🚀 go-pugleaf

In article <12950089@hpspdla.spd.HP.COM> ric@hpspdla.spd.HP.COM (Ric Peregrino) writes:
>
>Shooting from the hip I'd say information has the same units as entropy,
>which I thought was unitless. But what about Gibbs free energy:
>
>dG = dH + TdS
>
>S then has units of energy/temperture? Maybe I'm just being dense?

You'll have to check your specific gravity yourself.  The usual
definition of entropy a la stat mech is S = k ln Omega, where Omega is
the number of microstates in the given macrostate, a dimensionless
quantity.  I believe this formula appears on Boltzmann's grave, by the
way.  k here is his constant, which appears in the formula for the
probability that a state of energy E will be occupied at temperature T, 

exp(-E/kT)

(suitably normalized).  So since things in exponentials have to be
dimensionless, k has units of energy/temperature, so yes, you're right,
S has units of energy/temperature.  Information however is dimensionless
(as are "bits"); in Boltzmann's formula the information required to say
which microstate the macrostate is really in is ln Omega.  (If we use
base 2 log here the way computer scientists do, we get the number of
"bits".  If we use natural log, as Boltzmann did, this number is the
number of "nits".  Wasting energy arguing over small pieces of
information is therefore known as "nitpicking."