June 1993 - cypherpunks-legacy

Crypto API (was: Software infrastructure)
by gdale＠apple.com 05 Jun '93

05 Jun '93

I was thinking about a crypto project I have on the backburner (actually it's just in the concept stage), a program to hide encrypted data in raster image files. I really don't want to muck with writing or porting an encryption package, I just want to do the part that sticks the data into the image. HERE'S THE BEEF: I think a lot of the various mail readers and what not would be likely to encorporate encryption, if only it were mind-bogglingly simple. An API (Application Programming Interface) could be proposed by cypherpunks or others. The same folks should probably provide at least one encryption library for use. It would be nice if this API allowed for multiple encryption schemes (user selectable), like the way Apple's Communications Toolbox allows users to switch between various connection protocols. I'm much too busy to embark on something like this (largely because I'm not a cypher expert), but if I can help convince somebody to do it, or to provide feedback on the interface, I'm available. ________________________________________________________________________ Geoff Dale -- insert standard disclaimers here -- gdale(a)apple.com "Mind your nerve ends, love bunch" -- Dr. Caligari

1 0

The "WARLOCK" Cipher
by tcmay＠netcom.com 05 Jun '93

05 Jun '93

Cypherdroids, Earlier today I commented on the advisability of posting new ciphers to this group, or to sci.crypt. (I haven't gotten my message yet, so I don't know if it went through.) By coincidence, a very long posting on sci.crypt has appeared, announcing a new matrix-based crypto system, called "WARLOCK" by its inventors. They provide an extensive introduction to the mathematics used and offer various analyses. This is a first step toward making their system worth analyzing. I suggest following the debate on this system will be educational. And perhaps their system will even survive to become a reasonable alternative. -Tim May -- .......................................................................... Timothy C. May | Crypto Anarchy: encryption, digital money, tcmay(a)netcom.com | anonymous networks, digital pseudonyms, zero 408-688-5409 | knowledge, reputations, information markets, W.A.S.T.E.: Aptos, CA | black markets, collapse of governments. Higher Power: 2^756839 | Public Key: PGP and MailSafe available. Note: I put time and money into writing this posting. I hope you enjoy it.

2 1

Work the work!
by fergp＠sytex.com 05 Jun '93

05 Jun '93

-----BEGIN PGP SIGNED MESSAGE----- On Tue, 1 Jun 93 09:56:45 -0700, Eric Hughes <uunet!soda.berkeley.edu!hughes> wrote - > We are trying to build a sandbox, and the government is trying to > restrict the use of sand. We are indeed doing just this. Although the small minority of us compu-professionals are writing code, flailing congressmen, etc., it takes more than what is currently being acknowledged to get things changed. I will not, however, discontinue my diatribe with my elected representatives on the topic of our electronic rights to privacy under the first and fourth amendments. (I'm having too much fun doing something that I seriously, and perhaps foolishly, believe in.) Is there something going on with the EFF that we should know about? > That said, I also urge those who are writing code to continue. To > those of you not writing code, however, I say start talking to your > friends and neighbors and communities and newspapers. > Now. We are working on it. A vote of confidence towards crypto-freedom. Are we east-coast-niks welcome in this process? Is policy being drawn by a few EFF persona without consultation of the masses? Eric, before you say "Now,", you'd best detail us. I know what you mean, however, many of the crypto-warriors which may follow do not. It may be a good idea to _now_ place a broad policy statement. Cheers. -----BEGIN PGP SIGNATURE----- Version: 2.2 iQCVAgUBLAwNs5RLcZSdHMBNAQFlRAP7BSpktDz4URB0rhWQ5mxb2UcJqEZHdp+2 It+Whxh1MzYTLFi0SfvZRQYjPEZO1wN2ac8bQyl2zOpi7viAg8X+AfEZACWooqUQ y8Dyddup15MNj/p53fJQhzKYaX4K4xD2h6WTWO1X8Q2SPHo0WV48Hu+uO8nyeoqD PJj0d/IHvg4= =6GvE -----END PGP SIGNATURE----- Paul Ferguson | The future is now. Network Integrator | History will tell the tale; Centreville, Virginia USA | We must endure and struggle fergp(a)sytex.com | to shape it. Stop the Wiretap (Clipper/Capstone) Chip.

4 4

HELP: pgp, .forward, mh
by Karl Barrus 05 Jun '93

05 Jun '93

Situation: I'd like mail that arrives at my other account to be encrypted and then forwarded to this account. I have spent a few hours trying various things and nothing seems to work. I've tried this as my .forward file: "| /myhome/pgp -fea barrus | mail elee9sf(a)menudo.uh.edu" but all that arrives at elee9sf is a blank message. I've tried "| /myhome/remail.script" where remail.script was a one-liner similar to the above. Nothing. Any suggestions? Note: 1) barrus(a)tree.egr.uh.edu is a NeXT account. Is NeXTSTEP just too different for this stuff to work? 2) I'm running pgp2.1 on tree.egr, and don't have the old docs anymore. Was the -f option not present? 3) Anybody have a makefile for NeXT so I can upgrade? I compiled 2.1 by hand editing some lines, but the task looks pretty daunting with 2.2. I tried 'make mach' but that didn't get very far. So, if that's impossible, how do I get my elee9sf account to do it? I use mh on menudo, and have tried to barrus(a)tree.egr.uh.edu | A "/path/pgp -fea barrus | /path/rvcstore +tonext" in my .maildelivery but that doesn't seem to fly either. Any suggestions? /-----------------------------------\ | Karl L. Barrus | | elee9sf(a)menudo.uh.edu | <- preferred address | barrus(a)tree.egr.uh.edu (NeXTMail) | \-----------------------------------/

2 1

RE: Term software develo
by Doug Porter 05 Jun '93

05 Jun '93

Eric Hughes says: > FOSSIL is . . . the only abstraction for serial communications the PC has . . . There are at least 7 versions of INT 14. Then there's Ungerman-Bass's old INT 6B as well as two flavors of its successor, NASI/NCSI. There are also many variants of COMx device drivers. These are all serial line abstractions for the PC. There are others. FOSSIL may be popular on BBSs, but NASI/NCSI is number one in the market. The oldest version of INT 14 is number two. Who was it said "The nice thing about standards is that there are so many to choose from"? :-) Doug

3 2

Re: Term software dev
by Doug Porter 04 Jun '93

04 Jun '93

Eric Hughes asks: > What are NASI/NCSI? Does it cost to use them? Is source available? It's an API originally developed by Network Products Corporation and based on an earlier spec from Ungerman-Bass. NPC used it in building an async comm server for LANs. NPC called the API "Network Communications Server Interface" or NCSI. Novell licensed the technology from NPC and renamed the API to "Novell Asynchronous Server Interface" or NASI. Other async server vendors picked up on it about then as a result of Novell's evangelism. Serial communications software packages from Crosstalk to Procomm to tiny niche products started to support it. Then because of the end user package support, the standard was used for different kinds of serial connections. For example, we (CyberCorp) built a NASI/NCSI interface for intelligent Digiboards a while back. I've never heard of anyone being charged to use the spec for whatever they like. At least part of the spec, the part promoted by Ungerman-Bass, seems to be in the public domain. We originally got the spec from Novell when we built a Netware compatible async server. I don't know of any free source code, but code for either end of the spec is only one or two thousand lines of C. Doug

1 0

Re: VinCrypt
by Clark Reynard 04 Jun '93

04 Jun '93

Even as a former 'hacker' myself, the second to last person I would trust not to install a backdoor (next to the NSA) is a hacker. In addition, merely having been a systems hacker hardly qualifies one for writing complex crypto software. Without any assurance as to the authors' qualifications for writing a crypto package, or their integrity. Even if I could trust their integrity, I'm very leery of black-box software. ---- Robert W. Clark Just Say No! to the rclark(a)nyx.cs.du.edu Big Brother Chip

8 9

Term software, disk driver encryption
by greg＠ideath.goldenbear.com 04 Jun '93

04 Jun '93

Am suprised no-one's mentioned DISKREET, the encrypted disk driver software included with the Norton Utilities. It does DES, though the manual doesn't mention which flavor of DES it uses. I have been using it for roughly 2 years now without any trouble. It's been well-behaved and is probably already on lots of folks' disks, by virtue of being included with the rest of the Norton stuff. I think "TPU" refers to Turbo Pascal Units. "Async Pro" is, if I remember right, the name of an add-on async communications library for Turbo Pascal. Telix is not free, but is freely distributable, as it's shareware. -- Greg Broiles greg(a)goldenbear.com Golden Bear Computer Consulting +1 503 465 0325 Box 12005 Eugene OR 97440 BBS: +1 503 687 7764

1 0

subscribe
by bei＠dogface.austin.tx.us 04 Jun '93

04 Jun '93

subscribe l-cpunks(a)dogface.austin.tx.us

1 0

DOS disk encryptor
by Eric Blossom 04 Jun '93

04 Jun '93

Ryan, Good luck on building the DOS disk encryptor. I belive that what you need to do is write a standard DOS disk driver (that can be installed in CONFIG.SYS) that implements the READ and WRITE primitives. I belive that they use the same entry point (the STRATEGY entry) in the driver. You would basically just call the BIOS routines to do the actual i/o. You don't have to worry about the FAT etc, just encrypt everything. You will probably want to use DES or IDEA and run it in CBC mode or Counter Mode. You would use the DISK BLOCK NUMBER as a piece of the key material (or part of the Initialization Vector), hence, even if the same data appeared multiple places on the drive, it would appear different on the surface. There is a good description of operation modes in "Modern Cryptology: a Tutorial" by Gilles Brassard (Springer Verlag Lecture Notes in Computer Science #325, 1988). Denning's book covers it too. I'd probably start out getting it running on a floppy. After that, just use a separate partition to make life easier. The driver is handed physical (or logical) block numbers, and these map directly to the physical drive block number by adding the offset of the beginning of partition. At driver init time, you read the partition table on the hard disk, looking for a "system type" that identifies the partition as one of your encrypted ones. Prompt for the pass phrase, and store it in the driver. I assume that your concern is somebody physically grabbing the disk drive. I don't have a problem with the pass phrase in memory, as long I have physical control of the system. In some of the DOS references, there used to be a sample RAM DISK device driver. You could use it as the skeleton to get the entry points right, and then just encrypt the block and call the BIOS to do the i/o. Have fun, Eric Blossom

1 0