CJKV Information Processing, 2nd Edition

HTML—HyperText Markup Language

|

715

ere are three ways to specify charset information in CSS, specically the in-le method

(similar to the use of the

<META> tag in HTML), via the HTTP header, and from the refer-

ring document. Interestingly, the HTML standard states that the HTTP header has higher

precedence than the

<META> tag declarations, and contemporary web browsing applica-

tions adhere to this.

Automatic encoding detection issues

Most contemporary browsers provide a feature that automatically detects what encoding

is being used, and the algorithms that perform this have improved signicantly. EUC-JP

and Shi-JIS encodings used for Japanese text can oen be dicult to dierentiate. ere

is a useful trick that can help browsers more reliably detect the encoding by including an

HTML comment near the top of the HTML le. is comment includes a single charac-

ter whose encoding is unambiguously either EUC-JP or Shi-JIS encoding. I suggest the

following two meaningful characters for this purpose:  (tōkyō, meaning “Tokyo”).

e character codes for these two characters are <5 > and <5 > for EUC-JP, and

<93 8> and <8 9> for Shi-JIS. Note how the second kanji is unambiguously either

EUC-JP or Shi-JIS. e following is an example HTML comment that includes these

two kanji:

<!---  used for correct automatic encoding detection --->

is technique provides a reliable backu