On Monday, October 12, 2015 at 3:29:20 PM UTC-4, Harry Potter wrote:
> I thank you for mentioning LZSS.  It's *better* than LZW, *and* it's faster!.
 :D  Let me tweak it some more.  I'm good at tweaking.  :)

I fixed a bug and did some tweaking.  Now, it's almost as good as Deflate.  I
plan to add Vitter Adaptive Huffman code.

I also want to know other tweaks I can apply.  I added:

*  An efficiency check: After I find a LZSS match, I check the next two bytes. 
If they have a better match, I compress that way, instead.
*  I differentiate between small and large LZSS matches.
*  I use as few bits as possible to write a value.
*  If a byte is *not* compressed, I only take into account values already
found.

--- SoupGate-Win32 v1.05