L2/01-092

2001-01-30 UTC Meeting

Sandra O'Donnell additional text 

A limitation of the ISO/ANSI C model is its assumption that characters can always be processed in isolation.

Implementations that choose to go beyond the ISO/ANSI C model may find it useful to mix widths within their APIs.

For example, an implementation may have a 32-bit wchar_t and process strings in any of UTF-8, UTF-16 or UTF-32 forms.  Another implementation may have a 16-bit wchar_t and process strings as UTF-8 or UTF-16, but have additional APIs that process individual characters as UTF-32, or deal with pairs of UTF-16 code units.