On Exceptions and Rules: Best Practices for Where, When, and How to Handle Exceptions

If an exception is raised in the declaration section of one's block, that block's exception section cannot handle the exception. That doesn't seem right. Why does PL/SQL work this way, and what does it mean for the author's coding practices? If an exception is raised in the declaration section of one's block, that exception will propagate out of the block unhandled. PL/SQL behaves like this (or, to be more accurate, the Oracle PL/SQL development team decided to implement exception handling like this) because until local variables and constants are fully elaborated, one doesn't have a viable subprogram with which to work. Suppose the declaration-raised exception were handled inside that subprogram. To what could one refer inside the exception handler?