Robert Dodier wrote:
>
> I'm pretty much opposed to ifft at this point. Naming it the same
> seems pointless given all the other differences with Matlab/Octave,
> notably the definition of the transform differs with Octave
> (normalization is different). Wouldn't be surprised if Matlab is
> different too.
>
Ok. My point wasn't that we should use the same name as Matlab/Octave,
but that people needing to use fft are doing numeric computations and my
own experience/expectation is that the inverse is ifft. But fft_inverse
is ok too.
> I';m partial to big-endian names but inverse_fft is OK too.
> (I see there are already inverse_jacobi_foo functions.
>
I named those, I guess, because when I see sin^{-1}, I pronounce it
inverse sin, not sin inverse. (Well, maybe arcsin, but I don't think
anybody says arc jacobi sn.)
Ray