http://input.cpatch.org/code/big5.zip -> big5-1.txt


                        BIG-5 DATA
                        ==========
                (C) Copyright 1993, Guo Jin
                        (guojin@iss.nus.sg)

----------------------------------------------------
(1) BIG-5 double bytes coding
-----------------------------
First byte:     0xA1 - 0xF9
        symbols:        0xA1 - 0xA3     ( 3 sectors)
        common hanzi:   0xA4 - 0xC6     (35 sectors)
        undefined:      0xC7 - 0xC8     ( 2 sectors)
        rare hanzi:     0xC9 - 0xF9     (49 sectors)
        total defined:  3+35+2          (87 sectors)
Second byte:    0x40 - 0xFE
        part one:       0x40 - 0x7E     (63 codes)
        undefined:      0x7F - 0xA0     (34 codes)
        part two:       0xA1 - 0xFE     (94 codes)
        total defined:  63+94          (157 codes)
coding space:   87 x 157 = 13659 codes

(2) BIG-5 three sections
------------------------
symbols: 
        full:   3x157=471 codes
        undefined: 0xA3C0 - 0xFE        (47 codes)
        total defined:  471-47         (424 codes)
common hanzi:
        full:   35x157=5495 codes
        undefined: 0xC6A1 - 0xC6FE      (94 codes)
        total defined:  5495-94       (5401 codes)
rare hanzi:
        full:   49x157=7693 codes
        undefined: 0xD6 - 0xFE          (41 codes)
        total defined:  7693-41       (7652 codes)
defined char:
        pure hanzi:     5401+7652=13053
        all:    424+13053=13477

(3) original BIG-5 coding scheme
--------------------------------
        warning:
                the coding scheme is quite different from
                its implementation listed above.

        o  code:        double bytes
        o  each 7-bit byte: 0x21-0x7E
                to avoid ASCII control codes (0x00-0x1F),
                white space (0x20), and delete (0x7F)
        o  plane:
                Byte1-bit8      Byte2-bit8      no.     plane
                        0               0       0       ASCII
                        0               1       1
                        1               0       2       first
                        1               1       3       second
        o  symbols: (first plane)
              * section 0x21 - 0x41 (code 0x2121 - 0x417E)
                with (0x41-0x21)x94=3102 codes, but only
                651 defined.
              * since 651 < 7x94=658, only
                7 sectors (0x21 - 0x27) really defined.
              * for the last sector really defined, there are
                7=658-651 codes undefined.
        o  control codes:(first plane)
              * section 0x42 - 0x43 (code 0x4221 - 0x437E)
                with (0x43-0x42)x94=188 codes, but only
                33 defined.
              * since 33 < 1x94=94, only
                1 sector (0x42) really defined.
              * 61=91-33 undefined for the sector 0x42.
        o  first level hanzi:(first plane)
              * section 0x44-0x7D (code 0x4421 - 0x7D4B)
                with (0x7D-0x44)x94=58x94=5452 codes, but only
                5401 defined.
              * 51=5452-5401 undefined for the last secotr.
        o  second level hanzi: (11: seconde plane)
              * section 0x21 - 0x72 (code 0x2121 - 0x7244)
                with (0x72-0x21)x94=82x94=7708 codes, but only
                7652 defined.
              * 49=7708-7652 undefined.
        o  derived data:
                pure hanzi: 5401+7652=13053
                symbol+pure hanzi: 651+13053=13704
                all sectors defined: 7+58+82=147
                codes undefined in the defined sectors: 147x94-13704=107

(4) references
--------------
        Huang, Jack K. T. and Huang, Timothy D.,
        Republic of China's Big Fives. in
        An Introduction to Chinese, Japanese and Korean Computing,
        Chapter 5 Section 6.  World Scientific, 1989.

(5) tools for BIG-5 code table
------------------------------
        o  to display:
        use any BIG-5 Chinese environment.
        (for example, KC for DOS, ftp-able form
        NCTUCCCA.edu.tw [140.111.3.21]
        under /Chinese/DOS/Chinese-Systems/KC.)

        o  to print:
        use any BIG-5 print program.
        (for example, hz2ps for UNIX,
        cnprint for DOS, UNIX and VMS, both ftp-able from
        ifcss.org [129.107.1.155])

(6) C source code
-----------------
/*********************************************************************
*
* big5-table.c
* ============
*
* make: gcc big5-table.c -o big5-table
* usage:big5-table >! big5.dat
*
* author: Guo Jin (guojin@iss.nus.sg)
* date: Oct. 11, 1993
*
* function: to construct a BIG-5 coding table for easy reference.
*
* algorithm: simply list all valid BIG-5 codes (including a few undefined,
* see also BIG-5 coding scheme above) in a section by section format.
*********************************************************************/
#define CHAR_PER_LINE   16
main()
{
        unsigned char   sec, off;
        printf("%s", head);
        for(sec=0xA1; sec<=0xF9; sec++) {
                for(off=0x40; off<=0x7E; off++) {
                        if (!((off-0x40)%CHAR_PER_LINE)) {
                                printf("%2x%2x\t", sec, off);
                        }
                        printf("%c%c ", sec, off);
                        if (!((off-0x40+1)%CHAR_PER_LINE)) {
                                printf("\r\n");
                        }
                }
                                printf("\r\n");
                for(off=0xA0; off<=0xFE; off++) {
                        if (!((off-0xA0)%CHAR_PER_LINE)) {
                                printf("%2x%2x\t", sec, off);
                        }
                        if (off==0xA0) printf("%c%c ", 0xA1, 0x40);
                        else printf("%c%c ", sec, off);
                        if (!((off-0xA0+1)%CHAR_PER_LINE)) {
                                printf("\r\n");
                        }
                }
                                printf("\r\n");
                                printf("\r\n");
        }
}