Automatic Japanese transliteration with a formalism for presyntactic analysis

A Japanese transliteration system, primarily used for the romanization of Japanese dictionary entries, is presented. As the system was implemented in a quick and dirty way, we show how to reimplement it cleanly and efficiently with the help of the Sumo formalism, a framework for the development of presyntactic applications. We describe Sumo, its foundation and control language, and outlines the implementation of the romanizer using the Sumo tool.