Re: The Most Profound Distillation of the Feedack from Modern Cyptography.



On 9/22/2011 10:05 AM, Jeffrey Goldberg wrote:
On 11-09-22 11:36 AM, Prof Craver wrote:

All he's saying is that mathematically impossible encryption will
remain mathematically impossible as long as digital computers are
digital.

Ah. So he is pointing out that computers must represent numbers using
finite representations. Therefore all numbers that computers use are
countable.

Technically that's a true statement.

Indeed, but he seems to see this as a problem (and presumably thinks
that he has a solution to the "problem").

Usually I run into those people criticizing proof of correctness on
the grounds of undecidability. I didn't know they were into cryptography, too.

Historically, there were some analog encryption schemes for
voice and video. They tended to suffer from "residual intelligibility"; some sense of the content still got through.
Turing once proposed using a device which did an analog modular
add to add a random noise stream to audio. The receiving end could then
reverse the operation and decrypt. Keeping the noise streams (one time
keys on phonograph records) in sync was too hard to do with 1940s
technology, so that idea didn't work out. This is not SIGSALY;
that was a competing digital system, with 15 racks of tube gear.
Turing knew about that and was trying to design something simpler.

There's no cryptographic advantage in doing this in analog.
Turing was just trying to get the system tube count down.

John Nagle
.



Relevant Pages

  • Re: Pursuit of the Graf Spee, May 6, TCM, 4:15 AM EDT
    ... out bombes and similar digital computers. ... That leaves analog ones. ... I seem to remember the bigger gun fire control system having such ... either for the powder magazine or the local ambient air. ...
    (sci.military.naval)
  • Re: 50 years later, Marvin Minsky still doesnt get it
    ... in algorithms, from step to step, or frame to frame. ... All digital computers are made up of 100% analog parts. ... AI is simply a problem of making a robot act like a human. ...
    (comp.ai.philosophy)
  • Re: 50 years later, Marvin Minsky still doesnt get it
    ... All digital computers are made up of 100% analog parts. ... AI is simply a problem of making a robot act like a human. ... They standardly compare successive frames to each ...
    (comp.ai.philosophy)
  • Re: Creative Word
    ... the languages. ... Digital computers can only go so far and only very ... Analog 'human' processes must intervene. ... Digital computers have only recently achieved speed levels ...
    (soc.religion.bahai)
  • Re: Alan Turing
    ... To head up an article about "Parry", ... extreme paranoia, Analog used the title "Turing Point" (which the eye ...
    (rec.arts.sf.written)