CJKV Information Processing, 2nd Edition

Miscellaneous Algorithms

|

593

}

}

else if (c <= 159)

whatcode = SJIS; /* Shift-JIS detected. */

}

else if (c >= 240 && c <= 254)

whatcode = EUC; /* EUC-JP detected. */

else if (c >= 224 && c <= 239) {

c = fgetc(in); /* Read next byte to c. */

if ((c >= 64 && c <= 126) || (c >= 128 && c <= 160))

whatcode = SJIS; /* Shift-JIS detected. */

else if (c >= 253 && c <= 254)

whatcode = EUC; /* EUC-JP detected. */

else if (c >= 161 && c <= 252)

whatcode = EUCORSJIS; /* Ambiguous (Shift-JIS or EUC-JP). */

}

}

}

return whatcode; /* Return the detected code. */

}

Appendix C provides Perl code for a much more exible way to automatically detect the

encoding of CJKV text les, not only those for Japanese. at Perl code shows how power-

ful regular expressions can be when used in specic contexts.

Half- to Full-Width Katakana Conversion—in Java

It sometimes is necessary to convert half-width katakana to their full-width counterparts.

is is most useful as a lter to ensure that no half-width katakana characters are included

within email messages. It is also useful when you need to move les from one platform to

another and the new platform does not support half-width katakana characters. Example

usage of this Java method is as follows:

String half = "\uFF76\uFF9E";

String full = KatakanaFilter.halfToFullWidthKatakana(half);

ere is no simple conversion algorithm that you can use to accomplish this ...