Perhaps "Perl Unicode" and "UTF8" are not the same thing as a "Wide String" as refered to by the Win32 API?

Unicode includes more than one encoding ("transformation format"). Perl uses UTF-8. Windows calls UTF-16 "UNICODE" or "wide" (or close to UTF-16 -- I don't think Windows supports multi-word characters which are part of UTF-16; [that is, Windows "UNICODE" appears to use only fixed-width 16-bit characters]). Windows calls UTF-8 "multi-byte".

Update: And Perl doesn't support UTF-16 directly (there may be modules that deal with it), which could also be expressed as "Perl doesn't support multi-word nor single-word characters" using some of my loose terminology from above.

[Updated]