bear <xxxxxx@sonic.net> writes:

> This is why I think "glyph=character" is more or less required
> for a language that wants to have "single character" arguments to
> and returns from case mapping functions.  You cannot do it in a
> fashion compliant to the Unicode standard if you use
> "codepoint=character" instead.

We cannot have single character arguments for case mapping functions.
You can't even do it if you have "glyph=character", because there are
case mappings which map multiple characters.  (Es-zet being the
classic, but hardly the only, example.)

> The best we can hope for, and the point of language design, is to make
> it easy to do the right thing; there is no way we can prevent
> programmers from doing the wrong thing if they're so inclined, nor
> even make it the slightest bit difficult.  Our duty is to make sure
> the right thing is no harder to do than the wrong thing.

We should also not provide functions which are automatically the wrong
thing.