* I can't remember where I saw precedent for quoting the class name.
* Built-in functions like `int` and `float` use the `foo()` style.
* Con: no longer unambiguous no matter what __name__ is set to.
* Pro: errors now communicate in a more conventional, familiar way.