It's DUDE-8! It's quick! It's easy! It does it all!
The DUDE-8 algorithm compresses 21-bit Unicode Scalar Values into bytes.
It is based on the DUDE(-6) algorithm currently being proposed for
i18n of DNS names.
1. Let prev = 0.
2. For each Unicode Scalar value usv, let prev = prev xor USV.
3. Let the value of prev be expressed in 21 bits as xxxxxxxyyyyyyyzzzzzzz.
4. Encode prev in 3 bytes as 0xxxxxxx 0yyyyyyy 1zzzzzzz.
5. Emit all non-zero bytes.
6. Repeat until done.
The optional signature of DUDE-8 is 03 FD FF, which is how U+FEFF appears at the
beginning of a DUDE-8 compression stream.
DUDE-8 is simpler and simpler than SCSU, but doesn't allow recovery from
garbles or even partial random or backwards access.
--
John Cowan [EMAIL PROTECTED]
One art/there is/no less/no more/All things/to do/with sparks/galore
--Douglas Hofstadter
