Alex,

> I followed your recipe and the result looks very promising (see second 
> attachment). I'm not sure if all the codes should generate displayable 
> characters,

  I think the all characters in arial.ttf in that range are shown
there.

>  but the ones that I know from Latin-2 (ISO-8859-2) are sure 
> in there. On XP I'm using a german keyboard and I switch to a serbian 
> keyboardscheme whenever I need to enter Latin-2 characters. Entering ć 
> (0107) in that way inside the MultiArial string results in a ae ligature 
> character (see first attachment).  This usually only happens with other 
> applications when I try to enter a Latin-2 character while they are 
> using Latin-1.
> A character (Cut from) & (Paste into) the MultiArial string results in a 
> DNU SmallInteger>>asciiValue.

  You have to write the keyboard input interpreter and clipboard
interpreter.  See the subclasses of InputInterpreter and
ClipboardInterpreter.

  The Windows VM just passes the bytes to the image, so the image
should be aware that in what encoding those keycode are.

-- Yoshiki