Using an IME

Now that we have a basic understanding of how transliteration works, we can discuss how someone actually inputs East Asian text, using Japanese as an example. Japanese text input is divided into three phases: conversion/transliteration, dictionary lookup, and candidate selection. During the conversion phase, a user inputs Latin characters that correspond to Kana characters. As each Latin character is typed the IME automatically converts them into either Hiragana or Katakana characters. For example, if we type the Latin letters 'ame' we obtain the Hiragana letters Hiragana letters .

In the dictionary lookup phase, conversion dictionary takes Kana characters, as input and converts them into suitable mixtures of Kanji and Kana characters. The dictionary uses a key, which is commonly known as a 'reading', or phonetic pronunciation, to find acceptable replacement Kanji characters. Typically, each reading or key may have several Kanji associated to it. In addition, many Kanji may have multiple readings. For example, the Kanji character for 'rain' rain has two possible readings ame Latin letters 'ame' or u Latin letter 'u'.

During the candidate selection phase the user is presented with a list of Kanji/Kana characters to select after a sequence of Kana characters has been input. The list of replacements is typically prioritized by most frequent use. For example, in Figure 6, the Hiragana letters Hiragana letters correspond to several Kanji/Kana characters. In this example the first choice is the input Hiragana characters while the last choice is the input characters represented in Katakana. Choices two through five are Kanji characters that all have the same reading. In particular, choice two is the Kanji character for 'rain', choice three is the Kanji character for 'candy', choice four is the Kanji/Kana characters for the verb 'to sew', and choice five is the Kanji character for 'heaven'.

After the user selects a candidate from the candidate list (by either typing the number of the choice, navigating to the choice with the arrow keys or by clicking the choice with the mouse), the characters are automatically inserted into the application. For convenience, most IME's automatically insert the most frequently used replacement into the application. If the IME selected the incorrect replacement, the user can force the replacement list to be displayed. Once the character or characters have been selected, the entry process starts over again.

Figure 6: Candidate selection list.

Candidate selection list

Contact IBM

Need assistance with your globalization questions?

Topic contents