithm. I don't see any reason to believe that triple DES is less secure than other currently popular algorithms such as IDEA, although 3DES does require somewhat more processor power.

1 0

Are 2048-bit pgp keys really secure ?
by danisch＠ira.uka.de 17 Dec '03

17 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- A 2048-bit pgp key ( n=p*q somewhere around 2^2048, p and q somewhere around 2^1024) is only as secure as it looks like, if both p and q are prime numbers. In fact p and q are only pseudo prime numbers, they are not proven to be prime numbers. It is known only that they have a high probability to be prime numbers. Usually a candidate number is send through a probabilistic prime test which says either "No, not a prime" or "a prime with a probability of at least 50% ". Usually this test is repeated 10 or 20 times, so after passing this iteration the probability of having a prime number is at least 1:2^10 or 1:2^20 . Would such a test be sufficient for generating 1024-bit prime numbers? Does it make sense to use pseudo-prime-numbers with a low probability of 1:2^10 only to generate a rsa key with a 2048 bit n ? Now have a look at pgp2.6.2: In genprime.c the prime numbers are generated. After testing the candidates with a table of small primes, they are passed to slowtest(). [Read slowtest and its comment...] slowtest() does not do one of the usual primality tests. It just passes the candidate through a Fermat test. Only four (4!) passes are done. The comment of slowtest() gives a probability of 10^-44 to fail for a number of about 512 bit. If this is true ( 10^-44 ~ 2^-146 ), about one of 10^44 keys is weak. This shouldn't be a problem, 10^44 is quite big. But at the moment I can't follow the arguments, why 4 Fermat tests should be enough to find good (pseudo-)primes. I can't see a reason why the iteration should already be stopped after the 4th loop. Generating a key should be worth to wait some minutes longer, especially when this doesn't need interactive work. I am also not convinced yet of the Fermat test. Why not use a Rabin-Miller-Test ? I have read only a very small piece of the pgp code yet, but if I understand the code of slowtest well (correct me if not...) the command mp_init(x, primetable[i]) for i=0,1,2,3 sets mpi x to the values 2,3,5,7 . If I understood this well, the slowtest() is nothing more than testing for a given p whether 2^(p-1) = 1 mod p 3^(p-1) = 1 mod p 5^(p-1) = 1 mod p 7^(p-1) = 1 mod p Any comments? BTW: The comment of slowtest() references "Finding Four Million Large Random Primes", by Ronald Rivest, in Advancess in Cryptology: Proceedings of Crypto '91. I have the "Advances in Cryptology - Crypto '91, Proceedings", Lecture Notes in Computer Science, 576, Springer, here. Call me blind or stupid, but I can't find the referenced Article. Neither the Title in the contents, nor R. Rivest in the Author Index. Can anybody tell me where to find the referenced Article ? Hadmut Danisch BTW 2: pgp2.6.2 doesn't work well if a key identified by its keyid is keychecked ( pgp -kc 0x... ). It stops after the first signature with a signators key shorter than the signed/checked key, because the global precision is changed and not changed back for testing the signature. -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQCVAwUBLwBtzGc1jG5vDiNxAQHi6wP/WS3afYhQ0ijJZfWbByjtvPrCZtCfDs1M 1p8Paqx0ZIIgCE2G6tY8JTlZ6tn5nEY4/qGHS3Q3TrO77HVheKq2bHMajGzSA3At CoX65ycg2Pn30q7PeLY89vtNosW568CqnmpPAmusD+o9CFO6RpFFZxIb5pgY5brF 8ll/F1ztdmM= =JZS6 -----END PGP SIGNATURE-----

2 1

swIPe pointer please
by cjl 17 Dec '03

17 Dec '03

I need a pointer to swIPe info and preferably an FTP site. Responses to me by E-mail please. C. J. Leonard ( / "DNA is groovy" \ / - Watson & Crick <cjl(a)welchlink.welch.jhu.edu> / \ <-- major groove ( \ Finger for public key \ ) Strong-arm for secret key / <-- minor groove Thumb-screws for pass-phrase / )

2 1