Servus Denis,
Am 05 Apr 97 um 13:06 schrieb Denis Boyles an All:
 DB> Does anybody have any suggestion, or guidlines to what huffman buffer 
size
 DB> is good for what data? I've been using the following:
 DB>   -JH16384
 DB>   -JH65535
 DB> Depending on the files, one way is usually better than the other. Though
 DB> I'm wondering what is best for what. Also, might other sizes be better 
for
 DB> certain data types than other sizes. ie. text files, binary files, etc.
Well, the Huffman buffer size is for making a data table how often a char is 
being used. Normally the biggest buffer size is always the best. But you have 
the problem, that the table is getting bigger and the time is going up. As I 
know the packer is getting slower but the pack rate is going to get up.
That's the theory. Perhaps Robert Jung can say something about his 
implemention of the huffmann algorism. How many chars are put together in the 
char table which is made when using the packer?
                    Gruss   /~\/~\
                           (_/\/\_)artin
... Soviet spy-dancer:  A ballet ruse...
--- FMail/386 1.02
---------------