What do you think?
Rate this book
The Data Compression Book
The Data Compression Book Second Edition The Data Compression Book is the most authoritative guide to data compression techniques available. This second edition has been updated to include fractal compression techniques and all the latest developments in the compression field. All the code in the previous edition has been updated to run with today's compilers and has been tested on multiple platforms to ensure flawless performance. You'll learn to write C programs for nearly any environment as you explore different compression methods. Nelson and Gailly discuss the theory behind each method and apply the techniques involved to shrink data down to a minimum. Each technique is illustrated with a complete, functional C program that not only demonstrates how data compression works, but it also can be incorporated into your own data compression programs. You'll also get detailed benchmarks demonstrating the speed and compression ability of each technique. The code in this book has been tested on a variety of platforms and compilers including Microsoft Visual C++ 1.5 with MS-DOS 5.0 and 6.22; Borland C++ 4.0 and 4.5 with MS-DOS 5.0 and 6.22; Symantec C++ 6.0 and 7.0 with MS-DOS 5.0 & 6.22; Interactive Unix System 3.2 with the portable C compiler; Solaris 2.4 with the SunSoft compiler; and Linux 1.1 with the Gnu C Compiler. Topics Include:
576 pages, Paperback
First published August 1, 1991
1 person is currently reading
72 people want to read
About the author
Mark Nelson
221 books13 followersLibrarian Note: There is more than one author in the Goodreads database with this name.
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 3 of 3 reviews
May 1, 2015
An enjoyable read, treating the problem of data compression at more or less the level I wanted it to. Has its share of rough edges, though.
I appreciate that the authors made the effort to avoid pseudo-code and provide working C code for every algorithm, but because everything sucked so much more in 1995 and because Nelson wanted DOS support, that code tends to have a lot of boilerplate noise; combined with the fact that whole-program implementations are provided for everything (including at least one program pretty unrelated to compression), something like half of the book is just code. That's a bit excessive, especially considering that the included floppy disk has another copy of that code anyway (presumably; I'm not even sure I own a working 3½-inch drive so I didn't check).
Another problem is that the focus is very much on how, and not so much on why. Sometimes the why is obvious—it's clear why Shannon-Fano and Huffman coding tend to compress data, and why Huffman is better at it than Shannon-Fano, for example—but when it's not, the authors make no attempt to explain—though I could certainly write my own implementation now, I still don't have a decent feel for why arithmetic coding compresses at all.
In the end, The Data Compression Book really doesn't deserve its reputation as the book on data compression, but if you just want an easy-to-follow discussion of the main algorithms that are unfiddly enough that you could conceivably want to implement them for fun, you could certainly do a lot worse.
(To be perfectly honest, though, you could probably also do better with just its table of contents and Wikipedia.)
I appreciate that the authors made the effort to avoid pseudo-code and provide working C code for every algorithm, but because everything sucked so much more in 1995 and because Nelson wanted DOS support, that code tends to have a lot of boilerplate noise; combined with the fact that whole-program implementations are provided for everything (including at least one program pretty unrelated to compression), something like half of the book is just code. That's a bit excessive, especially considering that the included floppy disk has another copy of that code anyway (presumably; I'm not even sure I own a working 3½-inch drive so I didn't check).
Another problem is that the focus is very much on how, and not so much on why. Sometimes the why is obvious—it's clear why Shannon-Fano and Huffman coding tend to compress data, and why Huffman is better at it than Shannon-Fano, for example—but when it's not, the authors make no attempt to explain—though I could certainly write my own implementation now, I still don't have a decent feel for why arithmetic coding compresses at all.
In the end, The Data Compression Book really doesn't deserve its reputation as the book on data compression, but if you just want an easy-to-follow discussion of the main algorithms that are unfiddly enough that you could conceivably want to implement them for fun, you could certainly do a lot worse.
(To be perfectly honest, though, you could probably also do better with just its table of contents and Wikipedia.)
August 2, 2012
One of the first books on information theory and programming I read. I remember that it blew my mind at the time.
August 5, 2022
An awesome introduction to techniques used for compression, loved it
Displaying 1 - 3 of 3 reviews