[Cryptography] how to detect breakage -- lures etc.??

[Ray Dillinger]

On Mon, 2020-01-06 at 07:21 -0500, Arnold Reinhold via cryptography
wrote:
> 
> If I understand your concept correctly, you are proposing a cipher
> based on a pseudo-random sequence generator made up of rotors of
> different periods with incommensurate lengths to produce a very long
> overall period, where the connections between the individual rotors
> can be varied according to a key so as to produces a very large set
> of different sequences. 

Yes.  Thank you for stating it more clearly and succintly than I think
I was capable of. 

> It’s my understanding that is pretty much what the NSA ended up doing
> in the 1950s, except they used binary electronic shift registers with
> feedback instead of alphabetic rotors. Your rotor idea may well have
> worked, but major military communications circuits were migrating to
> radio and landline teletype from Morse code, so binary made sense and
> electronics were the future. 

Right.  The problem with rotor machines in general isn't that they're
insecure.  Although many of the early ones were insecure, that's not
endemic to rotor machines as a class.  In fact, as they run on systems
that are known with absolute certainty to NOT be running any kind of
malware or spyware (because they can't), they can arguably be even more
secure than modern ciphers.

The *real* problem with rotor machines is that if you try to operate
them fast enough to manage an internet data link, they explode. Random
explosions, sadly, aren't really a feature consistent with the
harmonious operation of modern office environments. For starters you
have to explain to OSHA. 

I happen to have a soft spot for rotor machines in my heart, but I
don't mistake them for a viable alternative to modern ciphers for data
communication.  Written correspondence is pretty much the limit of
their applicable domain. 

I will have a look at the shift-register machines.

				Bear