[Zlib-devel] adaptive extensions?

[Krzysztof Jankowski]

Hi,
I've got a question, but I'm sure I'm not the first one asking it. I
did go through zlib archives, but didn't find any relevant thread
there.

Are you planning for any extensions to zlib format/implementation in
future? I'm thinking in terms of extensions which could make deflate
implementation more feasible for data streaming environments or some
HW implementations, i.e. single pass encoding/decoding or something
like that.

I know one could say, well go for static Huffman coding and you're
there. But some fully adaptive method would provide better compression
results I believe.

Currently RFC1951 specifies BTYPE=0b11 as reserved. Would you ever
consider redefining that to e.g.:

11 - compressed with adaptive X coding

I don't know what X should be. Possibilities are: adaptive Huffman,
arithmetic coding etc. I'm sure patent issues have to be cleared, as
zlib is meant to be a free library.

You might say, well adaptive approaches are usually slow to implement.
That's true but if the '11' option was there, people would have a
freedom of choosing it on per their needs.

What do you think?

Thanks,
Krzysztof