The Great Lempel-Ziv vs. Arithmetic Coding Debate

In order to do something different the next class will be arranged in the form of "The Great Lempel-Ziv vs. Arithmetic Coding Debate". Therefore participation to this class is obligatory except for a very good reason that should be mailed beforehand to Teemu and myself. This debate will happen Thursday April 11 from 4:15PM-6:30PM at HIIT premises (HTC). For the debate, half of the groups will be defending the use of arithmetic coding and the rest universal coding based on Lempel-Ziv coding. The point is to argue about the use of the methods from pragmatic point of view, i.e., why in practice one should prefer one method over the other. The debate will be structured as follows:

1) The defending groups will select among themselves people to give a brief 10 minute introduction (2-3 transparencies) to the method in question.

2) Both parties will present a list (5 items or more) of reasons why their method is better than the competing one.

3) A discussion will be intiated and hopefully lively continued.

There will be pizza and (soft) drinks available and the winners get even more pizza. The group assignments are as follows, and should not be changed without strong supporting argumentation:

Lempel-Ziv:
  - Team Aunimo
  - Team Haapasalo
  - Team Heino
  - Team Paivaniemi

Arithmetic coding:
  - Team Kaariainen
  - Team Makela
  - Team Pitkanen

Good starting points as reading material for the debate can be found in

ť MacKay's draft, Chapter 6
ť Lempel-Ziv in: Loben Chiang's thesis draft
ť data-compression.com: Lossless Data Compression
ť Steve Blackstock: LZW and GIF explained
ť Arturo San Emeterio Campos: Arithmetic coding
ť Mark Nelson: Arithmetic Coding + Statistical Modeling = Data Compression (Dr. Dobb's Journal, Feb. 1991)




3 Concepts: Information	2002