[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Localization in OOC



> Although these C-like facilities are minimally sufficient, my
> question is, can we do better?  Both Java and C++ provide locale
> classes, and other related classes, that allow multiple locales to
> be defined within a program, rather than relying on a global
> property.  For a quick comparison between these facilities in C and
> C++, see

Why is there a need for multiple locales within a program?  It seems
that the user can only use one locale at a time and the OS only supports
one locale at a time.  Just because Java & C++ do things one way has
rarely been a reason to do it the same in Oberon-2. ;-)
 
> 
> This also brings up the problem of multi-byte character support,
> including Unicode.  Some time ago, there was a little discussion on
> adding the type `LONGCHAR' to OOC, is this still under
> consideration?  If it is added, then localization should be built
> into LONGCHAR related modules.

I believe the result of this discussion was that multi-byte character
support should be added via a module rather than as a built-in type
since there are so many inconsistent standards for multi-byte
characters.  Using a module to provide this support will enable
the same compiler to support whatever standard happens to be in
vogue.  To date, no one has felt it necessary to add such a module
but I may do so in the future when fully supporting the BeOS library
calls.  This OS has multi-byte character support built in from the
ground up.

Michael