Given that the use of var will automatically assign undefined to the associated variable before execution of the script, it stands to reason that the script must be parsed before it's run procedurally by the interpreter, right?

As a follow-up, assuming the previous statement is true, and given that when you switched const username to let username and re-assigned an identical object to username thereby instructing the interpreter to create a new object (allocating new memory and creating more work for garbage collection) and a new reference to that object at the memory location represented by username, why can't the interpreter identify that the new object is identical to the old object and optimise by essentially... doing nothing?

In short, if the script is parsed before execution, why can't the interpreter produce an optimised set of instructions that simply does not re-instantiate and reassign objects (or reference types, generally) whose functional identities are equal?

Apologies for the verbose question; I hope it's sensical. I'm just curious whether this is by design or simply an impossibility, given the nature of interpreted languages. I'm a little rusty regarding my fundamentals of computer science.

Let and Const