A type system can refuse to turn a `bytes` into a `utf8str` until it's been appropriately parsed.

(It doesn't even need to be a very good or strongly-enforced type system - Go makes it dangerously easy to convert between `[]byte` and `string` by other-type-system standards, and yet everything works pretty well. It's enough to hitch your thinking and make you realize you need another step.)

How does the type system knows the bytes are utf8 and not CP850?

The same way you can parse a string into an integer and it knows you got an either[integer, err].

So it doesn't, and it's not about the type system. Any decode operation in python 3 raise an exception or return string.

It's about the runtime error handling making it mandatory to deal with the error and won't let you panic at run time for this specific error.

I'm sorry you don't see exceptions as part of a type system yet.

The compiler can raise a warning about bad characters. It can't detect all problems, but it certainly can help with some.

It's not the type system, but many dynamic languages are interpreted instead of compiled.

Those bad characters are fed into the program long long after the compiler has done its job.

Not necessarily? I'm talking about constant string literals in source code which can be validated at compile time.