right, a lot of python code either doesn't result in libraries (or even documentation), or library APIs often tend to focus on dicts or integral types. successful libraries like numpy and pandas focus on a very small set of well-defined objects and leverage duck typing for containers (and runtime casting for interop with native code).
I'm not saying arguments in favor of static typing and annotations are invalid, but rather it's a mindset that can make much of python look very foreign. the norvig spellchecker is a key example of swath of python code where static types are essentially irrelevant.
I'm not saying arguments in favor of static typing and annotations are invalid, but rather it's a mindset that can make much of python look very foreign. the norvig spellchecker is a key example of swath of python code where static types are essentially irrelevant.