This is an old-school C implementation to answer question 4 from tutorial
set 1 from my Information Theory course (2010). The C program writes a text
file and the plotting is done by gnuplot (I used gnuplot 4.2 patchlevel 6).

*** I wouldn't normally write code for this sort of problem in C. ***

I recommend using an interactive, feature-rich programming environment when
exploring new topics. I have also provided a Matlab/Octave answer to this
question as an example.

I provided this implementation because I got some questions from people
attempting to use more bare-bones programming environments. Plain ANSI C
seems like a lowest common-denominator that you could generalize to your
favourite programming environment. If you don't want to leave your
favourite programming language, output to text files and learn how to use
some plotting package that can work from text files. Gnuplot can be ok. Or
of course you could use Matlab/Octave just for plotting, or R, or
python+matplotlib...


The demo will work on most Unix-like distributions out of the box, although
you may have to install gnuplot (also required are a C compiler, bash and
ghostscript). All the dependencies are available for Windows.

       doit  bash script to run the whole demo eventually creating hist.pdf
gnuplot.cmd  gnuplot commands
     hist.c  ANSI C program to generate hist.txt containing bin centers,
             histogram counts, and mean predictions using the central
             limit theorem


Iain Murray, October 2010