Hi Folks,

I'm having some difficulty reading Howard & Vitter's 'practical
implementations of arithmetic coding' (http://www.gvu.gatech.edu/
~jarek/courses/7491/Arithmetic2.pdf). I'm immediately confused when I
start reading section 3.1 (reduced precision arithmetic coding).
Hopefully someone here can help me figure out what background context
I've managed to miss for reading this?

When they first refer to states, what are they? Initially I assumed
they were the set of all valid intervals in the range [0,n) (which
I've worked out to be (n^2 + n)/2 rather than the 3N^2/16 that they
list as the number of possible states -- For example: assuming n = 2,
I expected that we would have the following valid states [0,1), [0,2),
[1,2). )

In example 4, they say that the table at the bottom is made by
straightforward application of arithmetic coding -- but I have no idea
where that table has come from :-/ (as an aside, presumably the p in
the table is the probability of a 0, but that leaves me confused as to
what the alpha parameter is.).

What background do I need to read this paper, other than a good
understanding of the integer implementation of an arithmetic coder?

Thanks,
John G.