On Aug 30, 2020, at 6:42 PM, Shiro Kawai <xxxxxx@gmail.com> wrote:

If so, (nan= special-nan.f special-nan.d) doesn't tell anything except a specific implementation-dependent behavior. 

I agree.

Won't harm, but it seems just as well to say passing different precisions of NaNs won't raise an error but the result is implementation-dependent.

I believe that raising an error should also be a valid option (if one is trying to write a platform-independent library).

Brad