Ask HN: What's stopping us from having full static validation of Python code?

[diegojromero]

I have developed two mypy plugins[1][2] for Python to help with static checks.

I was wondering, how far are we with providing such a high level of static checks for interpreted languages that almost all issues can be catch statically? Is there any work on that on any interpreted programming language? What would you like to be statically-checked in Python?

[1] https://github.com/diegojromerolopez/mypy-pure

[2] https://github.com/diegojromerolopez/mypy-raise

[vintagedave]

I know there are standard answers to why, but by crikey do I agree this would be enormously valuable.

I’ll check out your repos. Can you share more about what they can and can’t do, your approach, your goals…?

[diegojromero]

Sure, in my workflow I make use of mypy and ruff extensively (but ruff does not allow plugins). So I decided to create several plugins with the intention of improving the developer experience and understanding of the code.

Now, what I am asking is, could we expand this by having static annotations that enforce runtime behaviours before running the code? I'm not advocating for the removal of the tests or some manual testing phase, but in the same vein of type errors being shown in our IDE, could we have static checks like contracts for functions or class invariants?

Full disclaimer: I leveraged AI technologies to create them.