atan2 as 2-arg method of atan

[simonbyrne]

Should atan2 become a 2-argument method of atan?

The main argument against would be convention: atan2 is well-established, going back at least to Fortran IV (1961).

The only languages I have found that do combine them are Mathematica ArcTan and Common Lisp atan.

[StefanKarpinski]

We could at least allow atan(y, x) as another way of writing atan2(y, x).

[jrevels]

The convention is cemented enough that it has its own wikipedia entry under https://en.wikipedia.org/wiki/Atan2. The discoverability factor there might be a point in favor of keeping the current name.

[StefanKarpinski]

I would favor keeping two-argument atan2 but also allowing atan(y, x) (non-breaking).

[jrevels]

I'm just bikeshedding here, but I don't think the proposed change renders a benefit that outweighs the detriment of exporting multiple names for the same operation. My preference is just to keep atan2 due to convention.

[Sacha0]

I'm just bikeshedding here, but I don't think the proposed change renders a benefit that outweighs the detriment of exporting multiple names for the same operation. My preference is just to keep atan2 due to convention.

I would begin arguing along the same lines as Jarrett, but instead conclude that discarding the relic of past programming languages' constraints that is atan2 is best 😉.