December 2003 - cypherpunks-legacy

GOPHER: AOL, CPSR, cypherpunks
by Karl Lui Barrus 17 Dec '03

17 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- Cypherpunks, Earlier, somebody (I beleive Carl Ellison, but I screwed up and deleted the file) asked about AOL and the gopher site. The problem may not be AOL or CPSR actually. What happened is this: Chael reorganized the gopher site, placing the Cypherpunks Gopher Archive into a subdirectory. So the archive is still at chaos.bsu.edu, but no longer immediately at the top level. Perhaps CPSR's pointer to the site is no longer valid! Or maybe the problem is that he also upgraded to a new gopher server, which is supposed to be backwards compatible ;) and one of these days I'll create some sort of index files (?) to take advantage of the new features, etc. So maybe this is the problem, a new version that AOL doesn't handle. So try gopher'ing directly to chaos.bsu.edu and then into the "Cypherpunks Gopher Archive" if going through somebody else isn't working. Hm... I'll inform CPSR about the change. Karl Barrus klbarrus(a)owlnet.rice.edu -----BEGIN PGP SIGNATURE----- Version: 2.6 iQCVAgUBLhnpG8SF/V8IjI8hAQGn0gP/YjRzP3OODE9e7pXdHhOpHtY6UddH75Fj 7z1T/RA9JuL3es926zNZP2K9C1p9AlgwKyPcea2GWkgT8q79Ink4VXKGJbwEzzwV Kyh5a3Efm7ZsewxlUnrkRtppH2qc//bhNkpaTSIlM1Ccl7+yBttONyq6qHFfpzLd m88J8APnLWM= =AbvY -----END PGP SIGNATURE-----

1 0

Re: Detweiler clone at WSJ
by Blanc Weber 17 Dec '03

17 Dec '03

From: Dave Hart <list of things to do to make the life of moral innocents safer> There! That should protect us from all the creeps in the world. Of course, there will be a few anarchist crybabies who will call these measures fascist, and a few victims of the new openess...(etc) .......................................................... Don't laugh - they'll probably agree these are really good ideas. The author might even commend you for being so thorough and stringing them together to make such a complete package. Blanc

1 0

Re: Password entropy
by wcs＠anchor.ho.att.com 17 Dec '03

17 Dec '03

"Nobody" asks whether you really get 128 bits of entropy out of MD5 if you put in fewer bits, and whether you need to put in more than 128 bits of entropy to get 128 bits of entropy out. (This is mainly relevant for the case where you iterate MD5 N times for large N.) Entropy = -sum ( p(Xi) * log2(p(Xi) ) , Xi { outcomes of a random event X } which is the sum of the amount of information each event gives you times the probability of the event occurring. In this application, the events are "the input to MD5 is" and "the output from MD5 is", and each input is one of many (presumably independent) values leading to the same output. You know that Entropy(MD5(Xi)) is <= 128, since there are only 2**128 possible outputs, and they're supposedly equiprobable given random input. If the distribution of the Xi's is known, and it has substantially lower entropy than 128 bits, then the output also has lower entropy, since the probability of MD5(Xi) appearing is the probability of Xi. There's a bit more entropy lost in the MD5 step - if MD5(Xi) = MD5(Xj), -p(Xi|Xj)*log(p(Xi|Xj) < -p(Xi)*log(p(Xi)) + -p(Xj)*log(p(Xj)). On the other hand, collisions are infrequent - the probability of a pair of numbers having the same MD5 value is presumed to be 2**-128, and the usual birthday paradox calculations apply, so you'll probably find one if you take 2**64 random samples. At this point, knowing the details of the MD5 algorithm *does* matter; you can analytically find a few pairs of inputs that have the same MD5 value - but if you're choosing random inputs it's not likely to happen. If you could analytically invert MD5 (it's presumed that you can't, even for the 128-bit-input case), or store the results in a 2**128 large lookup table (:-), you could find out exactly how much lossage there is. Don't worry about it :-) If you still *are* worried about it, however, you can scramble things a bit; since MD5 produces 128 bits of output but uses 448 bits of input+padding, you can add a different constant to the input at each step. If you're using it as a salt, put it at the beginning; if you're just doing it for multiple iterations it doesn't matter much. Bill Celebrate Independence Day the traditional way - overthrow a government!

3 2

MD5: hashing, > 1->1
by Karl Lui Barrus 17 Dec '03

17 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- >> is based upon the fact that *finding* two messages that hash to the >> same value is as difficult as a brute-force attack, which requires >> 2^128 trials (maybe it's 2^127, but I don't think that really > This is incorrect, with a large memory, this is the birthday paradox in > action, and it takes about 2^64 tries, which puts SHS right up there at > 2^80 same as skipjack. Geez, I did it again (deleted the original message - the one Derek sent). So from memory, I beleive that in the context in which Derek was describing the "finding two messages" above, his statement about the difficulty (2^128) is correct. The birthday paradox is the situation when you are looking for *any* two messages that hash to the same value. In this case, 2^64 is the expected work. However, if you are given a particular hash and you are looking for another message which has the same hash, then the difficulty is 2^128. This is the situation which is (more) important since it corresponds to forging MD5 hashes for a signed message. Say you are given a message and you want to find another which has the same hash. 2^128 applies. The birthday paradox situation corresponds to just finding two messages with the same hash. In this case the expected work is 2^64, but then the two messages that you discover with the same hash may be random (and thus worthless). Karl Barrus klbarrus(a)owlnet.rice.edu -----BEGIN PGP SIGNATURE----- Version: 2.6 iQCVAgUBLhnrj8SF/V8IjI8hAQGlmQP6AshYEwjoJGbN8cZZRiPAEdhZO9AAWG2Y P08YcQ/wUWNEAOAvi4WISPobIWxO6oRk+fBRvUMWv7wyU4eRA/7yj95nlDaui5oW rDaFrh+IBnC8Epce2hing6TqWdBxL5uKBCuq1CrKnUkDO2uESoZkN/aDpbnvueC9 05aqKfQ9P+U= =Lscb -----END PGP SIGNATURE-----

2 1

Re: BoardWatch & DigiCash.
by gtoal＠an-teallach.com 17 Dec '03

17 Dec '03

: > Ziff-Davis publications. Anyone care to suggest a magazine (or other form : > of trade rag) that gives unbiased information (or at least as close as : > possible)? Z-D lost my subscription to both PC Mag and PC Computing, for : * I like "Byte." Our own Peter Wayner has written good articles on : crypto. Good articles are few and far between. Byte has gone downhill in a *big* way, from the non-PC general computing/algorithm focus is had 10 years ago - when we would have the annual Languages edition, and programs you could actually do something new and useful with. (I wrote an image processing suite based on their 'Mona Lisa' article, and their one of finding multi-word anagrams turned into my scrabble program) Anyway all the blow-in cards suck. : * "Communications of the ACM" is also good. Maybe it got better recently, but last few years I was reading it, it too had turned into a comic. I was reasonable back around 76-80 but then it turned into a joke. I admit I haven't looked at it in the 90's. G

1 0

PC Expo Disk Available
by rarachel＠prism.poly.edu 17 Dec '03

17 Dec '03

Hey guys, I've just sent up the PC Expo disk package to ftp.wimsey.bc.ca incase anyone is interested in it. Give it some time as Mark has to clear it but it's in the incoming folder right now... I'll try to send it up to some other sites, but I guess I won't send it to csua.berkeley.edu as it isn't being taken care of... (if you know otherwise let me know and I'll send it there too...)

1 0

Re: Pass Phrases
by fasttech!zeke＠uu4.psi.com 17 Dec '03

17 Dec '03

We Americans have had many years of being forced to remember relatively high-entropy things. Roughly 23 or 33 bits. I refer, of course, to phone numbers. I, for one, have no problem remembering a few dozen of these, in various area codes, even though I don't dial some of them for years. Our neural cortexes have been "conditioned" to remember 7 digit (or 10 digit) numbers. Much more easily than, say, 8 or 9 digit numbers. So I claim that a simple way to get an easy 33 bits of entropy into your pass phrase would be to generate a "random" 10 digit number, but then remember it as an area code + phone number. You would, of course, combine this number with some other sources of entropy, such as words or phrases. If you really wanted to, you could forget the words or phrases and just remember four numbers of this sort. You would have your magical 128+ bits of entropy a lot more easily than remembering 40 random decimal digits.

1 0

AOL and CPSR gopher
by Carl Ellison 17 Dec '03

17 Dec '03

I tried out America OnLine last night (free disk in the mail) and saw that they had support for gopher -- so I poked around and found the CPSR site -- and from that, the pub/cypherpunks directory. However, there was no pgp directory and no code in the cypherpunks directory. Does anyone know if this is manipulation by AOL or some side effect of gopher? (I'm new to both AOL and gopher). - Carl

1 0

BoardWatch on digital cash
by Stanton McCandlish 17 Dec '03

17 Dec '03

See current (July 1994) _BoardWatch_, pp. 60-63. There's an article on an e-money scheme called NetCash. Unfortunately it is utterly stupid, but BW is giving it a semi-endorsement. Some of you d-c afficionados might like to disabuse them of some notions. Some of the flaws: 1) not cryptographically secure 2) someone can randomly guess the ser. # of your digicash and go spend it 3) non-anonymous 4) the person transferring the netbucks to you can actually spend it before you validate it with the central server (e.g. it would only be of use in cases where product/service has yet to be rendered, and customer gives you the netmoney, which you verify and only then serve them. Completely useless otherwise.) There are more, but those are the main ones. -- Stanton McCandlish * mech(a)eff.org * Electronic Frontier Found. OnlineActivist F O R M O R E I N F O, E - M A I L T O: I N F O @ E F F . O R G O P E N P L A T F O R M O N L I N E R I G H T S V I R T U A L C U L T U R E C R Y P T O

3 2

RE: (None)
by Mike Markley 17 Dec '03

17 Dec '03

---------- | From: Jeff Gostin <jgostin(a)eternal.pha.pa.us> | To: <cypherpunks(a)toad.com> | Subject: (None) | Date: Sunday, July 03, 1994 4:08PM | | bmorris(a)netcom.com (Bob MorrisG) writes: | | > I've heard rumors through the years that CIA/NSA/whoever can aim a | > parabolic antenna at your window, read the electronic pulses surrounding | > your computer, and thusly determine what you are typing. Is there any | > truth to this? | If this is a reference to the TEMPEST attack, yes, they can. From | what I hear, it's trivially easy for them to do, because they have The | Right Equipment. Rumor also suggests that that The Right Equipment isn't | so difficult to come by. I don't pretend to understand the intricacies of | TEMPEST, but I do know that the government requires all of their computers | (used at any level of security above none) to be TEMPEST shielded. This | tells us its a real concern. | | There are other ways to beat TEMPEST, like making it damn near | impossible to get close enough to the computer in question to get a | reading. | | --Jeff | -- | ====== ====== +----------------jgostin(a)eternal.pha.pa.us----------------+ | == == | The new, improved, environmentally safe, bigger, better,| | == == -= | faster, hypo-allergenic, AND politically correct .sig. | | ==== ====== | Now with a new fresh lemon scent! | | PGP Key Available +---------------------------------------------------------+ | How well does a Tempest attack work with multiple machines in the same room? It seems to me that it would take additional equipment to filter out the different machines that are being run. It also seems that it should be trivial to create a noise generator that would make Tempest useless. Kind of like using a spark plug to foil older radar guns. Mike. ===================================================== Mike Markley <mmarkley(a)microsoft.com> I'm not a Microsoft spokesperson. All opinions expressed here are mine. =====================================================

1 0