ZIP Compression Methods
The internal header attributes on file and directory lever allow the following compression methods:
- Original data will be embedded directly.
- Modified LZW (Lempel-Ziv-Welch) compression method with partial clearing and dynamical addressing of the table entries.
- Reduced (1 to 4)
- A dictionary based algorithm used for addressing byte sequences from former contents. Finally a variable length coding is applied according to the probability distribution of the codes. This method is scalable in four steps.
- A dictionary based algorithm accessing a 4 or 8 KByte large dictionary. Sequence lengths, addresses and uncompressed symbols (optional) use Shannon Fano codes.
- Tokenizing (reserviert)
- Tokenizing is only part of the parameter list. The implementation is not specified and it is not used by PKWARE.
- A dictionary based algorithm accessing an up to 32 KByte large dictionary. Additionally a Huffman coding will be used for redundancy reduction.
- Equivalent to Deflate, but using an up to 64 KByte large dictionary.
- BZIP2 is only part of the parameter list and not part of the "Appliction Note". PKWARE refers to the internet for further information.
It is not neccessary for a compatible ZIP application to support any of these compression methods for encoding. It is only required that a decoder is able to process them.
< ^ >