🚀 go-pugleaf

edf@sm.luth.se (Ove Edfors) writes:
|  Is there anyone out there, who have information on fast decoders for
| variable-length codes, e.g. Huffman codes.

My program "huf" quickly decodes Huffman codes.  It can be ftped from
ftp.cs.toronto.edu:/pub/hugh/huf*.

It uses a static code table, i.e. the code is the same for an entire
file.  This allows the decoder to "leap" into the code tree based on
the first n bits by using an auxilliary table with 2^n entries.

One reason that this is effective is that huf uses words as symbols,
not ASCII characters.  Thus each symbol is represented by a larger
string of bits.

I haven't thought hard about whether this technique would work well
with an adaptive Huffman program.

Hugh Redelmeier
hugh@mimosa.com or {utcsri, yunexus, utzoo, scocan}!redvax!hugh
When all else fails: hugh@csri.toronto.edu
+1 416 482-8253