Wooster

Or maybe `most_precise: bool` or `precisest: bool` or something?

I think this is starting to look pretty good now. We're using a builtin function that most users wouldn't think of to provide speed and accuracy. I think this is...

Alright. I too think it's perfectly fine to just go with the slightly "less precise" option even though it's still really accurate. In practical usage I can't really think of...

Perfect.

Alright so the first thing I can tell from the CI failures is that there are some problems with `f16` for Wasm here... This line in the "lerp" test in...

> * if an additional function for lerp for integers would be added, what might be its fully qualified name in the std lib, and what might its function signature...

I should also note that although the codegen for this looks great on x86, this may codegen poorly on platforms that don't have a native FMA: ```zig const std =...

Or actually, could we possibly make use of the currently set float mode here? There is enum `std.builtin.FloatMode` with values `Strict` and `Optimized`, and we have `@setFloatMode` to set the...

Vectors with a scalar value? So 2 vectors each with 1 scalar? In that case you can do something like this right? ```zig _ = lerp(f32, (@Vector(1, f32){0.5})[0], (@Vector(1, f32){1.0})[0],...

> I would just keep using my own implementation > I'm also in favor of using the more precise algorithm Clearly we're all so split on the implementation on this...