re: Unicode

Read SC41-5603-01 International Application Development.
Also get the  1998 DBCS Residency Project Document.

DBCS is NOT UNICODE.

UNICODE (UCS-2)  is ALL double byte.  Your disk usage doubles
but disk is cheap these days.  Applications are easier to code
in UNICODE as the various ways characters are encoded is not
so important to the programmer.

So what is DBCS.  Mostly it is a mix of SBCS and DBCS code
with special characters SO SI to tell the system where DBCS
data starts and ends.  If you have ALL DBCS and NO SBCS data
then you have a state that is similar to UNICODE.

Now when the system detects a SO/SI how does it know what the
16 bit patterns really mean.  Is it Chinese Japanese Korean
Arabic etc etc.  This is where the encoding tables come in.

You attach the encoding table to the file or field so then
you get the correct rendering of the proper character.

IMO DBCS sucks.  UNICODE is the way to go. Having said that
the iseries/AS400?? is not a UNICODE machine, so you will
need to do your application in DBCS.
Now I believe JAVA is all UNICODE enabled so I could be wrong.
W2000 is UNICODE and it handles multilanguages very well
go BILL.

IBM should take the plunge and make the Iseries completely
UCS-2 compliant with the option of having UCS-4 if needed.
Then largely all the CCSID that and DBCS this and SBCS the
other can join the dinosaurs and if IBM does not do this
it could make the iseries join the dinosaurs IMO.

Regards
Frank Kolmann

>date: Tue, 14 Sep 2004 20:25:11 -0500
>from: "Bob Kohlndorfer" <kathryn.kasper@xxxxxxxxxxx>
>subject: Unicode
>
>Hi, everyone;
>
>We have been given an assignment to find out what is involved in making a
generic >application we developed compatible with database files in
'unicode' format.  This >application was developed on a single byte box and
we are ignorant of DBCS and >Unicode.  Can anyone tells us what's involved?
Any hardware changes?  what changes in >our code? etc?  Where should we
start reading?
>
>Thanks for any advice.