There are five kind of characters in Japanese text files.
1. Kanji ( Chinese(Han) characters ):
2. Katakana ( hard phonetic ) :
3. Hiragana ( easy phonetic ) :
4. Gaiji ( user-defined characters )
5. Rohmaji ( Roman characters ) : abcABC123!"#...
In Shift-JIS code whitch is widely used on PC, they are classified intotwo classes.
1. Single-byte characters ( Hankaku(half width)-Katakana, Rohmaji )
2. Double-byte characters ( Kanji, Zenkaku(full width)-Katakana, Hiragana,Gaiji)
The first bytes of double-byte characters are 81-9F or E0-FF.
The second bytes are 40-FC.
Hankaku-Katakana is A0-DF
To identify if it is a first byte of a double-byte character or a second-byteofadouble-bytecharacter or a single-byte character, you should scanfromthetop of thetext or from the point where it is obvious.
For example a double-byte character '
' consists of 83 and 41. But it does not mean that consecutive 83 and 41alwaysmeans''. 83 can be the second byte of a double-byte character and 41 is a single-byte'A'.So,itis not easy to find/replace/convert japanese character strings.Itisalsodifficust to word-wrap (actually, we do character-wrap) Japanesesentences.[Input method]
To input Japanese text, we use very intelligent translation programs. OnDOS,translationprogramsare called FEP ( Front End Processor ). On MS-Windows,theyarecalled IME( Input Method Editor ). Some use MS-IME which comeswithWindows.Unfortunately,MS-IME is not very intelligent. So there aremanymore intelligentIMEs likeATOK, WX3, WP(WordPerfect)-IME and so on.Someof them comes withJapaneseword-processors.
When I input a text 'My name is Yoshinori Saigusa.', I input like this.
watashinonamaehasaigusayoshinoridesu
On the display, I will find a Japanese text transrated and displayed reversedlikethis.
Usually, you can query other candidates and make a choice. Your choicewillberegisteredin your dictionary and considered next time. You canuse yourpersonaldictionaryand grow it up. You may have several dictionariesforeach kindof job.
Unfortunately, my first name and my family name is not computer-friendly.SoIshouldcorrect it in a slightly time-comsuming manner, and registeritforthe future.
