Re: Confirmation of Shannon's Mistake about Perfect Secrecy of One-time-pad

On Nov 6, 11:46 pm, wangyong <hell...@xxxxxxx> wrote:
On 11 7 , 12 13 , William Hughes <wpihug...@xxxxxxxxxxx> wrote:



On Nov 6, 11:08 pm, wangyong <hell...@xxxxxxx> wrote:

how you can compute, how can you find the probability M not changed.
do not just the the result ,that is useless.

Compute the joint probability of M, K and C
(use the known probability distribtuions of M and K
the fact that M and K are independent, and the formula
for getting C from M and K)

Assume C fixed. Two cases.

For each case note the value of C and compute
the conditional probability of M using the known
joint probability of M, K and C and the definition
of conditional probability.

In both cases you get the same probability
distribution for M

- William Hughes >
- -

a change in form but not in content,you just use the probability when
c is not fixed.

No. I do not use the probability when
c is not fixed. I compute the conditional probability. This
can be computed from the joint probability, but
is not the same as the unconditional probability
(the probability when c is not fixed).

- William Hughes- -

- -

how can you compute the probability when c is fixed,
do not use that when c is not fixed,
if use, tell why you can use.
that is why you take mistake.

I do not use the probability when c is not
fixed to compute the probability
when c is fixed. I use the joint probability of
M,K and C. [Indeed you could use the joint
probability to calculate
the probability of C when C is not fixed
(the marginal) but I do not do this.]
I use the joint probability to
calculate the probability of M or K when
C is fixed (the probability when C is fixed
is defined in terms of the joint probability)

- William Hughes

.

Relevant Pages