🚀 go-pugleaf

In article <tom.712397505@cluster> tom@stallion.oz.au
(Thomas Essebier) writes about a proposed cipher algorithm:
> By far not being an expert on cryptology, I would appreciate some 
> comments on the crypto scheme below. 
>  [ use a standard linear congruential random number generator (= RNG),
>    primed with the key, to derive a stream of bits, which is then xor-ed
>    with the data ]

jonathan@chpc.utexas.edu (Jonathan Thornburg) follows up,
in article <1992Jul31.034244.9150@chpc.utexas.edu>, with:
> I'm sorry, but this is a very weak cipher -- it's very easy to break
>  [ via (1) brute forcing all 2^32 possibilities
>    or  (2) recovering the RNG output stream using known attack methods ]

I appreciated Hal Feinstein's later reply
(hal@gateway, in article <1992Jul31.124951.18341@linus.mitre.org>).

A further question I have on this apparently mundane topic is this:

How secure is such an x-or scheme when ammended with the following:
   a) additionally x-or using characters in the key, and
   b) perform character transposition within blocks of the message.

Since the ASCII set is fairly small (96 - 256 chrs), it seems that using
a suitably large block size, >> 256, with this ammended method might: 
   - render the brute force attack infeasible;
   - foil RNG output stream recovery attempts;
   - resist (to some satisfactory degree?) non-plaintext attacks; and
   - resist (to a lesser degree?) a single plaintext attack.
Multiple plaintext attacks would probably succeed.

Generally, I like the idea of RNG/PRNG x-or-ing and its "randomizing"
effect on the message data.  Transposition seems to complicate attack
methods, at the expense of requiring an encode versus decode mechanism
(so that the substitituion/transposition steps are performed in the
correct order).  However, like Tom in the original post of this thread,
I am no expert; but I am interested, curious, and playing with code in
this vein.

Comments?
______________________________ Mike Cepek ______________________________