something for later

[vh]

something to python:

this is the scenario. person a wants to tell person b a message. However, person a can't be assured that someone will not intercept the message and figure it out. to complicate things, person a and person b know nothing about each other. person a can encrypt the message but then person b cannot decrypt it; if he can, that means anyone who has intercepted the message can too.

solution: person a encrypts the message with x, sends it to person b. person b encrypts the x encrypted message with y and sends it back to person a who unencrypts x and sends it to person b to unencrypts y.

example: person a sends the number '11' to b
first a multiplies by 3
11*3=33
person b multiplies by 7
33*7=231
person b sends it back to person a
person a divides by 3
231/3=77
person b divides by 7
77/7=11
person b recieves the message

[atomic7732]

what if I wanted to send "what if I wanted to send"?

[Darvince]

THEN ENCODE IN 10-BIT AND THEN NUMBER.

[vh]

what if I wanted to send "what if I wanted to send"?

well yeah as darvince said just give all the characters a two number code which leaves plenty of space for punctuation and then convert that sentence into a number and send that

[blotz]

oh

[atomic7732]

what if I wanted to send "what if I wanted to send"?

well yeah as darvince said just give all the characters a two number code which leaves plenty of space for punctuation and then convert that sentence into a number and send that

but how does that person know what two number code corresponds to what?

For example for "extra" security I might decide that:
01 - z
02 - y
03 - x
04 - w
...
25 - b
26 - a

while you on the receiving end might think I just encoded it alphabetically

and that's just a simple encoding difference, what if I came up with a code to encrypt the order of characters?

in that case I might want to say "hi".

hi -> 1918... but how do you encrypt these characters anyway? if you assumed 0-9 was appended to the end in order of increasing value, you would get

28372836, this sent back to me, i'd decrypt it (0-9 in decreasing order because my encoding is that way) to get

8182 which you would decrypt into nothing. and backwards it's 2818 which is "1i"

Though actually... (100-81)(100-82) = 1918 which is "hi" but how would you come to a conclusion like that with encryption made so that people couldn't actually crack the code with 3 minutes of effort?

[vh]

but how does that person know what two number code corresponds to what?

well in the real world everything is binary so each letter has an official binary representation. if i were to this in python i would make the encoder and decoder use a standardized number-letter code, the extra security is not needed. so a is always 01 and b is always 02 etc.


if you wanted to send 'hi' which might correspond to 0809
>multiply it by 3
2427
person b recieves it, multiplies by 4
9708
you get it back, divide by 3
3236
send it to a, who divides by 4
0809
turns it back into letters

however, by doing 9708/2427 and 9708/3236 it would be possible to figure out the multipliers used.

instead of just multiplying a better idea is to figure out two commutative functions;
f(x)=x^3
f'(x)=root3(x)
g(x)=x^5
g'(x)=root5(x)

0809
apply f(x) to 0809
529475129
person b applies g(x):
41612884801984134225475389695239054060426649
person a applies f'(x):
346531411903049
person b applies g'(x):
0809

[atomic7732]

I'm talking seriously, if someone intercepted the message, your plans are foiled.

[vh]

how so?

[Darvince]

well yes but if they receive it on the way then they get the number 529475129, 41612884801984134225475389695239054060426649, or 346531411903049 which is gibberish

[atomic7732]

how so?

because a-01 b-02 is kind of extremely simple?

unless your functions are like 8th degree polynomials with decimals and radicals...

[vh]

well the functions thing can be debated.

"because a-01 b-02 is kind of extremely simple?"
well if you think this is simple you might as well think security is a joke because this is how pretty much all data is stored no matter what the encryption.
changing this would make the code a one-time pad which cannot be broken by anyone at all unless they have the key (and they don't have the key, thats the point).

[atomic7732]

quantum key distribution

[Darvince]

use atmospheric noise encryption
or just use the inverse of x^5 - x^4 + 5