Japanese language in EBCDIC

Several mutually incompatible versions of the Extended Binary Coded Decimal Interchange Code (EBCDIC) have been used to represent the Japanese language on computers, including variants defined by Hitachi, Fujitsu, IBM and others. Some are variable-width encodings, employing locking shift codes to switch between single-byte and double-byte modes. Unlike other EBCDIC locales, the lowercase basic Latin letters are often not preserved in their usual locations.

The characters which are found in the double-byte Japanese code used with EBCDIC by IBM, but not found in the first edition of JIS X 0208, also influenced the vendor extensions found in some non-EBCDIC encodings such as IBM code page 932 ("DBCS-PC") and Windows code page 932.

Single-byte codes

Similarly to JIS X 0201 (itself incorporated into Shift JIS), Japanese EBCDIC encodings often include a set of single-byte katakana. Several different variants of the single-byte EBCDIC code are used in the Japanese locale, by different vendors; a given vendor may also define two different single-byte codes, one favoured for half-width katakana and one favoured for Latin script. Variants of EBCDIC favoured by a given vendor for use for katakana are sometimes referred to as EBCDIK, standing for Extended Binary-Coded Decimal Interchange Kana code.

Code pages incorporating half-width kana are an exception to IBM's EBCDIC invariant character set, which specifies a set of characters which are usually encoded the same across all EBCDIC code pages. Most notably, they sometimes include katakana characters at code points which are used for lowercase letters of the Basic Latin alphabet in the invariant set. Encoding of lowercase letters when katakana characters are included at those locations, and encoding of katakana characters when lowercase letters are retained in their usual locations, can vary between vendors, as shown below.