← Back to context

Comment by nextaccountic

2 years ago

Yeah if a codebase is full of stuff like this, auditing it is awful. It's like, instead of employing computers to check the details your code, force it to be done manually (in an error prone way)

This is nonsensical. When you use a function, how do you know what it will do? You guess from its name?

> auditing it is awful.

If a function specifies a requirement, you look at the callers and see if that requirement is met. If it's easy to verify in code, you can assert. Is there an easier way to audit correctness?