Re: ideas on "reversing" the process for deflate code?

 There are two decisions you have to make when deflating: 

1. how big a chunk of data should I try to deflate at once?

2. should I create my own custom huffman table or use the built-in one?

 You can simplify things by always choosing, say,  a 2k block to compress
and by always using the built-in huffman table (or emiting a noncompressed
block if compressing with the built in table resulted in a growth
rather than a shrinkage).   This is sub-optimal
but would still give you some compression.