> In any case, it should be task for database-communitaction layer, i.e.
> IBO to use those facilities and give you the string converted to
> system character set.

Sorry, maybe that is simple terminology issue, but the IBO must
provide conversion into Unicode-string ("wide string"), not the system
charset - applications must be able to correctly process data in other
charsets, not only in the system charset. Latter is useful only when
using non-unicode widgets in the GUI.

Roman