Re: Order being born from disorder?-Question

[inet@intellisys.net (brian whatcott)]

At 04:01 PM 5/17/97 EDT, LUDWIK KOWALSKI wrote:
> ... I would like somebody to assign entropies
> to the following patterns. Each of them is ten bits long sequnce of zeros
> and ones.
>              0000011111   entropy= ?
>              1111100000   entropy= ? (probably the same as above)
>              0101010101   entropy= ?
>              0010011011   entropy=?
>              0001001100   entropy=?
>              0000000000   entropy=?
>              1100110011   entropy=?
>              0001000001   entropy=?
>
> And explain the algorithm of assignments. Thanks in advance, Ludwik
Out of respect for the effort that Bowman put into his exposition of 
entropy, which
spoke neither upwards, downwards or with any detectable side, I will attempt to
answer it in his mode:

Treating this entropy as a 'Shannon entropy' which measures how uncertain 
these sample outcomes of yours are,  Bowman asks me to consider what is the
 minimal amount of information needed to uniquely define each of these
 ten-binary-bit values.

 This appears to be a particular case to which he alludes, where there are
 1024 possible outcomes, values or possibilities of which EACH is equally likely
(i.e. 1111100000 is as likely as 1001000100)
so that the entropy S expressed in bits = log, base 2(1024) 
which is 10

Hence the explicit answer to each of your questions is ten (bits)

Regards
brian whatcott <inet@intellisys.net>
Altus OK