How does compression work?  Basically, all compression
programs look for common substrings in a file (we will deal only with
bitstrings here, since that's how computers store their data, so a
file will be viewed as nothing more than a sequence of bits).

[Insert stuff about run-length encoding here, to introduce the
 concepts.]

[Then discuss variable-length encoding here, and close with the
 problem of how to find the most efficient alphabet given frequency
 statistics.  Lead into Huffman encoding from that question.]

[Then talk about other methods, normally used in addition to Huffman
 encoding.]

        The question naturally arises: for a given file, is there a
minimum size to which it can be compressed?  The answer has to by
"Yes", because certainly no file (of size greater than one bit) can be
compressed into a single bit, right?  No practical compression method
actually reaches the theoretical limit, however, because doing so
involves very lengthy computations.  In essence, you'd have to do
this:

        [What the hell *would* you have to do, anyway?]

        Here are some methods commonly used to approach this limit
without taking forever to compress the file:

        [Here they are, come and get 'em...]