[R] Subject: Re: BUG: atan(1i) / 5 = NaN+Infi ?

[Duncan Murdoch]

On 2024-09-12 11:08 a.m., Jonathan Dushoff wrote:
>> In this case, I do think we should look into the consequences of
>> indeed distinguishing
>>    <double> * <complex>
>>    <complex> * <double>  and
>>    <complex> / <double>
>> from their respective current {1. coerce to complex, 2. use complex arith}
>> arithmetic.
> 
> I'm wondering whether – if this indeed gets opened up – it might also
> make sense to calculate <double> x / <complex> y using real arithmetic
> (as x*y / |y|²)

That's not the correct formula, is it?  I think the result should be x * 
Conj(y) / Mod(y)^2 . So that would involve <double> * <complex> and 
<complex> / <double>, not just real arithmetic.

Duncan Murdoch