December 2003 - cypherpunks-legacy

Re: PGP 2.6 uses RSAREF version 1
by sommerfeld＠localhost.medford.ma.us 17 Dec '03

17 Dec '03

Has anyone been able to FTP a copy of RSAREF from rsa.com recently? I get "permission denied" when I attempt to cd into the directory named by the README file.. - Bill

1 0

Patches to make PGP2.3a compatible with 2.6
by Alan Barrett 17 Dec '03

17 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- Here's a set of patches relative to PGP 2.3a to make it do the following: * Display and accept hexadecimal key IDs with 8 digits. * Accept input "packets" with version 2 or 3. * Produce output "packets" with either version 2 or 3, under control of the new "version_byte" variable in the config file or on the command line. * Produce ASCII armour (or "armor" for USAans) with an desired version string, under control of the new "armor_version" variable in the config file or on the command line. This stuff needs testing and documentation. No warranty, etc. Please send comments to me. If it works the way I hope, then pgp +armor_version=2.6 +version+byte=2 should be compatible with MIT PGP 2.6 before September, and pgp +armor_version=2.6 +version+byte=2 should be compatible with MIT PGP 2.6 after September. Enjoy, - --apb (Alan Barrett) diff -u3 -r pgp/src/armor.c pgp-apb/src/armor.c - --- pgp/src/armor.c Sat Jul 3 00:32:38 1993 +++ pgp-apb/src/armor.c Wed May 25 17:38:32 1994 @@ -29,6 +29,8 @@ #include "crypto.h" #include "armor.h" +char armor_version[20] = ""; /* version text in armor output */ + static int dpem_file(char *infile, char *outfile); static crcword crchware(byte ch, crcword poly, crcword accum); static int pem_file(char *infilename, char *outfilename, char *clearfilename); @@ -508,7 +510,8 @@ else fprintf (outFile, "-----BEGIN PGP MESSAGE, PART %02d/%02d-----\n", 1, noSections); - - fprintf (outFile, "Version: %s\n",rel_version); + fprintf (outFile, "Version: %s\n", (armor_version[0] != '\0' ? + armor_version : rel_version)); fprintf (outFile, "\n"); init_crc(); diff -u3 -r pgp/src/config.c pgp-apb/src/config.c - --- pgp/src/config.c Mon Jun 14 02:44:57 1993 +++ pgp-apb/src/config.c Wed May 25 18:00:33 1994 @@ -84,7 +84,7 @@ MYNAME, TEXTMODE, TMP, TZFIX, VERBOSE, BAKRING, ARMORLINES, COMPLETES_NEEDED, MARGINALS_NEEDED, PAGER, CERT_DEPTH, CHARSET, CLEAR, SELF_ENCRYPT, - - INTERACTIVE, PKCS_COMPAT, + INTERACTIVE, PKCS_COMPAT, ARMOR_VERSION, VERSION_BYTE, /* options below this line can only be used as command line * "long" options */ #define CONFIG_INTRINSICS BATCHMODE @@ -96,7 +96,7 @@ "MYNAME", "TEXTMODE", "TMP", "TZFIX", "VERBOSE", "BAKRING", "ARMORLINES", "COMPLETES_NEEDED", "MARGINALS_NEEDED", "PAGER", "CERT_DEPTH", "CHARSET", "CLEARSIG", "ENCRYPTTOSELF", - - "INTERACTIVE", "PKCS_COMPAT", + "INTERACTIVE", "PKCS_COMPAT", "ARMOR_VERSION", "VERSION_BYTE", /* command line only */ "BATCHMODE", "FORCE", }; @@ -106,7 +106,7 @@ STRING, BOOL, STRING, NUMERIC, NUMERIC, STRING, NUMERIC, NUMERIC, NUMERIC, STRING, NUMERIC, STRING, BOOL, BOOL, - - BOOL, NUMERIC, + BOOL, NUMERIC, STRING, NUMERIC, /* command line only */ BOOL, BOOL, }; @@ -392,6 +392,20 @@ case INTERACTIVE: interactive_add = flag; + break; + + case ARMOR_VERSION: + strncpy(armor_version, str, + sizeof(armor_version)); + armor_version[sizeof(armor_version)-1] = '\0'; + break; + + case VERSION_BYTE: + version_byte = value; + if (version_byte < VERSION_BYTE_MIN) + version_byte = VERSION_BYTE_MIN; + if (version_byte > VERSION_BYTE_MAX) + version_byte = VERSION_BYTE_MAX; break; case BATCHMODE: batchmode = flag; break; diff -u3 -r pgp/src/crypto.c pgp-apb/src/crypto.c - --- pgp/src/crypto.c Fri Jul 2 23:55:07 1993 +++ pgp-apb/src/crypto.c Wed May 25 17:53:53 1994 @@ -59,6 +59,7 @@ #define USE_LITERAL2 +int version_byte = VERSION_BYTE_DEFAULT; /* PGP packet format version */ /* This variable stores the md5 hash of the current file, if it is available. It is used in open_strong_pseudorandom. */ @@ -313,7 +314,15 @@ */ int version_error(int val, int checkval) - -{ if (val != checkval) +{ return version_range_error(val, checkval, checkval); +} + +/* Return nonzero if val isn't in range from minval to maxval, after + * printing a warning. + */ +int +version_range_error(int val, int minval, int maxval) +{ if (val < minval || val > maxval) { fprintf (pgpout, PSTR( "\n\007Unsupported packet format - you need a newer version of PGP for this file.\n")); return(1); @@ -786,7 +795,7 @@ put_word16((word16) ske_length, certificate+certificate_length); certificate_length+=2; /* advance past word */ - - certificate[certificate_length++] = VERSION_BYTE; + certificate[certificate_length++] = version_byte; /* Begin fields that are included in MD calculation... */ @@ -1367,7 +1376,7 @@ goto badcert; /* complain and return bad status */ version = *certificate++; - - if (version_error(version, VERSION_BYTE)) + if (version_range_error(version, VERSION_BYTE_MIN, VERSION_BYTE_MAX)) goto err1; mdlensave = mdlen = *certificate++; /* length of material to be added to MD */ @@ -1807,7 +1816,7 @@ goto badcert2; /* complain and return bad status */ version = *certificate++; - - if (version_error(version, VERSION_BYTE)) + if (version_range_error(version, VERSION_BYTE_MIN, VERSION_BYTE_MAX)) goto err2; mdlensave = mdlen = *certificate++; /* length of material to be added to MD */ @@ -2361,7 +2370,7 @@ FALSE); /* Write version byte */ - - ver = VERSION_BYTE; + ver = version_byte; fwrite (&ver, 1, 1, g); writekeyID( n, g ); @@ -2745,7 +2754,8 @@ /* Read and check version */ fread (&ver, 1, 1, f); - - if (version_error(ver, VERSION_BYTE)) + if (version_range_error(ver, VERSION_BYTE_MIN, + VERSION_BYTE_MAX)) { fclose (f); return (-1); } diff -u3 -r pgp/src/crypto.h pgp-apb/src/crypto.h - --- pgp/src/crypto.h Mon May 10 01:38:27 1993 +++ pgp-apb/src/crypto.h Wed May 25 17:11:46 1994 @@ -75,6 +75,9 @@ /* Print an error message and return nonzero if val != checkval */ int version_error (int val, int checkval); +/* Print an error message and return nonzero if val not in [minval..maxval] */ +int version_range_error (int val, int minval, int maxval); + int check_key_sig(FILE *fkey, long fpkey, int keypktlen, char *keyuserid, FILE *fsig, long fpsig, char *keyfile, char *siguserid, byte *xtimestamp, byte *sigclass); diff -u3 -r pgp/src/keymgmt.c pgp-apb/src/keymgmt.c - --- pgp/src/keymgmt.c Wed Jun 23 22:53:53 1993 +++ pgp-apb/src/keymgmt.c Wed May 25 18:06:31 1994 @@ -137,7 +137,7 @@ char *bufptr; /* ptr to Key ID string */ static char keyIDbuf[2*KEYFRAGSIZE+1]; - - /* only show bottom 3 bytes of keyID */ + /* only show bottom 4 bytes of keyID */ bufptr = keyIDbuf; @@ -156,7 +156,7 @@ ** MSB-first keyID format */ - - for (i = KEYFRAGSIZE-3; i < KEYFRAGSIZE; i++) + for (i = KEYFRAGSIZE-4; i < KEYFRAGSIZE; i++) { sprintf( bufptr, "%02X", keyID[i] ); bufptr += 2; @@ -302,7 +302,7 @@ fwrite(&ctb,1,1,f); /* write key certificate header byte */ convert(cert_length); /* convert to external byteorder */ fwrite(&cert_length,1,sizeof(cert_length),f); - - version = VERSION_BYTE; + version = version_byte; fwrite(&version,1,1,f); /* set version number */ convert_byteorder(timestamp,4); /* convert to external form */ fwrite(timestamp,1,4,f); /* write certificate timestamp */ @@ -467,7 +467,8 @@ else if (is_ctb_type (ctb, CTB_SKE_TYPE)) { if (sigkeyID) { fread(&version,1,1,f); /* Read version of sig packet */ - - if (version_error(version, VERSION_BYTE)) + if (version_range_error(version, VERSION_BYTE_MIN, + VERSION_BYTE_MAX)) SKIP_RETURN(-6); /* Need a later version */ /* Skip timestamp, validity period, and type byte */ fread(&mdlen, 1, 1, f); @@ -493,7 +494,7 @@ if (n != NULL) set_precision(MAX_UNIT_PRECISION); /* safest opening assumption */ fread(&version,1,1,f); /* read and check version */ - - if (version_error(version, VERSION_BYTE)) + if (version_range_error(version, VERSION_BYTE_MIN, VERSION_BYTE_MAX)) SKIP_RETURN(-6); /* Need a later version */ if (timestamp) { fread(timestamp,1,SIZEOF_TIMESTAMP,f); /* read certificate timestamp */ @@ -1214,10 +1215,10 @@ } /* Here's a good format for display of key or signature certificates: - -Type bits/keyID Date User ID - -pub 1024/xxxxxx yyyy-mm-dd aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa - -sec 512/xxxxxx yyyy-mm-dd aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa - -sig 384/xxxxxx yyyy-mm-dd aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa +Type bits/keyID Date User ID +pub 1024/xxxxxxxx yyyy-mm-dd aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa +sec 512/xxxxxxxx yyyy-mm-dd aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa +sig 384/xxxxxxxx yyyy-mm-dd aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa */ if (moreflag) @@ -1406,7 +1407,7 @@ if (mcguffin && strlen(mcguffin) > 0) fprintf(pgpout,PSTR(", looking for user ID \"%s\"."),LOCAL_CHARSET(mcguffin)); } - - fprintf(pgpout,PSTR("\nType bits/keyID Date User ID\n")); + fprintf(pgpout,PSTR("\nType bits/keyID Date User ID\n")); } for ( ; ; ) { long fpos = ftell(f); diff -u3 -r pgp/src/pgp.h pgp-apb/src/pgp.h - --- pgp/src/pgp.h Fri Jun 11 09:44:11 1993 +++ pgp-apb/src/pgp.h Wed May 25 17:33:27 1994 @@ -118,7 +118,9 @@ #define CK_ENCRYPTED_BYTE 2 /* Conventional key is encrypted */ /* Version byte for data structures created by this version of PGP */ - -#define VERSION_BYTE 2 /* PGP2 */ +#define VERSION_BYTE_MIN 2 /* PGP2 to 2.5 */ +#define VERSION_BYTE_MAX 3 /* PGP2.6 */ +#define VERSION_BYTE_DEFAULT 2 /* PGP2 */ /* Values for trust bits in keycntrl packet after key packet */ #define KC_OWNERTRUST_MASK 0x07 /* Trust bits for key owner */ @@ -230,6 +232,8 @@ extern int compl_min; /* number of fully trusted signatures needed */ extern int max_cert_depth; extern char pager[]; /* file lister command */ +extern char armor_version[20]; /* version text in armor output */ +extern int version_byte; /* PGP packet format version */ /* These lists store hashed passwords for future use. */ /* passwds are passwords of as-yet-unknown purpose; keypasswds diff -u3 -r pgp/src/zipup.c pgp-apb/src/zipup.c - --- pgp/src/zipup.c Mon May 10 01:39:19 1993 +++ pgp-apb/src/zipup.c Wed May 25 17:49:09 1994 @@ -37,7 +37,7 @@ # define fhow (O_RDONLY|O_BINARY) # else /* !MSDOS */ #ifndef AMIGA - - long lseek(); + off_t lseek(); #endif /* AMIGA */ # define fhow 0 # endif /* ?MSDOS */ -----BEGIN PGP SIGNATURE----- Version: 2.whatever iQCVAgUBLeN86t7alOJsS1cfAQFS1gQArASHvKV51lLRIuaSiyAqF6h9XXQpalZo jdeZpoCC7P8oEe4inKNbtmFqPcQl8uTVlpTdUxJeErDLxSoDXlw04csW6gNssaFL 07+DpXqoogrOV9+kaPflNl+U3O1EWEDMGG064uDSSgJXLldYs8gGONOWpMV3EqZr tdQzYgc0rBM= =wsTt -----END PGP SIGNATURE-----

2 1

Re: Canadian site
by markh＠wimsey.bc.ca 17 Dec '03

17 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- Subject: Re: Canadian site > > For any other Canrem.com members reading this, I've uploaded a file > called PGP26INF.ZIP, that contains a copy of all files that can be > downloaded from net-dist.mit.edu, without recognition as a Canadian, > or U.S. Site. You should be able to get PGP 2.6 from either ftp.netcom.com:/pub/mpj or ftp.wimsey.bc.ca:/pub/crypto/software/dist/US_or_Canada_only_XXXXXXXX/PGP/2.6 for wimsey cd /pub/crypto/software read the README file if you agree to the conditions follow the instructions contained in that file. for ftp.netcom.com cd /pub/mpj read the file README.MPJ and follow the instructions. (the version of PGP 2.6 on wimsey is an image of the version from netcom). Mark -----BEGIN PGP SIGNATURE----- Version: 2.4 iQBVAgUBLeOfHmrJdmD9QWqxAQEZQwH9HpaZ6urIrcC2n7X9PrYLlHVExEBrCjx7 5Qrab4U/CD3A+lreO0KVPLBZuKmLLD2hEDZ2REPkiPbU+Rd54KPjrw== =gXzD -----END PGP SIGNATURE----- -- Mark Henderson markh(a)wimsey.bc.ca - RIPEM MD5: F1F5F0C3984CBEAF3889ADAFA2437433 ViaCrypt PGP key fingerprint: 21 F6 AF 2B 6A 8A 0B E1 A1 2A 2A 06 4A D5 92 46 low security key fingerprint: EC E7 C3 A9 2C 30 25 C6 F9 E1 25 F3 F5 AF 92 E3 cryptography archive maintainer -- anon ftp to ftp.wimsey.bc.ca:/pub/crypto

1 0

Trust Models
by peace＠BIX.com 17 Dec '03

17 Dec '03

marcos> Is there a mailing list explicitly talking about the "distributed > trust' issue? I am very interested in reputational mechanisms. If not > mailing lists, what papers should I download? rishab>I don't think there's a mailing list on this. I know that at least Bill Garland >wants to set up an experimental reputation market. Anyone else know about >papers or projects? I have a paper that I am preparing for publication called "Trust Models" that discusses the difference between PGP & PEM and looks towards a synthesis of these two models. Since it is not yet published, I will only send it to those who agree (1) not to rebroadcast it, and (2) send me their thoughts on it. Peace

1 0

Re: Graph isomorphism based PK cryptosystems?
by smb＠research.att.com 17 Dec '03

17 Dec '03

So, if JPP publishes it, it would be a matter of trust that he wouldn't patent it. Okay, I have no problems with that. However, it is still patentable. What could be done to make it unpatentable? Nothing. Under current U.S. practice -- and I think non-U.S. as well -- his system is quite patentable. Furthermore, patent applications in the U.S. are confidential. If jpp wants to patent it, he can. If he publishes first, he can only patent it in the U.S. If someone else has invented it first, they can file for a patent, or try to have jpp's thrown out if he should file. (The U.S., unlike the rest of the world, has a first-to-invent rule for patents.) But the one thing that's unconditionally barred is someone else filing a patent on something jpp invents and publishes.

1 0

Factoring with a quantum computer (Citation)
by jamesd＠netcom.com 17 Dec '03

17 Dec '03

First: Don't panic. So far no usable quantum computer has been built. It will be a long time before one is. Secondly a quantum computer capable of factoring 1024 bit keys will require polynomially high precision in its extremely tiny components, whereas a classical computer will only require order one precision, regardless of the number of bits. This may well not be feasible until we are close to achieving nanotechnology. (That is polynomially high precision, which achievable. Non polynomially high component precision is of course impossible for problems large enough to be interesting) Secondly: Yes, quantum computers will indeed be able to solve in polynomial time many problems that require non polynomial time for classical computers. In plain English that means that they can solve problem classes that ordinary computers cannot solve. Factoring is one of those problems. This result was established by E. Bernstein and U. Vazirani, and has recently been published as: Quantum Complexity Theory, Proc. 25th ACM Symp. on Theory of Computation, pp. 11--20 (1993). Barak Pearlmutter summarizes the current status as follows: The class of things a quantum computer can compute in polynomial time is called QRP. The class of things a regular randomized computer can compute in polynomial time is called RP. It is not known whether RP>P. But, under the usual assumptions, we know RP < QRP <= P^#P. It is not known whether QRP < NP. Factoring has not been shown to be in RP, nor has it been shown to be NP-complete. (If factoring were shown to be NP-complete, then we would have NP=coNP, a big deal, and also NP <= QRP, an immense deal assuming that accuracy problems don't make them impossible to build.) So it might be that NP <= QRP. Also plausible is RP < QRP < NP. In any case, the public key cryptosystems we thought were secure (RSA, discrete logs) has now had their viability brought into serious question. Even if QRP<NP (which it might not be) it might be the case that any public key cryptosystem can be broken in QRP, since no public key cryptosystem has been exhibited for which the inverse problem is NP-complete. (Some have been related to NP-complete problems, but with the reductino in the wrong direction, which proved fatal.) I will shortly give a hand waving explanation of why it is so, why a quantum computer is potentially capable of solving problems that a classical computer cannot solve within the lifetime of the universe. I am not able to cite Shor's result, showing that factoring is among the problems that a quantum computer can solve efficiently, because he is not yet published. I am attempting to reconstruct his results from information that people have given me about them, but so far I am making fairly heavy weather of it. He is applying the remarkable result of Bernstein and Vazirani to a particular interesting case. So if you want to go an prove that his method cannot possibly work, simply go and prove that the methods of Bernstein and Vazirani, cannot possibly work. The quantum magic part is in Bernstein and Vazirani's work, not Shor's work. (I can assure that they will work.) Some of you folk may remember the great flame war on this topic on the Extropians mailing list. I mentioned this result, without giving source or authority (it had not been published to my knowledge at that time), and Price flamed me vigorously, calling me an idiot, an ignoramus, a crackpot, and a fool, and so forth, and offered to bet 5000 British pounds, then about US$7500, on the matter. $7500 bet! I of course cheerfully accepted. Alas I wanted the judging to be done by somebody who had published on quantum computing in a refereed journal, (like Bernstein) and I wanted that person to hold the pot. Michael reluctantly agreed to having the bet judged by a qualified person but he wanted the pot held by a bunch of people who are involved in a business concerning which I have very grave and serious suspicions. I attempted my various means of persuasion to get him to put up the pot. Alas, Michael was not willing to let the pot for the bet go outside the control of him or these dubious people. Eventually I resorted to a vigorous attempt to shame him into putting up the money. My unkind comments concerning Michael and his pals eventually resulted in me being expelled from the Extropians list, for flaming Michael and defaming the Exi board. But since Bernstein and Vazirani are now published, and no further suckers seem willing to come forward and bet, I am now going to give an vague hand waving explanation of their results. A quantum event has many outcomes, all of them happening at once. Consider a quantum computer with 1024 bits of memory. (128 bytes, not 128 megabytes) A classical computer can only be in one state at a time. A quantum computer can be, and usually will be, 2^1024 states simultaneously. That is 10^306 states, or about 1 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 states simultaneously. Obviously a quantum computer can explore state space a lot more efficiently than a classical computer can. In fact a classical computer simply cannot explore a space that big, even if the computer were to fill then entire universe. This however is not much use, because the probability of observing it in the correct state is equally low. The clever trick of Bernstein Vazirani is to arrange for constructive interference between correct answers and destructive interference between incorrect answers. This requires the quantum computer to retain quantum coherence. It is evident that most of the people blowing off and saying "Quantum computers are bullshit" do not understand quantum coherence etc. (At least Michael Price had some faint grasp of quantum mechanics.) Thus a quantum computer has vast potential for problems that involve the combinatorial explosion. Unfortunately, because the transformation is unitary, a quantum computer cannot efficiently report back the one correct result that it has discovered to the any one particular reality Although getting the information out of a quantum computer is inherently inefficient, it is not always nonpolynomially inefficient. My judgment, for what it is worth, is that (for the vast majority of problems that are not NP complete, but are NP and are actually interesting, because they have an answer that is true rather than an answer that is marginally better than other answers) it will be reasonably easy to find ways of persuading a quantum computer to disgorge the necessary information that are only polynomially inefficient. These methods will be extremely general, rather than domain specific, and algorithms will improve in the direction of generality. A quantum computer will always operate very rapidly, for any algorithm that one would wish to program on a quantum computer. For many algorithms, the probability of it emitting the right answer is rather low. If the probability is non polynomially low, then the algorithm is of course useless. A number of algorithms have now been shown, including factoring, where the probability is only polynomially low, which is of course enormously better than non polynomially low. The precision of the components of the quantum computer must be proportional to the time taken by one shot of the algorithm, since in order to get the information out of the computer with a reasonably high probability, we require the class of true answers to interfere constructively, while the vastly larger class of incorrect answers interferes destructively. If the computer sits there for too long, random thermal noise will destroy coherence, and all the answers will interfere randomly, so the probability of getting the right answer out (observing the computer in the desired state) will be non polynomially low. Similarly, if the quantum computer takes too many steps, imperfections in its components will destroy quantum coherence. When the coherence is lost, a quantum computer is equivalent to a non deterministic computer. When it retains quantum coherence, it can surpass a non deterministic computer by an exponentially large factor. --------------------------------------------------------------------- We have the right to defend ourselves and our | property, because of the kind of animals that we | James A. Donald are. True law derives from this right, not from | the arbitrary power of the omnipotent state. | jamesd(a)netcom.com

1 0

Re: PGP 2.6 is dangerous in the long term ?
by bshantz＠spry.com 17 Dec '03

17 Dec '03

>What if (the cypherpunk community) comes out with a secure program that >doesn't rely on RSAREF? Can it be done? >NOTE: I very likely don't know what I'm talking about. >____ Robert A. Hayden <=> hayden(a)krypton.mankato.msus.edu >\ /__ -=-=-=-=- <=> -=-=-=-=- > \/ / Finger for Geek Code Info <=> Political Correctness is > \/ Finger for PGP 2.3a Public Key <=> P.C. for "Thought Police" >-=-=-=-=-=-=-=- >(GEEK CODE 1.0.1) GAT d- -p+(---) c++(++++) l++ u++ e+/* m++(*)@ s-/++ > n-(---) h+(*) f+ g+ w++ t++ r++ y+(*) Robert, The c-punk community could come up with a secure data encryption program that did not rely on RSAREF, but it would take time. When you look at the code for RSAREF, the algorithm really isn't all that bad. (In fact, IMHO, it's quite good.) Several questions that need to be brought up are, "How much time would it take?" "How solid is the algorithm", "Can it be accessible to the public?" Also, "Will whatever the c-punk community comes up with eventually turn into a PGP-like international version control nightmare as is happening now (Patents, international embargo's, NSA complaints, lack of backwards compatibility)?" It is my opinion, that since we do not live in a perfect world, eventually we'd run into the same problem we have now. Is it worth it? (Oooh, I smell a discussion about to start.) Brad :::::::::::::::::<<< NETWORKING THE DESKTOP >>>::::::::::::::::: Brad Shantz Internet : bshantz(a)spry.com SPRY Inc Ph# (206) 447-0300 316 Occidental Avenue S. 2nd Floor FAX (206) 447-9008 Seattle, WA 98104 ---------------------------------------------------------------- "In gopherspace no one can hear you scream." ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::

1 0

Re: IBM's NetSP
by jim＠bilbo.suite.com 17 Dec '03

17 Dec '03

> CDMF alleviates the worry of having your data flow across > the network in clear text. The degree of security is > equivalent to encryption using DES but with keys limited > to 40 bits. IBM has obtained approval from the US > Government to export CDMF in products without the > license required to export products containing DES. > Exportable and limited to 40 bits? Sounds like they're using RSA's RC2 algorithm. Jim_Miller(a)suite.com

1 0

Another AOL
by fnerd＠smds.com 17 Dec '03

17 Dec '03

CDMA -- Carrier (Collision?) Detect Multiple Access, Code Division Multiple Access. Grr. Tim has me started. -fnerd - - - - - - - - - - - - - - - To auditors without the code, calls seem indistinguishable from noise. --George Gilder -----BEGIN PGP SIGNATURE----- Version: 2.3a aKxB8nktcBAeQHabQP/d7yhWgpGZBIoIqII8cY9nG55HYHgvt3niQCVAgUBLMs3K ui6XaCZmKH68fOWYYySKAzPkXyfYKnOlzsIjp2tPEot1Q5A3/n54PBKrUDN9tHVz 3Ch466q9EKUuDulTU6OLsilzmRvQJn0EJhzd4pht6hSnC1R3seYNhUYhoJViCcCG sRjLQs4iVVM= =9wqs -----END PGP SIGNATURE-----

1 0

Re: Graph isomorphism based PK cryptosystems?
by smb＠research.att.com 17 Dec '03

17 Dec '03

This is one of the main reasons behind such publications as the IBM Technical Journal--the publication of results not worth patenting themselves, but definitely worth preventing others from patenting. Actually, no. IBM has a separate publication for just that purpose. The Technicnal Journal (and others like it) are intended as PR vehicles and as gold stars for researchers -- that's an important part of our salary, as it were.

1 0