Allocating code space

 Continuing on the system for secure, secret keys in the hand held smart card.

The decode is quantum, an analog logic stack h high, each level having a 100 pico second average delay with a 10 pico second sigma. If the stack is ten high then the lop layer of logic has a delay, total, of a thousand pico second with 100 pico seconds of sigma.

If my sampler on the quantum computer is five pico seconds then I get w * h * 2 samples. W is eight. treat that as a unit sample. For the simplest card passed around, it must train the decoder decode four keys, 256 bytes each.  This selector has to identify one of one thousand bytes with 1/5 accuracy.     The bytes are generally fixed, the fab, the correspondent bank, the self.   

There is a valid compression that uniquely does the job with minimum transactions. It needs an integrated, analog sample sthat is as fast as five pico seconds.   There exists a Huffman code that can do this with a ten stack.   Baring recoverable transmission errors, this solution is complete, and meets the first contract law. The analog logic runs under the kernel code, operated by the kernel in from spectre compatible cache contract.

Meaning

You have a quantum computer. I have a map that utilizes your computer for a finite group of messages.  My map is not complete, no one knows how to complete my map over the entire set of messages.  My message set is fixed, a finite set of secret keys other cards send me.    Whenever they send me a message, it is multiplied by the key they sent previous, but I trained on that key, so all our encrypted message can run through the irreversible quantum decoder, as these messages are multiplied by the trained key.

I have the general card to card encryption key, correspondent bank, Fab and self.  None of these keys need be kept, once trained the card no longer carries secret keys and in nearly impenetrable to countefeit.

Note. I have often thought of doing this with unique shot noise cources.

Shot noise or Poisson noise is a type of noise which can be modeled by a Poisson process. In electronics shot noise originates from the discrete nature of electric charge. Shot noise also occurs in photon counting in optical devices, where shot noise is associated with the particle nature of light.

That is what we are talking about, the excess or deficit of shot noise. The rate of shot noise is Gaussian, random enough if repeatable. So a single shot noise source couple to a high speed five pico second sampler works. Add it right to the chip substrate.

Over 1000 bytes we might have a bit less than 1% error ratio. Results in a retransmission, like having a dirty chip on your ATM card.