CJKV Information Processing, 2nd Edition

294

|

Chapter 4: Encoding Methods

Other Types of Encoding Repair

As opposed to being reencoded by a well-understood and proven transformation, some

data can become truly damaged in the sense that information is removed, either in terms

of entire bytes, or the values of specic bits. Such data appears as garbage, meaning it is

unreadable. is is referred to as mojibake ( mojibake) in Japanese.

We rst discuss the repair procedure for ISO-2022-JP–encoded les as an example of how

encodings may become damaged by having key characters removed, and then repaired.

In the past, one might have received Japanese email messages or attempted to display

Japanese articles from Usenet News, which had their “escape” characters stripped out by

unfriendly email or news reading soware. Sometimes the escape characters are simply

mangled—converted into a single “space” character (2) or into Quoted-Printable (dis-

cussed previously). is was a very annoying problem because one usually threw out such

email messages or articles, rather than suering through the manual and the somewhat

grueling task of manually restoring the escape characters. For example, look at the ISO-

2022-JP–encoded string in Table 4-91. It is rst shown as it should appear when displayed

properly, and then shown as it could appear if it were damaged, specically with its escape

characters missing. ere are certainly lots ...