Kuan-ming Lin
GS 321A homework (commented by Janet Maceda)
Feb. 9, 2005. All Rights Reserved.

Data Compression

  Data compression is defined as the process of encoding an input 
into a compact sequence which reserves sufficient information. 
It can be categorized into two classes: lossless and lossy. 
Lossless data compression guarantees that the original message 
can be exactly reconstructed from the compressed sequence.
The popular ZIP file format is a typical example in this class.
Other popular lossless compression schemes include RAR, gzip, bzip
(all used for compressing files), and PNG (used for compressing
images). In contrast, lossy compression does not necessarily 
reserve all information; therefore, the decompressed file is an 
approximated yet nearly indistinguishable duplication. Many 
multimedia compression standards, e. g. JPEG for images and MPEG
for video/sound, employ the lossy compression scheme, not only
because it yields shorter compressed stream, but also because
the small loss of quality is usually not perceivable.