Aprial 11th Wednesday (四月　十一日　水曜日)

最新推荐文章于 2018-08-09 16:39:32 发布

原创最新推荐文章于 2018-08-09 16:39:32 发布 · 684 阅读

0 ·

CC 4.0 BY-SA版权

文章标签：

#compression #semantic #dictionary #algorithm #encoding #comments

本文分享了作者花费四天时间研究LZ77算法源码的经历。文章详细介绍了LZ77算法的工作原理，包括其与K&M字符串匹配算法的相似之处，以及压缩过程中的动态字典更新机制。此外，还探讨了源码中一些晦涩难懂的部分，并记录了作者为理解这些部分所做的努力。

Today I finished reading the source of LZ77 algorithm. The most of the source was understood by me. I took
about 4 days to studying. Although the main frame of lz77 algorithm is not hard to understand from the related
to articles. However, no sooner had I seen that source files than I began to forget what I did. There are little
comment in this source. Many time I have to detect what those codes want to do.

The principle of lz77 is similar to the the principle of K&M string match algorithm. It is the main task to
compare those encoded phrases with those waiting encoding phrases. The whole compressing procedure is a dynamic.
Those encoded phrase set is called a dictionary. At the beginning of compression, this dictionary is empty.
With the process of compression, more and more phrase is added into.

The function output a gamma code also puzzled me once. Some codes is used to work out exponent for 2. The
author, obviously, seems to be a hack. The semantic of some codes is not trenchant.

So, I have to add comments many where. Finally, I adjusted the source as my wish, and build it. The binary
program can run correctly. I tested it by a big file about 4.8mb. The speed of compression is not quick, but
decompress its compressed file so quickly.