hip.txt

来自「My (so called) HiP compression algorithm」· 文本 代码 · 共 64 行

------------------------------------------------------------------
 HiP - A single file compressor
------------------------------------------------------------------
by yoda


version:          beta 1
coder:            yoda
project start:    16th September 2k2
coding language:  C++
compiler:         VS7
E-mail:           [e-mail]
website:          y0da.cjb.net


Prolog
~~~~~~
In the last time I tryed to create an own compression algorithm. So
here's my HiP compression algorithm as console mode utility.
The algorithm is a hybrid of Lempel-Ziv 77 and Huffman encoding (Huffman
+ LZ77 ZIP = HiP).
I don't know what I'll do with this nice peace of code. Maybe I'll
code an archiver or a compression library, when it's better on time.
Who knows?

behind the lines
~~~~~~~~~~~~~~~~
The HiP compression algorithm includes up to now the following features:
- LZ77 style string referencing with lazy evolution mechanism
- implicit RLE (Run-Length-Encoding)
- literal encoding via adaptive Huffman table
- non-bruteforce string comparison

Command line
~~~~~~~~~~~~
	HiP.exe (/[compress[X]/decompress]) (infile) (outfile)
	X - compression level (1..6)

History
~~~~~~~
beta 1 - 20th October 2k2:
- much literals in a row are encoded as literal block
- corrected silly bug that made occur errors if the output
  had been bigger than the input (e.g. archives)
- now using adaptive Huffman model
- killed embarrassing problem of worse compress ratio in
  bigger levels by ripping the highest 3, now 1..6
- compression process can now be cancel via ESC

alpha 3 - 14th October 2k2:
- serveral slight improvements
- hash search algorithm's performance slightly improved
- RLE is now an implicit LZ77 feature - works for any chain element size

alpha 2 - 4th October 2k2:
- little prerelease

Thanks to
~~~~~~~~~
Jeremy Collake, X-Lock, Baader


Please drop comments, improvements, etc...
Visit: y0da.cjb.net