Hitomi is a huffman encoder/decoder, written in huffman encoded bytecode.

    Hitomi takes two arguments, first argument is the input file and second
    argument is the output file.  Depending on the content of the input file,
    four things can happen:

    1. Input file is already encoded   Output file is decoded from input file.
                                       "dec" is written to stdout.
    2. Input file is not encoded       Huffman compression is applied to input
                                          file to produce output.
                                       "enc" is written to stdout.
    3. Input file is empty             Empty output file is created.
                                       "err" is written to stdout.
    4. Input file fails to open        "err" is written to stdout.

    Would be nice to have more descriptive output than "enc" or "dec", but I
    was really pressed for space.  Also, somewhere in the coding process I
    messed up, so that Hitomi is *only* guaranteed to compress text properly.
    Encoder works okay but decoder fails to recognize the header, so that
    non-text files can't be decoded :(

    Hitomi takes a long time to encode a file (about 3:30 for 250K).  After
    all, you can't expect much performance from a 7000 bit program, executed
    one bit at a time.  IOCCC places strict restrictions on file size but no
    limit on execution time, so I figured...  ^_^;;

    Program compiles with few warnings (string too long, undefined functions),
    and probably will not compile properly with anything other than gcc.