File: /~heha/hs/dos/doslfn/doslfn.zip/TBL.TXT

Description of Volkov Commander table and DBCS extended table format

1..32?	BYTE	Description of table (for displaying in VC)
2	BYTE	0x0D,0x0A, =Terminator
1	BYTE	Table Format Designator (0, 1, or 2)

SWITCH (Table Format Designator)
 CASE 0		SingleByte to SingleByte conversion table (not used by DOSLFN)
256	BYTE	other character codes for characters 0x00 to 0xFF

 CASE 1		SingleByte to 16bit Unicode conversion table
128	WORD	Unicodes for characters 0x80 to 0xFF

 CASE 2		DBCS to 16bit Unicode table (not defined by Volkov)
1	BYTE	lowest TrailByte (>=0x40)
1	BYTE	count of TrailBytes per LeadByte (<=0xC0) M
128	WORD	[0x0000		unused (reserved) LeadByte
		[0x0001-0x007F	LeadIndex into following 1-based DB table
		[0x0080-0xFFFE	Unicode for Single Byte Character
		[0xFFFF		unused Single Byte Character U+0xFFFF
		The DB table following is made of columns indexed by
		(TrailByte - lowest TrailByte).
		The rows of each column M = see above
		The number of columns N = (Maximum Lead Byte Index +1)
N	M WORD	Unicodes for double-byte characters (=DB table)
		Not assigned code points default to Unicode U+0xFFFF


The DBCS=>Unicode table is defined to be reasonably compact.
Conversion to Unicode can be done by one or two table accesses.
Backward conversion requires only a single "repne scasw" assembly
instruction, that's the fastest way for such a small table. The index
result must only be subtracted and divided to get the DBCS characters.

Typical sizes for this table are:
SJIS	CP932	17.3K	"Shifted" Japanese Industry Standard
GB2312	CP936	15.5K	Public Republic of China's "small" standard
			(GB stands for Guojia Biaozhun = National Standard)
GBK	CP936	48.4K	"full" standard (K stands for Kuozhan = Extended)
			Note that GB2312 is a true, code-compatible sub-set of
			GBK, containing the most often used 7500 characters.
			Both GB2312 and GBK use Simplified Chinese
Korea	CP949	47.4K	"Extended Wansung" ???
BIG5	CP950	33.5K	Traditional Chinese used in Taiwan (and Hong Kong,
			maybe yet)

All tables fit into DOSLFNs memory management, at least it just fit to
leave 1K free in CDROM mode.
Maybe there is a reduced code table for Korea's CP949 too, please feedback!
GBK seems to be not used in DOS, it's for Windows.
At least the screen redirector "TechWay SCS V3.2" does only support GB2312.
Detected encoding: ASCII (7 bit)2