Lecture 7

Chapter 5

· Data Compression – Makes files smaller

· Advantages/Disadvantages

· Uses less external storage

· Read/Write time faster to device with smaller size

· Fewer seeks/rotational latency with fewer tracks

· Requires more CPU time to compress/decompress

· Compact notation – Redundancy reduction

· Example

· State names require about 14 bytes (max)

· State abbreviations require 2 bytes

· State sequence numbers require 6 bits

· Costs

· Binary codes can’t be read (easily) by humans

· Encoding/decoding costs time

· Must include encode/decode logic in all programs requiring state names

· Is the cost worth it? It depends!

· Suppressing Repeating Sequences

· Run-length encoding

· Include run-length indicator, length value and target byte

· Can result in longer string than original (if not careful)

· Variable-length Codes

· Use shortest encodings for most frequently occurring characters

· E.g. Morse Code – “e” and “t” get . and - respectively

· Huffman Codes

· Instantaneous Code – You know you are at the end of a coded character without having to examine the next character

· No code is a prefix for another code

· The characters with the highest probabilities of occurrence are assigned the shortest codes

· Can be codes using a binary tree (see if you can figure out the algorithm!)

· Irreversible Compression Techniques

· Some information is lost and thus the original cannot be accurately reconstructed

· Voice, pictures, etc.

· The Master File Update Problem

· Given:

· a sorted old master file

· a sorted transaction file with

· adds, changes and deletes

· Output a sorted new master with the transactions applied to the old master file

· Write the pseudo-code for this algorithm