On (20 Mar 97) Herman Schonfeld wrote to Jerry Coffin...
 HS> I've seen some of the strangest graphics format compressions. One
 HS> which I saw uses huffman encoding, and I must admit, it is pretty
 HS> tricky.  If you've ever seen a huffman coded bitmap you'd be surprised
 HS> by the fact that it uses Binary Trees.
FWIW, all Huffmann coding is based on binary trees, regardless of what
you're actually encoding.
 HS> LZW compression is also pretty good, but the best overall I think is
 HS> arithmetic compression.
Arithmetic compression does work extremely well.  However, it has a
couple of shortcomings.  First of all, it's considerably slower than the
more common LZ and Huffman based compression schemes.  Worse, nearly the
only practical method of implementing it is patented by IBM, so if you
use it in a commercial product, you have to be prepared to pay
royalties.  AT&T also holds a patent on an improvement in IBM's method,
so to use that you pay royalties to both. (Actually, I don't know if
that's still held by AT&T, or if it went to Lucent when they broke up.)
    Later,
    Jerry.
... The Universe is a figment of its own imagination.
--- PPoint 1.90
---------------