From jim@acm.org Sat Jan 21 14:48:53 1995
Path: comlab.ox.ac.uk!uknet!str-ccsun!news.dcs.warwick.ac.uk!hgmp.mrc.ac.uk!sunsite.doc.ic.ac.uk!agate!overload.lbl.gov!lll-winken.llnl.gov!uwm.edu!vixen.cso.uiuc.edu!howland.reston.ans.net!usc!rand.org!mycroft.rand.org!not-for-mail
From: Jim Gillogly <jim@acm.org>
Newsgroups: sci.crypt
Subject: Re: Identify this scheme?
Date: 16 Jan 1995 14:25:02 -0800
Organization: Banzai Institute
Lines: 110
Sender: jim@mycroft.rand.org
Message-ID: <3fernu$iv5@mycroft.rand.org>
References: <3femrb$cjo@on-ramp.ior.com>
Reply-To: jim@acm.org
NNTP-Posting-Host: mycroft.rand.org

In article <3femrb$cjo@on-ramp.ior.com>,
Charles J. Fishburn <fishburn@on-ramp.ior.com> wrote:
>I am somewhat new in my study of encryption algorithms and terminology et 
>al.  And I am curious about what the following encryption scheme is 
>called (i.e. does it have a specific name (e.g. MD4, RC5), or is an 

Yes, it's called "weak".

>clear.  And here is the pertinent C code (as implemented from "What's the 
>Code", by David Stafford, Computer Shopper, July 1993:
>
>#define MULTIPLIER 0X15A4E35L
>#define INCREMENT 1
...
>  RandomSeed = MULTIPLIER * RandomSeed + INCREMENT;
>  return (RandomSeed);
...
>  while ((Ch = getc(InFile)) != EOF)
...
>    fputc(Ch^RandomNumber(256), OutFile);
...

>On my 486DX-50, 8MB RAM, en/decrypting a 100K file takes probably 15-20 
>seconds.  (I haven't taken a very recent estimate... so it may be a 

Cracking it (ciphertext-only) takes considerably less time, since you don't
need much ciphertext to find a workable key.  I append the response I wrote
when this thing was published in comp.souces in Aug 93.

	Jim Gillogly
	Hevensday, 25 Afteryule S.R. 1995, 22:20
____________________________________________________________________________

From: jim@rand.org (Jim Gillogly)
Newsgroups: comp.sources.d
Subject: Re: syf - file encryption program
Date: 31 Aug 1993 19:27:21 GMT
Organization: Banzai Institute

Here's a program to break the encryption on an ASCII file, in case you
forget your password.  :)  As I note in the comments, this cipher gives
you security nearly an order of magnitude better than ROT-13!  Not nearly
as good as simple substitution, though.

	Jim Gillogly
	Highday, 9 Halimath S.R. 1993, 19:27

--------------------------------------------
/* syfcrack: decrypt the output of syf.c without the key.
 *
 * Usage: syfcrack < encrypted-file
 * Output: an equivalent decryption key
 *
 * Reference: v39i066 of comp.sources.misc, author blackman@cs.buffalo.edu
 *
 * Assumes the plaintext is ASCII.
 * The relevant observation is that the program uses the low-order 8 bits
 * of a linear congruential PRNG, which cycles in 256 steps... the same
 * 256 for each key.  The result is that there are 256 possible keys,
 * meaning he gets almost an order of magnitude improvement over ROT-13!
 *
 * 31 Aug 93, SCRYER
 */

#include <stdio.h>

#define MULTIPLIER 0x015A4E35L
#define INCREMENT 1

long RandomSeed;

int GetRandomNumber( int Range )
{
	RandomSeed = MULTIPLIER*RandomSeed+INCREMENT;
	return(RandomSeed%Range);
}

#define CTSIZE 512              /* Don't need much ciphertext to break it */

char ciphertext[CTSIZE];
int ctlen;

main()
{
	long Start;
	int i;

	for (ctlen = 0; ctlen < CTSIZE; ctlen++)
		ciphertext[ctlen] = getchar();
	for (Start = 0; Start < 256; Start++)
	{
	    RandomSeed = Start;
	    for (i = 0; i < ctlen; i++)
		if ((0x80 & (ciphertext[i] ^ GetRandomNumber(256))) != 0)
		    goto leave2;
	    printf("%d is a reasonable key: ", Start);
	    RandomSeed = Start;
	    for (i = 0; i < 20; i++)    /* Here's a sample */
		putchar(ciphertext[i] ^ GetRandomNumber(256));
	    putchar('\n');
	  leave2:;
	}
}
-- 
	Jim Gillogly
	Highday, 9 Halimath S.R. 1993, 19:27
____________________________________________________________________________
-- 
	Jim Gillogly
	Hevensday, 25 Afteryule S.R. 1995, 22:25