> any task that speed is even remotely a consideration for anymore
How do you know whether or not speed is a consideration?
Yes, OP delivered impressive efficiency gains. I'm sure he could improve the efficiency even more by dropping into pure Assembly.
But is it worth it?
The prime consideration is not execution speed but maintainability. The further that OP got away from pure Python, the more difficult to maintain the code became. That's a downside.
Now, OP describes an important technique because in the real world, you have a performance budget. Code needs to execute at speeds that return quickly enough to the user, or long execution is financially expensive (i.e. cloud computing resources), etc. But optimizing beyond what the budget requires is wasteful in terms of time needed to do the optimization as well as harmful in terms of negatively impacting future maintainability.
> The prime consideration is not execution speed but maintainability.
Why? And how did you measure this drop in maintainability? I'm asking because I see developers prioritize _perceived_ maintainability over _measurable_ things that matter to the user (like performance).
How do you know whether or not speed is a consideration?
Yes, OP delivered impressive efficiency gains. I'm sure he could improve the efficiency even more by dropping into pure Assembly.
But is it worth it?
The prime consideration is not execution speed but maintainability. The further that OP got away from pure Python, the more difficult to maintain the code became. That's a downside.
Now, OP describes an important technique because in the real world, you have a performance budget. Code needs to execute at speeds that return quickly enough to the user, or long execution is financially expensive (i.e. cloud computing resources), etc. But optimizing beyond what the budget requires is wasteful in terms of time needed to do the optimization as well as harmful in terms of negatively impacting future maintainability.