RE: More Stream File and DBCS issues...

I'm in no way an expert on this stuff, Brad, so take it with a grain of
salt, but in general UCS-2 values require two bytes, and in those two bytes
DBCS data gets converted to code points that don't clash with standard
single-byte data.

For the most part, standard ASCII is mapped to 00xx, where xx is the
original ASCII hex code.  DBCS data, on the other hand, gets converted to
values > 256.  Thus you can't have collisions.

Note that UCS-2 is ALWAYS 16 bits per character.  Meanwhile, UTF-8 and
UTF-16, the other popular Unicode format, can contain characters of up to
four bytes which allow for over a million characters.  UTF is way confusing
for me, since it also allows for special "combining characters" (a way of
specifying, for example, an umlaut along with a U).

Joe

From: Brad Stone

In a nutshell, it sounds like this may convert all
characters, either SCBS or DBCS to unique values so that
there is no way a DBCS value can contain a byte that is
equal to a SBCS value and this problem would be eliminated?