[Python-Dev] Type hints -- a mediocre programmer's reaction

[Antoine Pitrou]

On Tue, 21 Apr 2015 18:27:50 +0300
Paul Sokolovsky <pmiscml at gmail.com> wrote:
> Let me try: MicroPython already uses type annotations for statically
> typed functions. E.g.
> 
> def add(x:int, y:int):
>     return x + y
> 
> will translate the function to just 2 machine instructions.

That's quite nice.

> Oh really, you care to support single-precisions in Numba?

Numba is quite specialized. In our area, single-precision arithmetic
can sometimes give double the speed on modern CPUs (especially when
LLVM is able to vectorize the code). The speedup can be even larger on
a GPU (where double-precision execution resources are scarce). I agree
it doesn't make sense for a general-purpose Python compiler.

> Anyway, back to your example, it would be done like:
> 
> SFloat = float
> DFloat = float
> 
> For a random tool out there, "SFloat" and "DFloat" would be just
> aliases to floats, but Numba will know they have additional semantics
> behind them. (That assumes that typedefs like SFloat can be accessed in
> symbolic form - that's certainly possible if you have your own
> parser/VM, but might worth to think how to do it on "CPython" level).

Fortunately, we don't have our own parser :-) We work from the CPython
bytecode.

Regards

Antoine.