December 1992 - cypherpunks-legacy

Re: Destroying Data (Re: Remailer Policies)
by wcs＠anchor.ho.att.com 21 Dec '92

21 Dec '92

Yanek points out that "rm" doesn't destroy data on disks, just inodes, and that you have to overwrite the data to destroy it. However, he suggests that you do this by > 'cat /dev/null > /dev/hdxxx' which overwrites with zeroes all as well as overwriting swap space and powering off RAM. Won't work, and it's the hard way to do it: cat /dev/null produces zero bytes of output (there's a /dev/zero on some systems which produces lots of bytes of zeros.) If you're dealing with amateurs (i.e. local cops or maybe FBI, but not NSA/KGB, who can take the disk apart and use fancy magnetic techniques), you can get away with overwriting the data once with zeros or other stuff; a simple approach is to create a file large enough to fill all the unallocated space on the disk (either using /dev/zero or writing your own); be sure to be root so you get the last 10% of a BSD file system. For professionals, overwrite it once with 0s, once with 1s, and once with some test pattern like Xs or /etc/termcap. This still won't help if there are bad blocks which the disk controller is mapping out transparently to your system, of course, but amateurs can't read them. Swap space is tougher - you'll need to look at how your kernel allocates it to know how to fill it all up, or else zap it all after rebooting your system, if this is practical. The government's rules for handling, um, special data say that you either need an officially approved software program, or an Official Big Magnet, or else you need to physically destroy the magnetic media from the disk. I don't like having Big Magnets near my lab, but sandblasting has a real physicality to it, and floppy disks make this satisfying scrunch in a paper-shredder :-) Needless to say, unless you're Ollie North, you shouldn't be shredding your data when the cops are busting you ... Bill Stewart, somewhere in New Jersey

1 0

DES source for the HP48SX calculator.
by Bill Sommerfeld 21 Dec '92

21 Dec '92

I have implemented a version of the core of the DES algorithm for the HP 48SX calculator; I mentioned this in passing on this list a couple of weeks ago. Anyhow, in the general cypherpunks spirit of things, I'm willing to share this code... Note that this only ECB-mode encryption and key schedule computation, and not decryption. Both operations take about 12 seconds.. mostly because they're written in user RPL rather than system RPL or machine code. If the person at Berkeley who maintains the cypherpunks FTP archive wants to get in touch with me, I'd be happy to contribute this file to the archive; also, I'll mail it to anyone with a US or Canadian email return address. - Bill

1 0

Avoiding snags in the Law : Remailers.
by DrZaphod 20 Dec '92

20 Dec '92

In Message 19 Dec 92 19:45:33 EST, Hal <CompuServe.COM!74076.1041(a)netcomsv.netcom.com> writes: >I suspect that instead the approach would be to claim that remailer >operators are responsible for the material their remailers produce, >regardless of its original source. >Such a policy would be a plausible extension of current Internet >policies, IMO. RFC 822, the document which describes the format of >Internet mail message, in session 4.4.2 discusses the "Sender:" field, >and says, "Since the critical function served by the 'Sender' field >is identification of the agent responsible for sending mail and >since computer programs cannot be held accountable for their behavior, >it is strongly recommended that when a computer program generates a >message, the HUMAN who is responsible for that program be referenced >as part of the 'Sender' field mailbox specification." [Capitalization >in the original.] The need for a person to take responsibility for >each piece of mail that is sent would tend to lead to the policy >I mentioned. > Perhaps the solution to this is to run encrypted remailers, anonymously. The remailer could either run on it's own, or under instructions by an anonymous source [i.e PGPed instructions and software updates signed by the owner who has created a second key for his identity of remailer operator [Let's call him Oz]. Example: Joe creates an account on a random system, installing the nessicary software to run a PGP remailer. This software would be modified to accept instructions from a specific anonymous identity, for which Joe creates one, leaving the public key with the remailer. The remailer also has it's own key pair to decrypt and forward messages [of course]. People could now route their messages through Joe's remailer to anywhere they wish. The remailer should be set up to NOT keeps logs, except errors. Now Joe decides he wants to update his remailer with the latest scripts. He mails te scripts [thru various remailers], encrpyted with the remailers key and signed by Oz's key. The remailer decrypts the mail, verifies that it belongs to Oz, and updates itself. In this way no person can be held responsible for mail passing thru it's remailer. The only way to stop such a remailer would be to downright shut it down, or to communicate to Oz via the remailer [the scripts would allow Oz to pick up any mail left directly to the remailer withour a forwarding header -- of course it would reroute the messages, encrypted, thru various other well-known remailers]. TTFN! DrZaphod [AC/DC] / [DnA][HP] [drzaphod(a)ncselxsi.uucp] Technicolorized

1 0

Re: Destroying Data
by bart＠netcom.com 20 Dec '92

20 Dec '92

In audio recording which uses magnetic media similar to disks the only way to thoroughly erase a tape is with a so-called bulk eraser. This applies a very strong a.c. magnetic field and then is gradually withdrawn from the tape to allow the residual magnetism to "follow the hysteresis curve" down to zero. This would be difficult to apply to disks except floppies. Consider also the technique used in the Norton Utilities program "Diskwipe" which takes a /g switch which "Follows certain government rules for wiping". I can't find the manual but I think it specifies how many repetitions are used and different values to write in each.

1 0

Remailers and liability
by miron＠extropia.wimsey.com 20 Dec '92

20 Dec '92

Here is a post which is somewhat relevant: >Newsgroups: alt.censorship,news.admin.policy,comp.org.eff.talk >From: nagle(a)netcom.com (John Nagle) >Subject: Appeals court declares child pornography law unconstitutional >Message-ID: <1992Dec17.232618.19727(a)netcom.com> Summary: New decision by Ninth Circuit may affect BBS operators >Keywords: law pornography >Organization: Netcom - Online Communication Services (408 241-9760 guest) >Date: Thu, 17 Dec 1992 23:26:18 GMT The Ninth U.S. Circuit Court of Appeals ruled, in US vs X-Citement Video Inc., (CA No. 89-50556), that the sections of the Protection of Children Against Sexual Exploitation Act of 1977 that deal with distribution, transportation, and receipt of sexually explicit materials are invalid on First Amendment grounds. The court let stand the prohibition on the production of child pornography. This ruling needs to be looked at more closely, but it may help to remove sysop worries about someone posting such material on their system without their knowledge, or about network nodes being liable for material passing through them. This is a step forward, because it was one of the very few real legal risks a US sysop had to face, and now it appears to be dead. Do the lawyers out there agree with this analysis? Producing child pornography remains illegal, which is reasonable. John Nagle Source: Wall Street Journal, 17DEC92, p. B10 (Western edition)

1 0

Remailers.
by Hal 20 Dec '92

20 Dec '92

-----BEGIN PGP SIGNED MESSAGE----- Responding to Edgar Swank's message: > We've already discussed use of PGP by pedophiles to encrypt > incriminating diaries. It gets better. PGP and remailers can > be used to encrypt and ship "kiddy porn" GIF files. Even private > possession of such material is a serious crime in the USA. A chain > of anonymous and encrypted remailers could be used to drop kiddy porn > into newsgroups like > > alt.sex.pictures > > I'm sure that would create quite a stir! This could easily happen soon, as news of these remailers spreads. There's been talk on the net about a student at an Ivy League college who is being investigated by the FBI for allegedly posting illegal images containing child pornagraphy. Once people find out about anonymous remailers, especially ours with their chaining capabilities and encryption, they will realize that such posts can be made with almost total safety. Presumably this may increase the number of people who would attempt it. > Perhaps Duncan or another attorney would like to comment on legal > liability of remailers in these kinds of situations. I don't think > there's currently any law on what records remailers would have to > keep; if you didn't have them, you couldn't give them up. > > If this becomes a problem, expect legislation regulating or even > outright banning remailers. However, if remailer code has spread > over many national boundaries, such legislation may not be effective. My guess is that it will not be feasible to ban remailers, since it would be hard to draw the line between completely automated remailers, and simple manual forwarding of a message that someone sent you, which happens all the time and can hardly be made illegal. I suspect that instead the approach would be to claim that remailer operators are responsible for the material their remailers produce, regardless of its original source. So if child porn comes out, I am guilty of sending child porn. I can argue that my remailer was automatic and that I shouldn't be held responsible for what comes out of it, but my guess is that this argument will be rejected on the grounds of personal responsibility, and because no one forces me to run a remailer which sends out anything that comes in to it. Such a policy would be a plausible extension of current Internet policies, IMO. RFC 822, the document which describes the format of Internet mail message, in session 4.4.2 discusses the "Sender:" field, and says, "Since the critical function served by the 'Sender' field is identification of the agent responsible for sending mail and since computer programs cannot be held accountable for their behavior, it is strongly recommended that when a computer program generates a message, the HUMAN who is responsible for that program be referenced as part of the 'Sender' field mailbox specification." [Capitalization in the original.] The need for a person to take responsibility for each piece of mail that is sent would tend to lead to the policy I mentioned. With such a policy in place, if enforced by law, I don't think people would run remailers in this country because of the legal risks. There would still be the international remailers, though. > By the way, no-one has yet responded to my request to add features > to Eric's remailer to 1)remove the "sig" after "--" and 2)Anonymously > post to newsgroups available at the remailer site. Both these > features are present in the Australian remailer at Pax. Unfortunately > that remailer seems less secure than Eric's in that it keeps more or > less permanent records of its users at the remailer site and it > doesn't seem to be able to chain to other remailers. > > -- > edgar(a)spectrx.saigon.com (Edgar W. Swank) > SPECTROX SYSTEMS +1.408.252.1005 Silicon Valley, Ca It should be possible to chain from Pax to our remailers, getting the best features of both. Pax could be the first remailer in the chain, stripping the sig. The message could then go through our remailers, providing non-traced security, at least if you can find someone who does not keep logs. Anonymous posting can be done already. Anyone can post to most newsgroups by sending mail to certain addresses. One such system is at Berkeley. Send to, e.g., sci-crypt(a)ucbvax.berkeley.edu to post to sci.crypt. There is another such system running in Austin, TX, I believe, but I don't have the corresponding net address handy. So this capability is already provided by our (or anyone's) remailers. One reason I haven't looked at stripping the .sig is due to an email discussion I had about a month ago with Eric Hughes. Eric strongly felt that message bodies should be preserved as much as possible by the remailers, in accordance with the general principle of Internet mail forwarding. Too much mangling is done already by mail gateways, and adding more changes might be harmful. Obviously, the remailers can't avoid doing some processing of the message body, what with the "::" pasting tokens and PGP decryption, but Eric felt that unnecessary changes should be minimized. I know Eric is working on some new remailer concepts so I want to defer to him on this issue for now. Hal Finney 74076.1041(a)compuserve.com -----BEGIN PGP SIGNATURE----- Version: 2.1 iQCVAgUBKzOXf6gTA69YIUw3AQEvQQQArHXSwwrkQ3mSGOD60ZBV/p1R9RONqv2x S64iJjdJB43G2nxLHB2t9xXAKSdCT00shEtil8LWu20XW0rWXqmMXyYudtt3qf3t 2arqlvxdGXlPt3eISFJ97U6jaI1EhswPg29fjuXMMaKv5OO/gBKp1i9Cig7suZDt EkUDJihLQxc= =q6zH -----END PGP SIGNATURE----- Distribution: CYPHERPUNKS >INTERNET:CYPHERPUNKS@TOAD.COM

1 0

TEMPEST
by gnu 19 Dec '92

19 Dec '92

> There are plenty of other government conspiracies to focus on ! This one > seems a bit lame. Our government should not be our opponent when we try to find out how to provide high levels of privacy for ourselves, our computers, and our communications. Ignoring that, yes, TEMPEST doesn't look like a deep dark conspiracy. It's legal to shield your equipment and probably legal to export it (if you didn't use the classified TEMPEST specifications in building it, but reverse engineered them yourself). The government is not our enemy here (like it *is* our enemy in the drug war), but it isn't our friend either. John Gilmore

1 0

Re: positive rep
by yanek＠novavax.nova.edu 19 Dec '92

19 Dec '92

> but how do you know who to add to the positive rep list? take netnews, > e.g. i generally scan a group for names i know (+ rep), then read For example, all mail/news reader software would have a function where after reading the article you could press a key to indicate you found the article useful/informative/interesting. What it would do is send the person a "certificate", which in essence says "I recommend reading messages by this person", signed cryptographically with your signature. These certificates could be appended to messages by that person, or posted to some special newsgroup just for that purpose (similar to "control"), or distributed through some other means. If several people to whom you have assigned high credibility recommend messages by a person, then your news software would automatically sort his messages to appear higher in the list before other people's messages, and also that person's recommendations would have a higher weight in computing other people's sort values. All this pre-supposes widespread use of authenticated messages, so someone could not forge himself some certificates from well-known people. Also, it would not allow a J. Random to forge a post that appears to be from a person with high reputation. The technology for such communication exists now (PGP, MD5, RSAREF, etc) it is just not integrated into most news/mail software. Unlike negative reputations, positive reputations are immune to subversion by creating many anonymous or pseudonymous identities. You might imagine someone creating two hundred "people" and have them all send him their "certificates". But this would not work, since no-one has recommended these people, their certificate would carry no weight whatsoever. -- Yanek Martinson mthvax.cs.miami.edu!safe0!yanek uunet!medexam!yanek this address preferred -->> yanek(a)novavax.nova.edu <<-- this address preferred Phone (305) 765-6300 daytime FAX: (305) 765-6708 1321 N 65 Way/Hollywood (305) 963-1931 evenings (305) 981-9812 Florida, 33024-5819

1 0

Remailer Policies
by edgar＠spectrx.Saigon.COM 19 Dec '92

19 Dec '92

I would like to expand on Hal's suggestion that remailer operators post their policies on a)keeping logs and b) divulging same. Hal says he keeps logs "for a few days" and might divulge logs of "really vicious, racially or sexually harrassing" messages. Presumably this would also cover anonymous threatening messages, blackmail, etc. We've already discussed use of PGP by pedophiles to encrypt incriminating diaries. It gets better. PGP and remailers can be used to encrypt and ship "kiddy porn" GIF files. Even private possession of such material is a serious crime in the USA. A chain of anonymous and encrypted remailers could be used to drop kiddy porn into newsgroups like alt.sex.pictures I'm sure that would create quite a stir! Perhaps Duncan or another attorney would like to comment on legal liability of remailers in these kinds of situations. I don't think there's currently any law on what records remailers would have to keep; if you didn't have them, you couldn't give them up. If this becomes a problem, expect legislation regulating or even outright banning remailers. However, if remailer code has spread over many national boundaries, such legislation may not be effective. By the way, no-one has yet responded to my request to add features to Eric's remailer to 1)remove the "sig" after "--" and 2)Anonymously post to newsgroups available at the remailer site. Both these features are present in the Australian remailer at Pax. Unfortunately that remailer seems less secure than Eric's in that it keeps more or less permanent records of its users at the remailer site and it doesn't seem to be able to chain to other remailers. -- edgar(a)spectrx.saigon.com (Edgar W. Swank) SPECTROX SYSTEMS +1.408.252.1005 Silicon Valley, Ca

1 0

positive reps and paranoia
by deboni＠diego.llnl.gov 19 Dec '92

19 Dec '92

I'm just a lurker on this list, trying to pick up on what's happening in this subset of computing and communications, but my chain's been yanked, and the subject merits a reply. On the matter of the discussion that's been going on vis-a-vis reputations versus kill files, I'm afraid we're regressing to the bad old days when everyone was considered bad and worthy of suspicion until they demonstrated that they were good and trustworthy. I'd personally rather believe people are basically good than otherwise. Even if I must occasionally suffer getting burned, it's easier on the nerves, attitude, and karma to assume the best in those I interact with. I think it's significant that there are really so few of us on the net who are actually insufferable and refuse to be shouted down to reasonable behavior by the civil rest of us. Those few who are will not be prevented from troubling us by the measures being advocated - positive reps, scores on 1-to-10 scales, etc. - any more than weapon makers are deterred by manufacturers of armor. someone who really wanted to could still flood our group with vitriol, using multiple artificial identities vouched for by other artificial identities. If such neurotic vengeful behavior were really likely on the net, we'd have seen it already. What, other than good sense and a low threshold of boredom, prevents any of us from flooding any and all news groups with garbage? And if it ever becomes a problem, we'll just have to appoint a moderator, perhaps on a rotating basis, from among those of us who are personally acquainted with each other. My point here isn't that we shouldn't prepare for the worst, but that we shouldn't get crazy about it. The theoretical aspects of the discussion are interesting to me, but I just thought it was getting a little close to the edge not to comment. Tom DeBoni (Once I figure this stuff all out, I'll get a "protected" identity, too.) deboni(a)llnl.gov

3 2