( The Canterbury Corpus )
home | purpose | summary | details | corpora | methods | related | credits | faq

Description of the char method

Note: this is not a compression method. It is simply a testbed for arithmetic coding with a very simple model. To see how well arithmetic coding can perform in real compression systems, see the results for methods such as ppm and ppmz. Eventually the results for methods such as "char" that focus on coding will be put into a separate section from the "real" compression methods.

Char is a very simple order zero character model driving an arithmetic coder.

Code and documentation is readily available. It includes two other models, a word based model and a bit based model.

The authors are John Carpinelli, Alistair Moffat (alistair@cs.mu.oz.au), Radford Neal (radford@ai.toronto.edu), Wayne Salamonsen, Lang Stuiver (langs@cs.mu.oz.au), and Ian Witten (ihw@waikato.ac.nz).

This page last updated Monday, January 08, 2001 by Matt Powell Department of Computer Science University of Canterbury