EddyHawk's Info List
---
FILE (ENCRYPTER/CIPHER), VERIFIER
---
-TinyIDEA
 By: Fauzan Mirza (England)
 Year: 1995-1999
 V3 [1996]
  heavy size optimization
 V4 [1999]
  heavy size optimization
  by Mark Andreas
  disadv: PGP-crypted, key only available for USA resident
 Note: the smallest IDEA cipher implementation

-TinyFish
 by: Dutra Lacerda
 Note: the smallest BlowFish cipher implementation

-BFA (BlowFish Advanced)
 By: Markus Hahn
 Compiler: BP V7.0 (?)
 V7 [1996?]
  uses Blowfish, Blowfish32, TDES & Cobra cipher & mixing of them
  GUI
  shareware
   max 5 char passphrase
  note: the doc mentions that this triple DES implementation isn't real
   because it removes initial & final permutation (considered as
   cryptographically worthless) to speed things up

-UCRYPT (UltraCRYPT)
 By: AIP-NL (The Netherlands)
 Compiler: Borland C++ [1991] + assembly
 V2.37b [1996]
  UCrypt is part of UC2 archiver and is designed to crypt uc2 archive only,
   but we can trick it to crypt anyfile by putting UC2 ID header ('UC2'+ 1Ah)
   in front of the file-to-be-crypted
  UCrypt uses 2 encryption algorithms:
  -3DES (Triple DES) + MD5
   apparently AIP-NL doesn't aware that MD5 is (half) broken in 1996
   .uses MD5 output as 2 random DES keys
   .crypt with key1, decrypt with key2, crypt with key1 (112 bit)
    but this isn't? 3DES:
    UCrypt doesn't process each block 3 times, but the whole file 3 times
     .since UCrypt passes my no-pattern-file test, it surely uses a
      chaining mode (ex: CBC). since DES-crypted block isn't the same
      with 3DES-crypted block while serves as IV to next block,
      it will generates different? ciphertext than standard 3DES
   .crypt keys with password & crypted file
    it means that the keys are stored in crypted archive
    -equal to trap door :)
    while imho, the password should be hashed with MD5 to be 2 DES keys, not
     current time/date? hashed through MD5
    and does UCrypt use 3DES again to crypt those keys? I don't think so
     apparently the password is directly used as the crypt key
   .the (de)cryption is very slow (of course, it's DES :)
    for 2,636kb (2,700,000 byte) file on my 1st computer:
    crypt: 242 second, decrypt: 233 second
    or about 80+ second for single DES
    while my Mr.Hyde V0.46a can attain 13 second
   .UCrypt is claimed to be enhanced against brute-force attack
    -maybe it adds salt on password to 'frustate on-line password guesser'
  -MD5 based one time pad
   .of course, this is not OTP, because the key must be:
    -truly random numbers
    -as large as plaintext
    -used only once
   .password -> MD5 hash -> PRNG?
   .the (de)cryption is slow
    for 2,700,000 byte file on my computer: crypt/decrypt=80/78 second
  UCrypt has 'destroy plaintext' option:
  -very slow wiping (wipe,fill,scramble,wipe,scramble,delete)
  -doesn't erase file entry
  -I think SWAG/WipeFile (used in Mr.Hyde) is faster & more secure :)
  src is available only to be verified

-PKZIP -S
 By: PKWARE (USA)
 V1.0
  broken by Biham & Kocher's attack < 1 day with few 100b of known plaintext
 V2.04g [1993?]
  stream cipher
  variable length key
 V2.50  [1999]
 Note: part of PKZIP archiver (can only crypt ZIP archive)

-PGP (Pretty Good Privacy)
 By: Philip Zimmerman (USA) et al/Network Associates
 year: 1990-2000
 Type: pub-key.(ciph/sign/verifi)er, free
 V2.3 [1993?]
  at that time, USA regulation classified its strong crypt as weapon, so
   USA residents was unallowed to export them. PGP uses RSA and becoming
   very popular until it spreads overseas. To be safe, 2 PGP version is
   made. USA PGP user can continue using usual PGP, while non-USA PGP
   user are advised to only use PGP international release (PGPi) to help
   the author
 V2.6.3 [1996]
 V5.0 [1997]
  DOS32 ver (last? DOS ver)
  uses Diffie-Hellman, IDEA & CAST cipher
 V6.53 [2000]
  Win32 ver
  Bussiness version

-CRY
 By: Harry J. Smith
 Year: 1987-1992
 Compiler: BP V6.0, BC++ [1991]
 V6.00 [Nov 1992]
  C & Pascal src is provided
  spent 1,400 hours of author time
 Note: to replace DES

-PEGWIT
 By: George Barwood
 Compiler: TopSpeed C [1989] (?)
 Type: pub-key.(ciph/sign/verifi)er, free
 V8.71
 Note: uses Elliptic Curve, Square cipher, SHA1 hash

-BLOWFISH-CBC
 By: Dutra Lacerda (Portugal)
 Compiler: BP V7.0
 V1.5a [27 Aug 1996]
  src is provided
   but prog x uses non-pub 8086 ASM routines -> faster
  uses BlowFish cipher
  crypt speed ~ 200,000b/sec (incl. I/O time)
  no longer updated?

-ARJ -G
 by: Robert K. Jung/ARJSoftware
 V2.75a [Oct 2000]
  crypt uses
   simple XOR (-g) -> easy? to break
   GOST 28147-89 cipher in 64bit CFB mode
    40bit key (-g -hg!) to meet USA export regulation (but not anymore?)
    separate module ARJCRYPT.COM, 256bit key (?) (-g -hg)
     only for USA user
 note:
  -g(arble) is simply transform passphrase into same length key and
  repeatly xor it with same length archive data
  part of ARJ archiver (can only crypt ARJ archive)

-Krypto
 by: Bathysphere
 V1.1 [1993]
  up to 255 char passphrase
 note:
  crypt uses date/time stamp (same plaintext -> different ciphertext)
  if U change crypted file d/ts, decrypt will fail regardless of correct key
 author: crypt uses novel chaos theory

-Mr.Hyde
 by: EddyHawk
 V0.54a [Nov 2000]
 EdH:
  anyone dares to review or should I? :)
  Morgan et al reported few security flaws, but I don't thoroughly test it

-Hermes
 by: Morgan (Poland)
 type: 386, fpu, 4Mb RAM, dos > v3.3
 V1.0b [22 Oct 2000]
  can't run under win
 note:
  pres uses SixPack
  crypt uses FFT, PRNGs, PNC, buffer permute, double feedback stream cipher,
   polymorphism

-VCH (VouCH)
 by: Awais M. Hussain (Pakistan?)
 year: 1993-1997
 type: 386, verifier
 compiler: tc v2.0 [1988]
 V1.3 [May 1997]
  lite ver