Shannon's Entropy v. Clausius's Entropy?

<[Only registered users see links. ]> wrote in message
news:%RPKf.15501$[Only registered users see links. ].prodigy.c om...
in

[Only registered users see links. ]
[Only registered users see links. ]
The correspondence between physical entropy and informational entropy is
conceptual. There is no energy involved in changing informational entropy,
and physical energy does not measure uncertainty or information.

The main resemblance is in the way the words are spelled.

"Analogies are like ropes; they tie things together well, but you won't get
very far if you try to push them." - Thaddeus Stout.


Tom Davidson
Richmond, VA

"tadchem" <[Only registered users see links. ]> wrote

involved
I


entropy,

How do you increase informational "order" (eg. alphabetize a list of names)
without doing _any_ work in the thermodynamic sense?

[Only registered users see links. ] wrote:

<snip>


You can't - but that's not a problem. It doesn't mean that energy has
been stored in the list, it just means that energy has been expended.

Are you waiting for someone to say something like "you cannot increase
informational order without expending physical (thermodynamic) energy"?



SCW

[Only registered users see links. ] wrote:

<snip>


Start by looking at what thermodynamic work *is*:

Work can be the integral of force over a distance, the integral of
pressure over a change in volume, or simply changing potential energy,
or other things:
[Only registered users see links. ]

Each of these definitions requires that the system under consideration
posess a property (mass, charge, pressure, tension, etc) that simply is
not defined for information sciences.

The term "work" exists in both contexts - information and physics -
just as "entropy." The definitions are context-sensitive and
irroconcileable between contexts.

The result is a trap that leads the unwary into the fallacy of
equivocation.

Tom Davidson
Richmond, VA

"Doune" <[Only registered users see links. ].uk> wrote


names)



It's interesting you should say that ;-) If it is true that you cannot
increase informational order without a thermodynamic cost, then why can't
you say that energy is somehow "stored" in the information in the same sense
that energy is stored in a rock that I carry up a hill? Also, what I find
particularly interesting is that while as you say any increase in
informational order comes at some cost in physical entropy, nevertheless, as
I understand it, the amount of energy required to perform computation can be
reduced to arbitrarily low levels. Therefore it seems to me that while
increasing informational order does comes at some thermodynamic cost, it
does not scale the same as it does with respect to physical work and kinetic
energy in the realm of matter and energy.

<[Only registered users see links. ]> wrote in message
news:NdsLf.24436$[Only registered users see links. ].prodigy.c om...

To transfer a bit of information you need some signal stronger than the
noise. Pushing down to the quantum the limit, the energy in each bit must be
some multiple of h since it must be sufficient to be detected above the
uncertainty. Shannon developed the channel capacity in terms of S/N ratio.

Even with S/N < 1 there is way of transmitting some information but speed
decreases.

Down at the quantum level, when noise is quantized S/N can not be less than
1.

As interesting as the parallel between Information and thermodynamic is,
in my opinion there is no physical substance to information, it is a
mathematical concept of order in sequence.

Physically speaking a random array of numbers and an ordered one are not
that different. Information is subjective, a symphony may sound better to
you and me than the same notes played randomly but physically the difference
is meaningless.

On the other hands a gas temperature has physical meaning.

Information has no mass, momentum, charge or any of the stuff that physic
deal with.

When we use electronic circuits to manipulate information, then energy
considerations come into play but this is because we use electrons like
beads on an abacus, we move than around like pieces of clay to keep count.

At this point we use physical quantities arranged in conventional pattern to
model the information symbols. Rearranging symbols does not require energy
per se, unless you do the "thinking" in a real physical brain, ether
biological or electronic.

Conceptually squaring a number requires no energy, does 64 contains more
energy than 8? NO. But placing 8 marbles on each axis and filling in the
square requires energy.

Just a few thought about the subject.

Mauro

"MG" <[Only registered users see links. ]> wrote


I liked your comments very much but here I would just point out that to the
extent that information has any impact on "reality" it must become embedded
in a physical system. To the extent that information is embedded (eg. in a
series of switches) it does have some minimal "physical" properties such as
the polarities of the switches.

"Doune" <[Only registered users see links. ].uk> wrote

can't
sense


What is it about information that makes it valuable to an organism? Maybe
one way to put it is this: for most animals, useful information increases
chances for survival and reproduction. This statistical quantity, this
augmented probability, is "lost" when the corresponding information is
degraded. Therefore, what I will prove for my Nobel Comedy Prize is this:
what is "stored" in information, and what you can get out of it, is not the
statistically measured stuff called "energy" but rather that statistically
evaluated phenonemon called probability of survival and reproduction.

"tadchem" <[Only registered users see links. ]> wrote

entropy,

You cited two Wikipedia articles in your post, but you failed to cite the
one that will make you eat your words:

[Only registered users see links. ]

Check it out. I'll await your apology ;-)

[Only registered users see links. ] wrote:

<snip repost>


Hold your breath, please.

Purple people are *so* amusing to watch, and I haven't tried to trap
purple people eaters in *decades*.

Tom Davidson
Richmond, VA